Blog

📚 Econometric Analysis Methods and Regression Models | 计量分析方法与回归模型

Econometrics is the application of statistical methods and mathematics to economic data, enabling economists to estimate relationships, test theories, and forecast future outcomes. For A-Level Economics students, understanding the basic tools of regression analysis is essential for interpreting research articles and for answering data-response questions on the effects of variables such as price, income, and taxation.

计量经济学是将统计学方法与数学方法应用于经济数据的一门学科，使经济学家能够估计经济关系、检验理论并对未来结果进行预测。对于 A-Level 经济学学生而言，理解回归分析的基本工具，对于解读研究文献以及回答涉及价格、收入、税收等变量影响的数据分析题至关重要。

---

1. The Role of Econometrics in Economics | 计量经济学在经济学中的作用

Economics is often concerned with cause and effect: how does a change in the interest rate affect investment, or how does a rise in the minimum wage affect employment? Econometrics provides a framework for translating economic theory into testable equations, and for quantifying the size of these effects from observable data.

经济学经常关注因果关系：利率变动如何影响投资？最低工资上升如何影响就业？计量经济学为把经济理论转化为可检验的方程、并从可观测数据中量化这些效应的大小提供了框架。

In A-Level Economics, you are not expected to derive advanced estimators, but you should be able to read a regression output, interpret coefficients, and recognise the difference between a statistical relationship and a real economic effect.

在 A-Level 经济学中，并不要求你推导高级估计量，但你应该能够阅读回归输出、解释系数，并认识到统计关系与真实经济效应之间的区别。

---

2. Correlation and Causation | 相关性与因果性

A scatter diagram can show whether two variables move together. Correlation measures the strength and direction of a linear relationship, but correlation alone does not prove causation. A finding that ice-cream sales and drowning deaths are strongly correlated does not mean ice-cream causes drowning; the common cause is hot weather.

散点图可以显示两个变量是否一起变动。相关性衡量线性关系的强度和方向，但仅有相关性并不能证明因果关系。如果发现冰淇淋销量与溺水死亡人数高度相关，这并不意味着冰淇淋导致溺水；其共同原因是天气炎热。

The two main pitfalls are omitted variables and reverse causation. An omitted variable is an unmeasured factor influencing both x and y, while reverse causation means y causes x rather than x causes y. When interpreting regression results, always ask whether a third factor or a feedback effect may be present.

两个主要陷阱是遗漏变量和反向因果。遗漏变量是指同时影响 x 和 y 的未测量因素；反向因果则意味着是 y 引起了 x，而不是 x 引起 y。在解读回归结果时，要经常思考是否可能存在第三个因素或反馈效应。

---

3. Scatter Diagrams and the Correlation Coefficient | 散点图与相关系数

The Pearson correlation coefficient r summarises the direction and strength of a linear association. It always lies between -1 and +1. A value of +1 indicates a perfect positive linear relationship, -1 a perfect negative linear relationship, and 0 no linear relationship.

皮尔逊相关系数 r 概括了线性关联的方向和强度。它的取值始终在 -1 与 +1 之间。+1 表示完全正线性关系，-1 表示完全负线性关系，0 表示没有线性关系。

r = Σ(xᵢ − x̄)(yᵢ − ȳ) / √[Σ(xᵢ − x̄)² Σ(yᵢ − ȳ)²]

Values close to +1 or -1 suggest a strong linear association, while values close to zero indicate a weak association. Note that r is sensitive to outliers and is only valid for linear relationships; a curved relationship may produce a low r even when the variables are closely linked.

接近 +1 或 -1 的值表明线性关联很强，而接近 0 的值表明关联较弱。注意 r 对离群值很敏感，并且只对线性关系有效；即使变量之间联系紧密，曲线关系也可能导致较低的 r 值。

---

4. The Simple Linear Regression Model | 简单线性回归模型

The population relationship between an independent variable x and a dependent variable y is written as y = β₀ + β₁x + ε, where β₀ is the intercept, β₁ the slope, and ε the error term capturing all other influences. In practice we estimate the sample regression line: ŷ = b₀ + b₁x.

自变量 x 与因变量 y 之间的总体关系写为 y = β₀ + β₁x + ε，其中 β₀ 是截距，β₁ 是斜率，ε 是捕捉所有其他影响的误差项。在实践中我们估计样本回归线：ŷ = b₀ + b₁x。

y = β₀ + β₁x + ε

The estimated slope b₁ is the key economic quantity. It tells us the expected change in y for a one-unit increase in x, holding other factors constant. For example, if y is consumption and x is disposable income, b₁ is the marginal propensity to consume.

估计出的斜率 b₁ 是核心经济数量。它告诉我们，在其他因素保持不变的条件下，x 每增加一个单位，y 预期改变多少。例如，若 y 是消费，x 是可支配收入，则 b₁ 就是边际消费倾向。

---

5. The Method of Least Squares | 最小二乘法

The ordinary least squares (OLS) method chooses b₀ and b₁ to minimise the sum of squared vertical distances between the data points and the regression line. Each vertical distance is called a residual, eᵢ = yᵢ − ŷᵢ.

普通最小二乘法选择 b₀ 和 b₁，使数据点与回归线之间垂直距离的平方和最小。每个垂直距离称为残差，eᵢ = yᵢ − ŷᵢ。

minimise Σ(yᵢ − b₀ − b₁xᵢ)²

The resulting OLS estimates are unbiased and consistent under the Gauss-Markov assumptions: zero mean error, constant variance, no correlation among errors, and no perfect multicollinearity. In A-Level terms, this means the ordinary straight line fitted through the data is the best linear approximation in the least squares sense.

在高斯-马尔可夫假设（误差均值为零、方差恒定、误差之间不相关、不存在完全多重共线性）下，OLS 估计是无偏且一致的。用 A-Level 的语言说，这意味着穿过数据的普通直线是最小二乘意义上的最佳线性近似。

---

6. Goodness of Fit: R² and Adjusted R² | 拟合优度：R² 与调整后 R²

The coefficient of determination, R², measures the proportion of the variation in y that is explained by the regression. It ranges from 0 to 1 in simple regression. An R² of 0.80 means 80% of the sample variation in y is accounted for by x.

决定系数 R² 衡量 y 的变动中被回归所解释的比例。在简单回归中它介于 0 与 1 之间。R² = 0.80 意味着 y 的样本变动中有 80% 可由 x 解释。

One weakness of ordinary R² is that adding more explanatory variables always raises it. The adjusted R² penalises additional predictors and is useful when comparing models. A high R² is desirable, but it does not guarantee that the model is economically meaningful or that the coefficients are unbiased.

普通 R² 的一个弱点是增加更多解释变量总会提高它。调整后 R² 会对新增预测变量施加惩罚，因此在比较模型时很有用。R² 高是好事，但它并不能保证模型具有经济含义或系数无偏。

---

7. Hypothesis Testing and Statistical Significance | 假设检验与统计显著性

To test whether x really has an effect on y, we set the null hypothesis H₀: β₁ = 0 against the alternative H₁: β₁ ≠ 0. The t-statistic is t = b₁ / SE(b₁), where SE(b₁) is the standard error. If the p-value is below the chosen significance level (usually 0.05), we reject the null and call the coefficient statistically significant.

要检验 x 是否真的对 y 有影响，我们设置原假设 H₀: β₁ = 0，备择假设 H₁: β₁ ≠ 0。t 统计量为 t = b₁ / SE(b₁)，其中 SE(b₁) 是标准误。若 p 值低于所选的显著性水平（通常为 0.05），我们就拒绝原假设，称该系数在统计上显著。

t = b₁ / SE(b₁)

Significance tells us that the relationship is unlikely to be due to random sampling, but it does not tell us whether the effect is economically important. A tiny coefficient with a very small p-value can be statistically significant yet practically irrelevant.

显著性告诉我们这种关系不太可能由随机抽样造成，但它并不能告诉我们效应在经济上是否重要。一个很小的系数即使 p 值非常小，也可能在统计上显著但在实际中无关紧要。

---

8. Types of Economic Data | 经济数据类型

Economic data come mainly in three forms. Cross-sectional data record many units at one point in time, such as household incomes across regions in 2023. Time-series data track one unit over time, such as UK GDP from 1990 to 2023. Panel data combine both, following many units over time, such as GDP per capita for several countries over two decades.

经济数据主要有三种形式。截面数据在某一时点记录许多单位，例如 2023 年各地区的家庭收入。时间序列数据在时间上追踪一个单位，例如 1990 至 2023

Published by TutorHao | Economics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

📚 Differential Equations Fundamentals and Solution Strategies | 微分方程基础与求解策略

A differential equation is a mathematical equation that relates a function with its derivatives. In A-Level mathematics, differential equations are a core topic in both pure mathematics and applied contexts, forming the bridge between algebraic manipulation and real-world modeling.

微分方程是联系函数与其导数的数学方程。在A-Level数学中，微分方程是纯数学与应用情境中的核心考点，是连接代数运算与现实建模的桥梁。

---

1. What Is a Differential Equation? | 什么是微分方程？

A differential equation involves an unknown function and one or more of its derivatives. The order of a differential equation is determined by the highest derivative present. For example, dy/dx = 2x is a first-order equation, while d²y/dx² + y = 0 is second-order.

微分方程包含一个未知函数及其一个或多个导数。阶数由方程中最高阶导数决定。例如，dy/dx = 2x 是一阶方程，而 d²y/dx² + y = 0 是二阶方程。

The degree of a differential equation is the power of the highest-order derivative after the equation has been made rational and integral in all derivatives. Most A-Level questions involve first-degree (linear) differential equations.

次数是在方程化为有理整式后最高阶导数的幂次。大多数A-Level题目涉及一次（线性）微分方程。

A general solution contains arbitrary constants; a particular solution is obtained by substituting initial conditions or boundary conditions to find those constants.

通解含有任意常数；通过代入初始条件或边界条件求得具体常数后，得到特解。

---

2. Classifying Differential Equations | 微分方程的分类

In the A-Level syllabus, differential equations are classified primarily in two ways: by order (first-order or second-order) and by linearity (linear or non-linear). A linear differential equation has the unknown function and its derivatives appearing only to the first power, with no products of these terms.

在A-Level大纲中，微分方程主要从两个角度分类：按阶数（一阶或二阶）和按线性（线性或非线性）。线性微分方程中未知函数及其导数只以一次幂出现，且不包含它们的乘积项。

A first-order linear differential equation can be written in the general form:

dy/dx + P(x)y = Q(x)

where P(x) and Q(x) are functions of x only. Similarly, a second-order linear differential equation has the form a·d²y/dx² + b·dy/dx + c·y = f(x), where a, b, c are constants.

一阶线性微分方程可写成一般形式：

dy/dx + P(x)y = Q(x)

其中 P(x) 和 Q(x) 仅是 x 的函数。类似地，二阶线性微分方程具有形式 a·d²y/dx² + b·dy/dx + c·y = f(x)，其中 a、b、c 为常数。

---

3. Method of Separation of Variables | 分离变量法

The separation of variables method is applicable when a first-order differential equation can be rearranged so that all terms involving y (and dy) are on one side, and all terms involving x (and dx) are on the other side. This technique is a cornerstone of the A-Level 微分方程 unit.

分离变量法适用于可以整理为一边只含 y（和 dy）、另一边只含 x（和 dx）的一阶微分方程。这一技巧是A-Level微分方程单元的基石。

Step 1: Rearrange the equation into the form f(y) dy = g(x) dx.

步骤一：将方程整理为 f(y) dy = g(x) dx 的形式。

Step 2: Integrate both sides: ∫ f(y) dy = ∫ g(x) dx.

步骤二：两边同时积分：∫ f(y) dy = ∫ g(x) dx。

Step 3: Add the constant of integration C and solve for y explicitly if possible.

步骤三：加上积分常数 C，尽可能显式解出 y。

Consider the worked example: dy/dx = 3x²y. Separating variables gives (1/y) dy = 3x² dx. Integrating both sides yields ln|y| = x³ + C, hence y = Ae^(x³), where A = ±e^C.

看一个经典例题：dy/dx = 3x²y。分离变量得到 (1/y) dy = 3x² dx。两边积分得 ln|y| = x³ + C，因此 y = Ae^(x³)，其中 A = ±e^C。

Key examination tip: always include the absolute value in ln|y|, and rewrite ±e^C as a single constant A before finalizing the answer.

关键考试提示：ln 中始终保留绝对值 |y|，并在最终答案前将 ±e^C 合并为单一常数 A。

---

4. Integrating Factor Method | 积分因子法

When a first-order differential equation of the form dy/dx + P(x)y = Q(x) cannot be separated, the integrating factor method is the standard approach. The integrating factor is defined as:

当形如 dy/dx + P(x)y = Q(x) 的一阶微分方程无法分离变量时，积分因子法成为标准解法。积分因子定义为：

I = e^(∫ P(x) dx)

Multiplying both sides of the differential equation by I converts the left-hand side into the derivative of I·y:

将微分方程两边同乘以 I，可将左边化为 I·y 的导数：

d/dx (I·y) = I·Q(x)

Then integrate both sides: I·y = ∫ I·Q(x) dx + C, and finally divide by I to obtain y.

然后两边积分：I·y = ∫ I·Q(x) dx + C，最后除以 I 得到 y。

Let us examine an example: dy/dx + 2y/x = x³. Here P(x) = 2/x. The integrating factor is I = e^(∫ 2/x dx) = e^(2 ln|x|) = x². Multiplying through gives x² · dy/dx + 2xy = x⁵, which simplifies to d/dx (x²y) = x⁵. Integrating: x²y = x⁶/6 + C, so y = x⁴/6 + C/x².

我们来看一个例子：dy/dx + 2y/x = x³。这里 P(x) = 2/x。积分因子为 I = e^(∫ 2/x dx) = e^(2 ln|x|) = x²。两边同乘得 x² · dy/dx + 2xy = x⁵，即 d/dx (x²y) = x⁵。积分：x²y = x⁶/6 + C，所以 y = x⁴/6 + C/x²。

---

5. Second-Order Homogeneous Equations | 二阶齐次方程

A second-order homogeneous linear differential equation with constant coefficients has the form a·d²y/dx² + b·dy/dx + c·y = 0. To solve it, we assume a solution of the form y = e^(λx), which leads to the auxiliary equation:

常系数二阶齐次线性微分方程具有形式 a·d²y/dx² + b·dy/dx + c·y = 0。求解时假设解具有形式 y = e^(λx)，从而得到辅助方程：

aλ² + bλ + c = 0

There are three cases to consider based on the discriminant Δ = b² − 4ac:

根据判别式 Δ = b² − 4ac 分为三种情况：

• If Δ > 0: two distinct real roots λ₁, λ₂, giving the general solution y = Ae^(λ₁x) + Be^(λ₂x).

  若 Δ > 0：两个不等实根 λ₁、λ₂，通解为 y = Ae^(λ₁x) + Be^(λ₂x)。

• If Δ = 0: one repeated real root λ, giving y = (A + Bx)e^(λx).

  若 Δ = 0：一个二重实根 λ，通解为 y = (A + Bx)e^(λx)。

• If Δ < 0: complex conjugate roots λ = p ± qi, giving y = e^(px)(A cos(qx) + B sin(qx)).

  若 Δ < 0：共轭复根 λ = p ± qi，通解为 y = e^(px)(A cos(qx) + B sin(qx))。

For example, the equation d²y/dx² − 5·dy/dx + 6y = 0 has auxiliary equation λ² − 5λ + 6 = 0 = (λ − 2)(λ − 3). Thus the general solution is y = Ae^(2x) + Be^(3x).

例如，方程 d²y/dx² − 5·dy/dx + 6y = 0 的辅助方程为 λ² − 5λ + 6 = 0 = (λ − 2)(λ − 3)。因此通解为 y = Ae^(2x) + Be^(3x)。

---

6. Second-Order Non-Homogeneous Equations | 二阶非齐次方程

A non-homogeneous equation has the form a·d²y/dx² + b·dy/dx + c·y = f(x). The general solution is the sum of the complementary function (CF, the solution of the corresponding homogeneous equation) and a particular integral (PI, any one solution of the full equation).

非齐次方程具有形式 a·d²y/dx² + b·dy/dx + c·y = f(x)。通解为互补函数（CF，对应齐次方程的解）与特解（PI，原方程的任意一个解）之和。

The table below summarises the standard trial functions for choosing a PI:

下表总结了选择特解时的标准试探函数：

f(x) 的形式	试探特解形式
多项式（如 x²）	同次数多项式（如 ax² + bx + c）
指数型（如 e^(kx)）	Ae^(kx)（若 k 是辅助方程的根，则需乘 x）
三角函数（如 sin(kx) 或 cos(kx)）	A sin(kx) + B cos(kx)

Consider d²y/dx² + y = 3 sin(2x). The CF is y = A cos(x) + B sin(x). For the PI, try y = p sin(2x) + q cos(2x). Substitution gives (−4p sin(2x) − 4q cos(2x)) + (p sin(2x) + q cos(2x)) = 3 sin(2x). Therefore −3p = 3 and −3q = 0, so p = −1, q = 0. The PI is y = −sin(2x), and the general solution is y = A cos(x) + B sin(x) − sin(2x).

考虑 d²y/dx² + y = 3 sin(2x)。CF 为 y = A cos(x) + B sin(x)。对特解尝试 y = p sin(2x) + q cos(2x)。代入得 (−4p sin(2x) − 4q cos(2x)) + (p sin(2x) + q cos(2x)) = 3 sin(2x)。因此 −3p = 3、−3q = 0，所以 p = −1、q = 0。特解为 y = −sin(2x)，通解为 y = A cos(x) + B sin(x) − sin(2x)。

Special caution: if the trial function duplicates a term in the CF, multiply the trial by x (or x² in resonance cases) to avoid linear dependence.

特别注意：若试探解与CF中的某一项重复，需将试探解乘以 x（共振情形下乘 x²）以避免线性相关。

---

7. Initial Value Problems and Boundary Conditions | 初值问题与边界条件

An initial value problem specifies conditions at a single point, typically y(x₀) = y₀ and y′(x₀) = y₁. A boundary value problem specifies conditions at two different points. Both are resolved by substituting the conditions into the general solution to find the arbitrary constants.

初值问题在单个点处给出条件，通常为 y(x₀) = y₀ 和 y′(x₀) = y₁。边值问题在两个不同点处给出条件。两者的求解方法都是将条件代入通解以确定任意常数。

Worked example: Solve dy/dx = 6x² − 4x with y(1) = 5. Integrating gives y = 2x³ − 2x² + C. Substituting x = 1, y = 5 gives 5 = 2 − 2 + C, so C = 5. Therefore y = 2x³ − 2x² + 5.

例题：求解 dy/dx = 6x² − 4x，且 y(1) = 5。积分得 y = 2x³ − 2x² + C。代入 x = 1、y = 5 得 5 = 2 − 2 + C，即 C = 5。因此 y = 2x³ − 2x² + 5。

For second-order initial value problems, both y and dy/dx at x₀ are used to determine the two constants A and B simultaneously.

对于二阶初值问题，利用 x₀ 处的 y 和 dy/dx 两个条件联立求出常数 A 和 B。

---

8. Applications in Real-World Modelling | 微分方程的实际建模应用

Differential equations are powerful tools for modelling real-world phenomena. In A-Level examinations, common applications include population growth (unrestricted and restricted), radioactive decay, Newton’s law of cooling, and simple harmonic motion.

微分方程是模拟现实世界现象的有力工具。在A-Level考试中，常见应用包括人口增长（无限制与有限制）、放射性衰变、牛顿冷却定律和简谐运动。

Exponential growth/decay: dN/dt = kN, where k is the growth (k > 0) or decay (k < 0) constant. The solution is N = N₀e^(kt), where N₀ is the initial population.

指数增长/衰减：dN/dt = kN，其中 k 为增长（k > 0）或衰减（k < 0）常数。解为 N = N₀e^(kt)，其中 N₀ 为初始数量。

Restricted growth (logistic model): dN/dt = kN(1 − N/M), where M is the carrying capacity. This yields a sigmoidal growth curve that plateaus at N = M.

有限增长（逻辑斯蒂模型）：dN/dt = kN(1 − N/M)，其中 M 为环境容量。其解为一条在 N = M 处趋于平稳的S形增长曲线。

Newton’s law of cooling: dθ/dt = −k(θ − θₑ), where θ is the object temperature and θₑ is the ambient temperature. The solution is θ = θₑ + (θ₀ − θₑ)e^(−kt).

牛顿冷却定律：dθ/dt = −k(θ − θₑ)，其中 θ 为物体温度、θₑ 为环境温度。解为 θ = θₑ + (θ₀ − θₑ)e^(−kt)。

For second-order applications, simple harmonic motion is described by d²x/dt² = −ω²x, whose solution is x = A cos(ωt) + B sin(ωt), or equivalently x = R cos(ωt + φ).

在二阶应用中，简谐运动由 d²x/dt² = −ω²x 描述，其通解为 x = A cos(ωt) + B sin(ωt)，等价形式为 x = R cos(ωt + φ)。

---

9. Common Pitfalls and Exam Strategies | 常见错误与考试策略

Students frequently lose marks on differential equation questions due to avoidable errors. Being aware of these pitfalls before the exam is half the battle won.

学生常在微分方程题目中因可避免的错误而失分。考前了解这些常见陷阱，等于赢了一半。

• Forgetting the constant of integration: always include C after every integration in indefinite integrals.

  忘记积分常数：不定积分后务必加上常数 C。

• Incorrectly separating variables: if the equation cannot be written as f(y)dy = g(x)dx, do not force it; use the integrating factor instead.

  不当分离变量：若方程无法写成 f(y)dy = g(x)dx 的形式，不要强行分离，应改用积分因子法。

• Sign errors when substituting the integrating factor: double-check that P(x) is transcribed correctly and the exponent sign matches.

  代入积分因子时符号错误：仔细核对 P(x) 是否抄写正确，指数符号是否对应。

• Wrong trial function for PI: if the standard trial form is also a solution of the homogeneous equation, multiply by x (or x²) until it is independent of the CF.

  特解试探函数选择错误：若标准试探形式恰好是齐次方程的解，则需乘以 x（或 x²）直到与CF线性无关。

• Not applying both initial conditions for second-order equations: both y and y′ conditions must be used to determine A and B.

  二阶方程未同时利用两个初始条件：必须使用 y 和 y′ 两个条件联立求出 A 和 B。

Exam strategy: always state the auxiliary equation explicitly, show the general solution before applying conditions, and verify the particular solution by substituting back where feasible.

考试策略：始终明确写出辅助方程，先写出通解再应用条件，并在可行时将特解代回验证。

---

10. Worked Example Combining Multiple Skills | 综合技能例题

Let us now solve a comprehensive problem that integrates several techniques studied above.

下面我们求解一道综合运用以上多种技巧的完整题目。

Problem: Solve the differential equation d²y/dx² − 4y = 8x, subject to y(0) = 2 and y′(0) = 4.

题目：求解微分方程 d²y/dx² − 4y = 8x，满足条件 y(0) = 2 和 y′(0) = 4。

Solution:

解答：

Step 1 — complementary function. The auxiliary equation is λ² − 4 = 0, giving λ = ±2. Hence CF = Ae^(2x) + Be^(−2x).

第一步——互补函数：辅助方程为 λ² − 4 = 0，得 λ = ±2。因此 CF = Ae^(2x) + Be^(−2x)。

Step 2 — particular integral. Since f(x) = 8x is linear, try PI = ax + b. Then d²(PI)/dx² = 0, and substituting gives −4(ax + b) = 8x, so −4a = 8 and −4b = 0, yielding a = −2, b = 0. PI = −2x.

第二步——特解：因 f(x) = 8x 为一次多项式，尝试 PI = ax + b。则 d²(PI)/dx² = 0，代入得 −4(ax + b) = 8x，即 −4a = 8、−4b = 0，解得 a = −2、b = 0。PI = −2x。

Step 3 — general solution: y = Ae^(2x) + Be^(−2x) − 2x.

第三步——通解：y = Ae^(2x) + Be^(−2x) − 2x。

Step 4 — apply conditions. y(0) = A + B = 2. Also y′ = 2Ae^(2x) − 2Be^(−2x) − 2, so y′(0) = 2A − 2B − 2 = 4, giving A − B = 3. Solving the system A + B = 2 and A − B = 3 yields A = 5/2, B = −1/2.

第四步——代入条件：y(0) = A + B = 2。又 y′ = 2Ae^(2x) − 2Be^(−2x) − 2，所以 y′(0) = 2A − 2B − 2 = 4，即 A − B = 3。联立 A + B = 2 与 A − B = 3，得 A = 5/2、B = −1/2。

Final answer:

最终答案：

y = (5/2)e^(2x) − (1/2)e^(−2x) − 2x

This example demonstrates the full workflow: finding the CF, constructing the PI, combining them, and solving for constants using boundary conditions.

该例题展示了完整流程：求互补函数、构造特解、合并为通解，并利用边界条件求解常数。

---

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

📚 Introduction to Functions: Concepts and Graphs | 初级函数的概念与图像

In mathematics, functions are one of the most fundamental ideas, connecting input values to output values through a precise rule. Understanding what a function is — and how to interpret its graph — lays the foundation for algebra, calculus, and beyond.

在数学中，函数是最基本的概念之一，通过一个精确的规则将输入值连接到输出值。理解什么是函数，以及如何解读它的图像，为代数、微积分以及更高级的数学打下基础。

---

1. What Is a Function? | 什么是函数？

A function is a relation that assigns exactly one output value to each valid input value. In other words, for every input (x) in the domain, there corresponds exactly one output (y) in the range.

函数是一种关系，它为每一个有效的输入值分配唯一的一个输出值。换句话说，对于定义域中的每一个输入 (x)，在值域中恰好有一个对应的输出 (y)。

f: x → y, 其中 y = f(x), 且每个 x 有且仅有一个 y

Think of a function as a machine: you feed it an input, it processes the rule, and out comes exactly one result. If a single input can produce two different outputs, the relation is not a function.

可以把函数想象成一台机器：您输入一个值，它按照规则运算，输出唯一的结果。如果同一个输入产生两个不同的输出，那么这个关系就不是函数。

---

2. Domain and Range | 定义域与值域

The domain of a function is the complete set of possible input values, while the range is the complete set of possible output values. For example, in the function f(x) = x², the domain is all real numbers, but the range is only non-negative numbers.

函数的定义域是所有可能输入值的集合，而值域是所有可能输出值的集合。例如，在函数 f(x) = x² 中，定义域为全体实数，但值域仅为非负实数。

• Domain / 定义域：The set of all allowable x-values / 所有允许的 x 值的集合
• Range / 值域：The set of all resulting y-values / 所有由此产生的 y 值的集合

When finding a domain, watch for restrictions: denominators cannot be zero, and expressions under a square root must be non-negative.

在求定义域时，需注意限制条件：分母不能为零，根号内的表达式必须非负。

---

3. Function Notation | 函数记号

Function notation uses the symbol f(x) instead of y. This allows us to clearly express the rule and to evaluate the function at specific points. For instance, if f(x) = 2x + 3, then f(4) = 2(4) + 3 = 11.

函数记号用 f(x) 代替 y。这样可以清楚地表达运算规则，并在特定的点上求函数值。例如，若 f(x) = 2x + 3，则 f(4) = 2(4) + 3 = 11。

f(x) = 2x + 3, 那么 f(4) = 11

Note that f(x) does not mean f multiplied by x. The parentheses are used to group the input variable and show that f is a rule acting upon x.

请注意，f(x) 并不表示 f 乘以 x。括号用于括住输入变量，表示 f 是作用于 x 的规则。

---

4. The Vertical Line Test | 垂直线检验法

The vertical line test is a visual way to determine whether a curve in the coordinate plane is a function. If any vertical line intersects the graph at more than one point, then the graph is not a function.

垂直线检验法是一种视觉方法，用来判断坐标平面上的曲线是否为函数。如果任意一条垂直线与图像交于多于一个点，则该图像不是函数。

垂直线与图像至多相交一次 ⇒ 是函数；多于一次 ⇒ 不是函数

For example, a circle fails the vertical line test because a vertical line can cut it twice, meaning one x-value corresponds to two y-values. A straight line with non-zero slope, by contrast, passes the test easily.

例如，圆无法通过垂直线检验，因为垂直线最多可以与它相交两个点，意味着一个 x 值对应两个 y 值。相比之下，一条斜率非零的直线则能轻松通过检验。

---

5. Linear Functions | 线性函数

A linear function has the general form y = mx + c, where m is the slope and c is the y-intercept. The graph of a linear function is always a straight line.

线性函数的一般形式为 y = mx + c，其中 m 为斜率，c 为 y 轴截距。线性函数的图像总是一条直线。

y = mx + c, 斜率 m = Δy/Δx = (y₂ − y₁)/(x₂ − x₁)

• Slope m / 斜率 m：Measures steepness and direction / 度量陡峭程度和方向
• Intercept c / 截距 c：The point where the line crosses the y-axis / 直线与 y 轴的交点
• Positive slope / 正斜率：Line rises from left to right / 直线从左到右上升
• Negative slope / 负斜率：Line falls from left to right / 直线从左到右下降

A horizontal line has slope 0 and represents a constant function y = c. A vertical line, however, is not a function because it fails the vertical line test.

水平线的斜率为 0，表示常函数 y = c。然而，垂直线不是函数，因为它无法通过垂直线检验。

---

6. Quadratic Functions | 二次函数

A quadratic function is given by y = ax² + bx + c, where a ≠ 0. Its graph is a parabola that opens upward if a > 0 and downward if a < 0.

二次函数形式为 y = ax² + bx + c，其中 a ≠ 0。其图像是抛物线，当 a > 0 时开口向上，当 a < 0 时开口向下。

顶点 x 坐标 = −b/2a, 对称轴为 x = −b/2a

The vertex is the turning point of the parabola, and the axis of symmetry is the vertical line that passes through the vertex. The y-intercept is simply c.

抛物线的顶点是其转折点，对称轴是通过顶点的垂直线。y 轴截距就是 c。

• a > 0 / a 大于 0：Parabola opens upward, vertex is a minimum / 开口向上，顶点为最小值
• a < 0 / a 小于 0：Parabola opens downward, vertex is a maximum / 开口向下，顶点为最大值

---

7. Absolute Value Functions | 绝对值函数

The absolute value function is written as y = |x|. It returns the distance of x from zero, so the output is always non-negative.

绝对值函数写作 y = |x|。它返回 x 到零的距离，因此输出总是非负的。

|x| = x 当 x ≥ 0; |x| = −x 当 x < 0

The graph of y = |x| has a characteristic V shape. It decreases with slope −1 for x < 0, reaches a sharp vertex at the origin (0,0), and then increases with slope +1 for x > 0.

y = |x| 的图像呈典型的 V 形。当 x < 0 时，它以斜率 −1 下降，在原点 (0,0) 处出现尖锐顶点，然后当 x > 0 时以斜率 +1 上升。

---

8. Reading Graphs: Key Points | 读图：关键点

Understanding a function’s graph means being able to identify intercepts, turning points, intervals of increase and decrease, and overall shape. The x-intercepts occur where y = 0, and the y-intercept occurs where x = 0.

理解函数图像意味着能够识别截距、转折点、增减区间以及整体形状。x 轴截距出现在 y = 0 处，而 y 轴截距出现在 x = 0 处。

Feature / 特征	How to Find / 如何找到
x-intercept(s) / x 截距	Set y = 0 and solve for x / 令 y = 0，解出 x
y-intercept / y 截距	Set x = 0 and solve for y / 令 x = 0，解出 y
Turning point / 顶点	Find where the graph changes direction / 找出图像改变方向的位置
Increasing / 递增区间	y-values rise as x increases / x 增大时 y 值上升
Decreasing / 递减区间	y-values fall as x increases / x 增大时 y 值下降

Always label axes with x and y, and mark key coordinates clearly. This makes interpretation and verification much easier.

务必用 x 和 y 标记坐标轴，并清晰标出关键坐标。这样能让解读和验证变得更加容易。

---

9. Transformations of Graphs | 图像的平移变换

Graphs of functions can be shifted, stretched, or reflected. The most common transformation is the vertical or horizontal shift.

函数图像可以平移、伸缩或翻转。最常见的变换是垂直平移和水平平移。

y = f(x) + k 向上平移 k 个单位; y = f(x − h) 向右平移 h 个单位

• y = f(x) + k / 加 k：Shifts graph up if k > 0, down if k < 0 / k > 0 时上移，k < 0 时下移
• y = f(x − h) / 减 h：Shifts graph right if h > 0, left if h < 0 / h > 0 时右移，h < 0 时左移
• y = −f(x) / 取负：Reflects graph across the x-axis / 关于 x 轴翻转
• y = f(−x) / 变负 x：Reflects graph across the y-axis / 关于 y 轴翻转

These transformations are easy to remember if you focus on the input (inside the parentheses) versus the output (outside the parentheses).

只要关注输入（括号内）与输出（括号外）的区别，这些变换就很容易记忆。

---

10. Inverse Functions: A Glimpse | 反函数简介

An inverse function reverses the operation of the original function. If f maps a to b, then the inverse f⁻¹ maps b back to a. Not every function has an inverse; the function must be one-to-one.

反函数是原函数运算的逆过程。如果 f 将 a 映射到 b，那么反函数 f⁻¹ 将 b 映射回 a。并非所有函数都有反函数，函数必须是一一对应的。

f⁻¹(f(x)) = x, 且 f(f⁻¹(x)) = x

Graphically, the inverse function is the reflection of the original graph across the line y = x. This means that the roles of x and y are swapped.

在图像上，反函数是原函数图像关于直线 y = x 的镜像。这意味着 x 和 y 的角色互换。

---

11. Common Mistakes to Avoid | 常见错误警示

Students often confuse the concept of a relation with a function, or misidentify the domain. Here are the most frequent errors and how to avoid them.

学生常常混淆关系与函数的概念，或者错误地判断定义域。以下是最常见的错误及避免方法。

• Mistake 1 / 错误一：Treating all curves as functions — always apply the vertical line test / 把所有曲线都当作函数——务必使用垂直线检验
• Mistake 2 / 错误二：Forgetting that denominators cannot be zero when finding the domain / 求定义域时忘记分母不能为零
• Mistake 3 / 错误三：Confusing f(x) with multiplication / 将 f(x) 误解为乘法
• Mistake 4 / 错误四：Swapping x and y incorrectly when finding an inverse / 求反函数时错误地互换 x 和 y
• Mistake 5 / 错误五：Drawing graphs without labelling axes or key points / 画图时不标记坐标轴和关键点

Paying careful attention to these details will improve both accuracy and exam performance.

对这些细节保持高度关注，将同时提升准确率和考试成绩。

---

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

📚 High-Performance Computing and Parallel Processing | 高性能计算与并行处理

High-performance computing (HPC) refers to the use of supercomputers and parallel processing techniques to solve complex computational problems that exceed the capacity of a single standard computer. At its core lies parallel processing — the simultaneous execution of multiple computations using multiple processors, cores, or machines.

高性能计算（HPC）指利用超级计算机和并行处理技术来解决超出单台标准计算机处理能力的复杂计算问题。其核心是并行处理——即使用多个处理器、多核心或多台机器同时执行多个计算任务。

This revision guide breaks down the key concepts, architectures, programming models, and performance laws that appear in computer science examinations. Whether you are preparing for A-level, IB, or university entrance tests, mastering these fundamentals will give you a clear edge.

本复习指南系统梳理了计算机考试中涉及的关键概念、体系结构、编程模型与性能定律。无论你正在备考 A-level、IB 还是大学入学考试，扎实掌握这些基础内容都将为你带来显著优势。

---

1. Why Parallel Processing? | 为什么要并行处理？

For decades, computer performance improved through increasing clock frequencies and transistor density — a trend described by Moore’s law. However, this approach has hit fundamental physical limits. Higher clock speeds generate excessive heat and consume more power, a problem known as the “power wall.” Meanwhile, the “memory wall” refers to the growing gap between CPU speed and memory access speed.

几十年来，计算机性能的提升主要依靠提高时钟频率和晶体管密度——这一趋势由摩尔定律所描述。然而，这一路线已经触及基本的物理极限。更高的时钟频率会产生过多热量并消耗更多功耗，这一问题被称为”功耗墙”。而”内存墙”则指 CPU 速度与内存访问速度之间不断扩大的差距。

Parallel processing offers a way forward: instead of making a single processor faster, we use many processors working together. This approach is more energy-efficient and can scale to massive problem sizes. Modern computers — from smartphones to data centres — are all parallel to some degree.

并行处理提供了一条出路：与其让单个处理器变得更快，不如让多个处理器协同工作。这种方法更节能，并且能够扩展到超大规模的问题。现代计算机——从智能手机到数据中心——都在一定程度上采用了并行处理。

---

2. Levels of Parallelism | 并行的层次

Parallelism can be exploited at multiple levels within a computer system. Understanding these levels is essential for exam questions that ask you to identify where parallelism occurs.

并行性可以在计算机系统的多个层次上加以利用。理解这些层次对于考试中要求你判断并行发生在何处的题目至关重要。

• Bit-level parallelism: Increasing the word size reduces the number of instructions needed to process large data. For example, a 64-bit processor can process 64 bits of data in one operation, whereas a 16-bit processor would need four operations.
• Instruction-level parallelism (ILP): A single processor executes multiple instructions simultaneously through techniques such as pipelining, superscalar execution, and out-of-order execution.
• Data-level parallelism (DLP): The same operation is applied to many data elements simultaneously. Vector processors and GPUs excel at this type of parallelism.
• Task-level parallelism (TLP): Different tasks or threads run on different processors or cores, each performing independent work that contributes to the overall result.

• 位级并行：增大字长可以减少处理大数据所需的指令数量。例如，64 位处理器一次操作可以处理 64 位数据，而 16 位处理器则需要执行 4 次操作。
• 指令级并行（ILP）：单个处理器通过流水线、超标量执行和乱序执行等技术同时执行多条指令。
• 数据级并行（DLP）：同一条操作同时应用于多个数据元素。向量处理器和 GPU 尤其擅长此类并行。
• 任务级并行（TLP）：不同的任务或线程运行在不同的处理器或核心上，各自执行独立的工作并共同贡献于最终结果。

In modern systems, all four levels coexist. A CPU core uses ILP internally, while multiple cores in a chip exploit TLP, and SIMD (Single Instruction, Multiple Data) units within each core exploit DLP.

在现代系统中，这四种层次同时并存。CPU 核心内部利用指令级并行，芯片中的多个核心利用任务级并行，而每个核心内部的 SIMD（单指令多数据）单元则利用数据级并行。

---

3. Flynn’s Taxonomy | Flynn 分类法

Michael Flynn proposed a classification of parallel computer architectures in 1966 based on two dimensions: the number of instruction streams and the number of data streams. This taxonomy remains a cornerstone of computer architecture exams.

1966 年，迈克尔·弗林提出了基于两个维度——指令流数量和数据流数量——的并行计算机体系结构分类法。这一分类法至今仍是计算机体系结构考试的核心内容。

Classification	Instruction Streams	Data Streams	Example
SISD	One	One	Traditional single-core CPU
SIMD	One	Multiple	GPU, vector processors
MISD	Multiple	One	Fault-tolerant systems (rare)
MIMD	Multiple	Multiple	Multi-core CPUs, clusters

SISD (Single Instruction, Single Data) is the classic von Neumann architecture. SIMD executes one instruction on many data elements simultaneously — ideal for graphics and scientific computing. MISD (Multiple Instruction, Single Data) is rarely implemented in practice but appears in some fault-tolerant and pipeline designs. MIMD is the most common form of parallel processing today, found in multi-core processors, distributed systems, and cloud computing clusters.

SISD（单指令单数据）是经典的冯·诺依曼体系结构。SIMD 在多个数据元素上同时执行一条指令——非常适合图形处理和科学计算。MISD（多指令单数据）在实践中很少实现，但出现在某些容错和流水线设计中。MIMD 是当今最常见的并行处理形式，广泛应用于多核处理器、分布式系统和云计算集群中。

When answering exam questions, remember this key point: GPUs are fundamentally SIMD machines, while modern multi-core CPUs are MIMD machines that also contain SIMD units.

回答考试题目时务必记住关键一点：GPU 本质上是 SIMD 机器，而现代多核 CPU 是同时包含 SIMD 单元的 MIMD 机器。

---

4. Shared-Memory vs Distributed-Memory | 共享内存 vs 分布式内存

Parallel systems are also classified by how memory is organised. This distinction determines which programming model can be used and influences scalability and cost.

并行系统还根据内存的组织方式进行分类。这一区别决定了可以使用哪种编程模型，并影响系统的扩展性和成本。

Shared-memory systems feature multiple processors accessing a single, globally accessible memory space via a bus or interconnect. All processors can read and write the same variables, which simplifies programming. However, they suffer from contention and cache-coherence problems, and are limited in scalability. Multi-core CPUs are a classic example.

共享内存系统的特点是多个处理器通过总线或互连网络访问一个全局可访问的统一内存空间。所有处理器都可以读写相同的变量，这简化了编程。然而，它们面临争用和缓存一致性问题，并且可扩展性受限。多核 CPU 是典型例子。

Distributed-memory systems connect independent nodes via a network, and each node has its own private memory. Processors communicate explicitly by passing messages. This approach scales to thousands or millions of cores, but programming is more complex. Supercomputers and clusters use this model.

分布式内存系统通过网络连接多个独立节点，每个节点拥有自己的私有内存。处理器之间通过显式传递消息进行通信。这种方法可以扩展到数千甚至数百万个核心，但编程更加复杂。超级计算机和集群采用这种模型。

A hybrid design — clusters of shared-memory nodes — is increasingly common. Each node internally uses shared memory, while nodes communicate via message passing.

混合设计——由共享内存节点组成的集群——正变得越来越普遍。每个节点内部使用共享内存，而节点之间通过消息传递进行通信。

---

5. Parallel Programming Models: OpenMP and MPI | 并行编程模型：OpenMP 与 MPI

Two programming models dominate parallel computing: OpenMP for shared-memory systems and MPI (Message Passing Interface) for distributed-memory systems. Exam questions frequently ask you to compare them.

两种编程模型主导着并行计算领域：OpenMP 用于共享内存系统，MPI（消息传递接口）用于分布式内存系统。考试题目经常要求你比较这两种模型。

OpenMP is a set of compiler directives, libraries, and environment variables that extend C, C++, or Fortran. The programmer annotates code regions that should run in parallel, and the runtime system manages threads automatically. For example:

OpenMP 是一组扩展 C、C++ 或 Fortran 的编译器指令、库和环境变量。程序员标注出应并行运行的代码区域，运行时系统自动管理线程。例如：

#pragma omp parallel for
for (i = 0; i < N; i++) { … }

This directive tells the compiler to distribute loop iterations across multiple threads. OpenMP uses a fork-join model: the master thread forks a team of threads, they execute the parallel region, then join back together.

这条指令告诉编译器将循环迭代分配到多个线程上。OpenMP 采用 fork-join 模型：主线程派生出一个线程组，各线程执行并行区域，然后再汇合到一起。

MPI is a message-passing library standard. Processes are independent and communicate by calling functions such as MPI_Send and MPI_Recv. Common collective operations include MPI_Bcast (broadcast) and MPI_Reduce. MPI programs use the Single Program, Multiple Data (SPMD) pattern, where every process runs the same program but on different data.

MPI 是一个消息传递库标准。各进程相互独立，通过调用 MPI_Send 和 MPI_Recv 等函数进行通信。常见的集合操作包括 MPI_Bcast（广播）和 MPI_Reduce（归约）。MPI 程序使用单程序多数据（SPMD）模式，即每个进程运行相同的程序但处理不同的数据。

Feature	OpenMP	MPI
Memory model	Shared memory	Distributed memory
Communication	Implicit (shared variables)	Explicit (message passing)
Ease of use	Easier, incremental	Harder, more control
Scalability	Limited to one machine	Scales across thousands of nodes

In many HPC applications, OpenMP and MPI are combined: OpenMP handles intra-node parallelism while MPI handles inter-node communication.

在许多高性能计算应用中，OpenMP 和 MPI 会结合使用：OpenMP 负责节点内的并行，MPI 负责节点间的通信。

---

6. GPUs and Vector Processing | GPU 与向量处理

Graphics Processing Units (GPUs) have become the workhorse of high-performance computing. A GPU contains thousands of small, simple cores designed for massive SIMD parallelism. They are particularly effective for matrix operations, deep learning, and scientific simulations.

图形处理单元（GPU）已成为高性能计算的主力军。一个 GPU 包含数千个为大规模 SIMD 并行而设计的小型简单核心。它们在矩阵运算、深度学习和科学模拟中尤其高效。

The GPU programming model uses a host (CPU) and a device (GPU). The CPU transfers data to GPU memory, launches a kernel — a function executed in parallel by many threads — and then copies results back. CUDA (NVIDIA) and OpenCL are the standard programming frameworks.

GPU 编程模型采用主机（CPU）加设备（GPU）的结构。CPU 将数据传输到 GPU 内存，启动一个内核——由大量线程并行执行的函数——然后将结果拷贝回来。CUDA（NVIDIA）和 OpenCL 是标准的编程框架。

Vector processing is a related technique in which a single instruction operates on a vector (an array) of data. A vector processor contains deep pipelines for each arithmetic unit. The Cray-1 (1976) was the first successful vector supercomputer. Modern CPUs include vector extensions such as SSE and AVX.

向量处理是一项相关技术，指单个指令对数据向量（数组）进行操作。向量处理器为每个运算单元配置了深度流水线。Cray-1（1976 年）是第一台成功的向量超级计算机。现代 CPU 包含 SSE 和 AVX 等向量扩展。

Key GPU-specific terms you must know: threads (individual execution units), warps (groups of 32 threads executing in lockstep), and memory hierarchy (global, shared, and local memory).

你必须掌握的关键 GPU 术语：线程（单个执行单元）、线程束（以锁步模式执行的 32 个线程组）和内存层次（全局内存、共享内存和局部内存）。

---

7. Amdahl’s Law and Speed-up | Amdahl 定律与加速比

Amdahl’s law is the single most important formula in parallel computing. It quantifies the maximum speed-up obtainable when only part of a program can be parallelised.

Amdahl 定律是并行计算中最重要的公式。它量化了当程序中只有一部分可以被并行化时所能获得的最大加速比。

Let P be the fraction of execution time that can be parallelised, and N the number of processors. The speed-up S is given by:

设 P 为可以被并行化的执行时间占比，N 为处理器数量。加速比 S 由下式给出：

S(N) = 1 / ((1 − P) + P/N)

As N approaches infinity, the maximum speed-up converges to:

当 N 趋于无穷大时，最大加速比收敛于：

S(∞) = 1 / (1 − P)

This means that if 90% of a program is parallelisable (P = 0.9), the absolute maximum speed-up is 10×, no matter how many processors you add. The serial portion (1 − P) becomes the bottleneck. This is a classic exam point: doubling processors does not double performance.

这意味着，如果一个程序中 90% 可以被并行化（P = 0.9），那么无论增加多少处理器，绝对最大加速比仅为 10 倍。串行部分（1 − P）成为瓶颈。这是一个经典考点：处理器数量翻倍并不会使性能翻倍。

Efficiency is defined as E = S(N) / N, measuring how well processors are utilised. Perfect linear speed-up gives E = 1 (100%). In practice, efficiency decreases as N grows due to communication overhead and load imbalance.

效率定义为 E = S(N) / N，用于衡量处理器的利用程度。完美的线性加速比对应 E = 1（100%）。实际上，随着 N 增大，通信开销和负载不均会导致效率下降。

---

8. Designing Parallel Algorithms | 并行算法设计

Designing a parallel algorithm requires a systematic approach. The widely taught framework — PCAM — consists of four stages: Partitioning, Communication, Agglomeration, and Mapping.

设计并行算法需要系统化的方法。广泛使用的框架——PCAM——包含四个阶段：划分、通信、聚合和映射。

Partitioning breaks the problem into smaller tasks and data pieces. Two strategies exist: domain decomposition splits the data, while functional decomposition splits the computation into different tasks.

划分将问题分解为更小的任务和数据块。有两种策略：域分解分割数据，而功能分解将计算拆分为不同的任务。

Communication determines how tasks exchange data. The Goal is to minimise communication, as it is typically far slower than computation. In a stencil computation, for example, each task computes interior values independently but must receive boundary values from neighbouring tasks.

通信确定任务之间如何交换数据。目标是尽量减少通信，因为通信通常远慢于计算。例如，在模板计算中，每个任务独立计算内部值，但必须从相邻任务接收边界值。

Agglomeration combines small tasks into larger ones to reduce communication overhead and improve efficiency. Finally, Mapping assigns tasks to processors, balancing load and minimising inter-processor communication.

聚合将小任务合并为较大的任务，以减少通信开销并提高效率。最后，映射将任务分配到处理器上，实现负载均衡并最小化处理器间通信。

Load balancing is critical: if one processor finishes much earlier than others, it sits idle while the system waits — a waste of resources. Static load balancing assigns fixed distributions, while dynamic load balancing redistributes work at runtime.

负载均衡至关重要：如果一个处理器远早于其他处理器完成工作，它就会空转等待系统，造成资源浪费。静态负载均衡分配固定的任务分布，而动态负载均衡在运行时重新分配工作。

---

9. Measuring and Improving Performance | 性能测量与优化

Besides speed-up and efficiency, several other metrics help evaluate parallel systems. Scalability measures how well performance improves as both problem size and processor count increase. Gustafson’s law observes that in practice, when more processors are available, scientists tend to solve larger problems rather than the same problem faster.

除了加速比和效率之外，还有其他几个指标可以帮助评估并行系统。可扩展性衡量当问题规模和处理器数量同时增加时性能的提升程度。Gustafson 定律指出，在实际应用中，当更多处理器可用时，科学家倾向于解决更大的问题，而不是更快地解决同一个问题。

Gustafson’s law expresses the scaled speed-up as:

Gustafson 定律将规模化加速比表示为：

S(N) = N − P × (N − 1)

Common performance bottlenecks in parallel programs include communication overhead, idle time from load imbalance, and serial sections that cannot be parallelised. Optimisation techniques include overlapping computation with communication, using non-blocking communication, and reducing synchronisation points.

并行程序中的常见性能瓶颈包括通信开销、负载不均导致的空闲时间以及无法并行化的串行部分。优化技术包括将计算与通信重叠、使用非阻塞通信以及减少同步点。

Profiling tools such as Intel VTune, NVIDIA Nsight, and the open-source tool gprof help identify hot spots — the code regions that consume the most execution time. Remember the optimisation principle: first measure, then optimise.

Intel VTune、NVIDIA Nsight 以及开源工具 gprof 等性能分析工具可以帮助识别热点——即消耗最多执行时间的代码区域。记住优化原则：先测量，再优化。

---

10. Applications, Challenges, and the Future | 应用、挑战与未来

High-performance computing powers many fields: weather forecasting, molecular dynamics, computational fluid dynamics, quantum chemistry, genome sequencing, financial risk modelling, and deep learning. In each case, the underlying calculations are inherently parallel — thousands of grid points, atoms, or data samples can be processed simultaneously.

高性能计算推动着众多领域的发展：天气预报、分子动力学、计算流体力学、量子化学、基因组测序、金融风险建模和深度学习。在每种情况下，底层计算本质上都是并行的——数千个网格点、原子或数据样本可以同时被处理。

Despite its successes, parallel processing faces persistent challenges. Energy consumption is a major concern for exascale systems. Programming complexity remains high, and fault tolerance becomes harder as systems grow — when one node in a million fails, the entire job may crash. New research explores domain-specific architectures, quantum computing, and neuromorphic chips as alternatives.

尽管取得了巨大成功，并行处理仍面临持续不断的挑战。能耗是百亿亿次（exascale）系统面临的主要问题。编程复杂性依然很高，而随着系统规模扩大，容错性变得更加困难——当百万节点中有一个节点发生故障，整个作业可能崩溃。新的研究正在探索领域专用架构、量子计算和神经形态芯片作为替代方案。

For your exam, remember the hierarchy: transistors → cores → nodes → clusters. And keep in mind that the trend is toward heterogeneous computing — combining CPUs, GPUs, and specialised accelerators to achieve the best performance per watt.

考试时请记住这个层级：晶体管 → 核心 → 节点 → 集群。同时牢记，当前趋势是异构计算——将 CPU、GPU 和专用加速器结合起来，以实现最佳的性能功耗比。

---

11. Key Exam Points Summary | 考试要点总结

To help you revise efficiently, here is a concise checklist of the most frequently tested concepts in this topic:

为了帮助你高效复习，以下是最常考概念的精简清单：

• Flynn’s taxonomy: SISD, SIMD, MISD, MIMD — be able to give examples of each.
• Shared memory vs distributed memory: advantages and limitations of each.
• OpenMP (thread-based, #pragma) vs MPI (process-based, message passing).
• Amdahl’s law: compute maximum speed-up given P; explain why serial sections limit scaling.
• Differences between GPUs and CPUs: thousands of cores, SIMD, memory hierarchy.
• PCAM design methodology and load balancing.

• Flynn 分类法：SISD、SIMD、MISD、MIMD——能够举出每种类型的例子。
• 共享内存 vs 分布式内存：各自的优点和局限。
• OpenMP（基于线程，使用 #pragma）vs MPI（基于进程，使用消息传递）。
• Amdahl 定律：给定 P 值计算最大加速比；解释为什么串行部分限制扩展。
• GPU 与 CPU 的区别：数千核心、SIMD、内存层次。
• PCAM 设计方法论与负载均衡。

Practice this calculation type: if P = 0.75 on 8 processors, Amdahl’s law gives S = 1 / (0.25 + 0.75/8) = 2.91. Notice that efficiency E = 2.91 / 8 ≈ 0.36 — only 36% utilisation, illustrating the cost of the serial portion.

请练习这类计算：若 P = 0.75，处理器数量为 8，Amdahl 定律给出 S = 1 / (0.25 + 0.75/8) = 2.91。注意效率 E = 2.91 / 8 ≈ 0.36——仅 36% 的利用率，这直观地说明了串行部分带来的代价。

---

High-performance computing and parallel processing are not just exam topics — they are the foundation of modern computing infrastructure. By understanding why parallelism matters, how systems are classified, how programs are written, and how performance is measured, you will be well prepared for both examinations and practical problem-solving.

高性能计算与并行处理不仅是考试主题，更是现代计算基础设施的基石。通过理解并行为何重要、系统如何分类、程序如何编写以及性能如何衡量，你将为考试和实际解决问题做好充分准备。

Published by TutorHao | Computer Science Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

📚 Newton’s Laws of Motion and Force Analysis | 牛顿运动定律与受力分析

Newton’s laws of motion form the bedrock of classical mechanics and are the single most heavily tested topic in A-Level Physics. Every mechanics question, from projectile motion to circular motion, ultimately reduces to a correct application of these laws and a rigorous free-body analysis.

牛顿运动定律是经典力学的基石，也是A-Level物理考试中考查密度最高的考点。无论是抛体运动还是圆周运动，一切力学问题的本质都是对牛顿定律的正确应用，以及对物体进行严谨的受力分析。

---

1. Newton’s First Law — Inertia | 牛顿第一定律——惯性

Newton’s first law states that an object will remain at rest or continue to move with constant velocity unless acted upon by a resultant (net) external force. This is the law of inertia. In mathematical terms, if the resultant force ΣF = 0, then the acceleration a = 0.

牛顿第一定律指出：若物体不受合外力作用，则静止的物体保持静止，运动的物体保持匀速直线运动。这就是惯性定律。用数学表达：若合外力 ΣF = 0，则加速度 a = 0。

The key exam implication is that constant velocity does not mean zero force — it means zero resultant force. A car cruising at 30 m/s on a straight road still has an engine force balancing air resistance and friction exactly.

考试中最重要的推论是：匀速运动不代表不受力，而是合外力为零。例如汽车以 30 m/s 匀速行驶时，发动机的驱动力恰好与空气阻力和摩擦力平衡。

• Inertia is a measure of an object’s resistance to change in its state of motion, quantified by inertial mass.
• 惯性是物体抵抗运动状态改变的能力，其大小由惯性质量来量度。

---

2. Newton’s Second Law — F = ma | 牛顿第二定律——F = ma

Newton’s second law quantifies the relationship between force and motion: the resultant force acting on an object equals the rate of change of its momentum, which for constant mass reduces to F = ma. The acceleration is directly proportional to the resultant force and inversely proportional to the mass.

牛顿第二定律定量描述了力与运动的关系：物体所受的合外力等于其动量的变化率。在质量恒定的情形下，该定律简化为 F = ma。加速度大小与合外力成正比，与物体质量成反比。

ΣF = ma  or  F = ma

ΣF = ma  即  F = ma

The unit of force is the newton (N), defined as the force required to accelerate a 1 kg mass at 1 m s⁻². Crucially, F = ma uses the resultant force — the vector sum of all forces acting on the object — not any individual applied force.

力的单位是牛顿（N），定义为使 1 kg 的物体获得 1 m s⁻² 加速度所需的力。注意，F = ma 中的 F 必须是合外力——即物体所受所有力的矢量和——而不是某个单独的施力。

• Acceleration is always in the same direction as the resultant force.
• 加速度的方向始终与合外力的方向相同。
• When the resultant force is zero, acceleration is zero — uniform motion.
• 当合外力为零时，加速度为零——物体做匀速运动。

---

3. Newton’s Third Law — Action and Reaction | 牛顿第三定律——作用力与反作用力

Newton’s third law states that if object A exerts a force on object B, object B exerts an equal and opposite force on object A. These forces are equal in magnitude, opposite in direction, and act on different objects.

牛顿第三定律指出：若物体 A 对物体 B 施加一个力，则物体 B 也同时对物体 A 施加一个大小相等、方向相反的力。这对力大小相等、方向相反，且作用在不同的物体上。

This is the most misunderstood law at A-Level. Students frequently confuse action-reaction pairs with balanced forces. The critical distinction: balanced forces act on the same object and cancel; action-reaction forces act on different objects and cannot cancel.

这是A-Level中最容易被误解的定律。学生经常把作用力与反作用力对和平衡力混淆。关键区别在于：平衡力作用在同一个物体上，彼此抵消；作用力与反作用力作用在不同物体上，根本不能互相抵消。

Feature	Balanced Forces	Action-Reaction Pair
特征	平衡力	作用力与反作用力
Objects involved	Same object	Two different objects
涉及物体	同一物体	两个不同物体
Effect	Can cancel (net force = 0)	Cannot cancel
效果	可相互抵消（合力为零）	不可抵消

---

4. Free-Body Diagrams — The Foundation | 受力分析图——解题的基础

A free-body diagram (FBD) isolates a single object and represents every external force acting on it as an arrow. The arrow’s direction shows the force’s direction, and its length represents magnitude (proportional scale). All other objects in the surroundings are omitted.

受力分析图（FBD）将某一物体单独隔离出来，用箭头表示作用在该物体上的每一个外力。箭头的方向表示力的方向，箭头的长度（按比例）表示力的大小。周围的其他物体一律不画出来。

Standard forces to identify in an FBD: weight (mg, always downward), normal reaction (perpendicular to the surface), tension (along a string, pulling away from the object), friction (parallel to the surface, opposing motion or attempted motion), and applied forces.

在受力分析图中需要识别的标准力包括：重力（mg，始终竖直向下）、支持力（垂直于接触面）、张力（沿绳子方向，背离物体拉）、摩擦力（平行于接触面，阻碍物体运动或运动趋势），以及外加推力或拉力。

Exam tip: always draw the FBD before writing equations. A correct FBD is worth half the marks in virtually every mechanics question.

考试技巧：先画受力分析图再写方程。在几乎所有的力学题中，一张正确的受力分析图已经拿到了将近一半的分数。

---

5. Weight and the Normal Reaction Force | 重力与支持力

The weight of an object is W = mg, where g is the gravitational field strength, taken as 9.81 m s⁻² on Earth’s surface (often rounded to 10 m s⁻² in approximations). Weight is a force — it is measured in newtons, not kilograms.

物体的重力为 W = mg，其中 g 是重力场强度，在地球表面取 9.81 m s⁻²（近似计算时常取 10 m s⁻²）。重力是一种力，其单位是牛顿（N），而不是千克（kg）。

W = mg

The normal reaction force N acts perpendicular to the contact surface. On a horizontal surface with no vertical acceleration, N = mg. However, on an inclined plane, the normal reaction is not equal to mg — it equals the component of weight perpendicular to the plane.

支持力 N 始终垂直于接触面。在水平面上且竖直方向无加速度时，N = mg。但在斜面上，支持力并不等于 mg，而是等于重力沿垂直斜面方向的分量。

N = mg cos θ (inclined plane)

N = mg cos θ（斜面情形）

In a lift accelerating upward, the normal reaction exceeds mg (apparent weight increases); accelerating downward, it is less than mg. If the lift cable breaks and free-falls, N = 0 — apparent weightlessness.

在加速上升的电梯中，支持力大于 mg（视重增大）；加速下降时，支持力小于 mg。若电梯缆绳断裂而自由下落，则 N = 0——出现完全失重。

---

6. Friction — Static and Kinetic | 摩擦力——静摩擦与滑动摩擦

Friction acts parallel to the contact surface and opposes relative motion or the tendency of motion. Two regimes exist: static friction (before motion starts) and kinetic (sliding) friction (during motion).

摩擦力方向平行于接触面，阻碍物体的相对运动或相对运动趋势。摩擦力分两种：静摩擦力（物体开始运动之前）和滑动摩擦力（物体运动过程中）。

fₛ ≤ μₛN   (static friction)

fₖ = μₖN   (kinetic friction)

fₛ ≤ μₛN   （静摩擦力）

fₖ = μₖN   （滑动摩擦力）

Here μₛ is the coefficient of static friction and μₖ is the coefficient of kinetic friction, with μₛ generally greater than μₖ for the same pair of surfaces. N is the normal reaction force — note that friction is proportional to N, not to the total weight.

其中 μₛ 为静摩擦系数，μₖ 为滑动摩擦系数。对于同一对接触面，μₛ 通常大于 μₖ。N 是支持力——注意摩擦力正比于 N，而不是正比于物体的总重力。

• Static friction is adjustable: it grows from zero up to a maximum value μₛN as the applied force increases.
• 静摩擦力是可变的：它随外力的增加从零逐渐增大到最大值 μₛN。
• Once motion begins, kinetic friction takes over at a constant value μₖN.
• 一旦物体开始滑动，滑动摩擦力就以恒定值 μₖN 取而代之。

---

7. Inclined Planes — Resolving Forces | 斜面问题——力的分解

The inclined plane is a classic A-Level scenario. Resolve weight into two components: mg sin θ parallel to the plane (down the slope) and mg cos θ perpendicular to the plane.

斜面是A-Level力学中的经典情景。将重力分解为两个分量：沿斜面方向的分量 mg sin θ（沿斜面向下）和垂直于斜面方向的分量 mg cos θ。

Parallel: mg sin θ    Perpendicular: mg cos θ

平行于斜面：mg sin θ    垂直于斜面：mg cos θ

For a block on a rough plane at angle θ to the horizontal, with motion down the plane:

对于放置在倾角为 θ 的粗糙斜面上的物体，若物体沿斜面下滑：

ma = mg sin θ − μₖN = mg sin θ − μₖmg cos θ

ma = mg sin θ − μₖN = mg sin θ − μₖmg cos θ

a = g(sin θ − μₖ cos θ)

When the block is just about to slip, static friction reaches its maximum: μₛ = tan θ. This gives a powerful experimental method for measuring μₛ — tilt the plane until motion just begins.

当物体刚好处于将要滑动的临界状态时，静摩擦力达到最大值，此时 μₛ = tan θ。这提供了一种极为实用的测量 μₛ 的实验方法——不断增大斜面倾角，直到物体恰好开始滑动。

---

8. Tension and Connected Particles | 张力与连接体问题

When two masses are connected by a light inextensible string, the tension is the same at both ends of the string. “Light” means negligible mass (so T is uniform along the string); “inextensible” means both masses share the same acceleration magnitude.

当两个物体通过轻质不可伸长的绳子连接时，绳子两端的张力大小相等。”轻质”意味着绳子的质量可以忽略不计（因此张力沿绳子处处相等）；”不可伸长”意味着两个物体的加速度大小相同。

Consider two masses m₁ and m₂ connected by a string over a frictionless pulley (Atwood machine). The heavier mass m₂ accelerates downward, m₁ accelerates upward, both with the same magnitude a.

考虑两个质量分别为 m₁ 和 m₂ 的物体，通过一根绳子跨过无摩擦的定滑轮相连（阿特伍德机）。较重的 m₂ 向下加速，较轻的 m₁ 向上加速，二者的加速度大小相等。

For m₂:   m₂g − T = m₂a

For m₁:   T − m₁g = m₁a

对 m₂：  m₂g − T = m₂a

对 m₁：  T − m₁g = m₁a

Adding both equations eliminates T:

将两式相加消去 T：

a = (m₂ − m₁)g / (m₂ + m₁)

When writing equations for connected particles, assign a consistent positive direction for the entire system. This is the single most reliable strategy for solving multi-body problems.

在列连接体问题的方程时，为整个系统规定一致的正方向。这是解决多物体问题最可靠的策略。

---

9. Resolving Forces at Angles — Composites | 斜向力的分解——复合情形

When a force F is applied at an angle θ to the horizontal, decompose it:

当一个力 F 与水平方向成 θ 角时，将其分解为：

Fₓ = F cos θ (horizontal)    Fᵧ = F sin θ (vertical)

Fₓ = F cos θ（水平分量）   Fᵧ = F sin θ（竖直分量）

A classic exam question: a box of mass m is pulled along a rough horizontal floor by a force F at angle θ above the horizontal. Here the normal reaction is reduced because the vertical component of F helps support the box.

经典考题：质量为 m 的箱子在水平粗糙地面上被一个与水平方向成 θ 角向上的力 F 拉着前进。此时支持力减小了，因为 F 的竖直分量帮助分担了箱子的部分重力。

N = mg − F sin θ

N = mg − F sin θ

ma = F cos θ − μₖN

If instead the force is pushing downward at an angle, N = mg + F sin θ and friction increases. Always state the normal reaction correctly before calculating friction.

如果力是斜向下推的，则 N = mg + F sin θ，摩擦力增大。在计算摩擦力之前，务必正确地写出支持力的表达式。

---

10. Applications — Lifts, Landing and Towing | 应用——电梯、着陆与牵引

Lift problems: A person of mass m in a lift experiences an apparent weight equal to the normal reaction N. With acceleration a upward: N = m(g + a). With a downward: N = m(g − a). A candidate who writes these two expressions correctly can usually solve the entire question.

电梯问题：电梯中质量为 m 的人所感受到的视重等于支持力 N。若电梯以加速度 a 上升：N = m(g + a)；若以加速度 a 下降：N = m(g − a)。考生只要正确写出这两个表达式，通常就能解出整道题。

Towing problems: A car tows a trailer with a tow rope of tension T. For the car and trailer as a single system: F − total resistance = (m_car + m_trailer)a. Then for the trailer alone: T − resistance_on_trailer = m_trailer × a. The system approach eliminates internal forces (the tension).

牵引问题：汽车通过拖绳牵引一辆拖车，拖绳上的张力为 T。将汽车和拖车视为一个整体系统：F − 总阻力 = (m_车 + m_拖车)a。然后单独分析拖车：T − 拖车所受阻力 = m_拖车 × a。整体法可以消去内力（即张力）。

---

11. Common Mistakes and Examiner Tips | 常见错误与考官建议

• Forgetting to include all forces in the FBD — always ask: is there gravity, normal contact, tension, friction, applied force?
• 遗漏受力分析图中的力——每次都要问自己：重力、支持力、张力、摩擦力、外加力都画上了吗？

• Using F = ma with an individual force instead of the resultant force. The equation is always ΣF = ma.
• 列 F = ma 时错误地使用某个分力而不是合外力。方程必须始终写成 ΣF = ma。

• Confusing the direction of friction: friction always opposes relative motion between surfaces, not necessarily the direction of motion. A box on an accelerating truck experiences forward friction because its tendency is to slide backward.
• 搞错摩擦力方向：摩擦力总是阻碍物体间的相对运动，而不一定是阻碍物体的运动方向。放置在加速卡车上的箱子受到向前的摩擦力，因为箱子有相对向后滑动的趋势。

• Forgetting to convert units — mass in kg, distances in m, time in s. A speed of 72 km h⁻¹ is 20 m s⁻¹.
• 忘记统一单位——质量用 kg、距离用 m、时间用 s。72 km h⁻¹ 等于 20 m s⁻¹。

• Not specifying a positive direction in multi-particle problems, leading to sign errors.
• 在多物体问题中未规定正方向，导致符号错误。

The examiner’s golden rule: draw the free-body diagram, choose a positive direction, write Newton’s second law for each object or for the whole system, then solve simultaneously. This structured approach converts an intimidating mechanics problem into routine algebra.

考官的黄金法则：画出受力分析图，选定正方向，对每个物体或整个系统分别写出牛顿第二定律，然后联立求解。按这种结构化流程操作，再复杂棘手的力学题也会变成常规的代数运算。

---

12. Exam-Style Worked Example | 考试风格例题精解

Question: A block of mass 4 kg is projected up a rough plane inclined at 30° to the horizontal. The coefficient of kinetic friction is 0.25. Calculate the deceleration of the block while it is moving up the plane.

题目：一个质量为 4 kg 的物体沿倾角为 30° 的粗糙斜面向上运动（初速度沿斜面向上）。已知滑动摩擦系数为 0.25。求物体沿斜面上升过程中的减速度。

Solution: While moving upward, the friction opposes the motion — friction acts down the plane, in the same direction as the component of weight.

解答：物体向上运动过程中，摩擦力阻碍其运动——摩擦力沿斜面向下，与重力沿斜面方向的分量方向相同。

Forces parallel to the plane:

沿斜面方向的力：

ma = mg sin θ + μₖN

N = mg cos θ

a = g(sin θ + μₖ cos θ)

Substituting g = 9.81 m s⁻², θ = 30°, μₖ = 0.25:

代入 g = 9.81 m s⁻²，θ = 30°，μₖ = 0.25：

a = 9.81 × (sin 30° + 0.25 × cos 30°) = 9.81 × (0.5 + 0.2165) ≈ 7.03 m s⁻²

The deceleration is approximately 7.0 m s⁻², directed down the plane.

减速度约为 7.0 m s⁻²，方向沿斜面向下。

Notice that the mass cancels out — if the question had asked for the deceleration, the mass was a distractor. Many A-Level questions deliberately include irrelevant data to test conceptual understanding.

注意到质量被消去了——题目给出 4 kg 的质量，实际上是一个干扰项。许多A-Level考题会故意加入无关数据，以考查学生的概念理解能力。

---

Published by TutorHao | Physics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

📚 Analysis of Basic Program Control Structures | 程序基本控制结构解析

In nearly every A-Level Computer Science syllabus, program control structures form the foundation of algorithmic thinking. Whether you are writing pseudocode for Paper 2 or debugging code in a high-level language, mastering sequence, selection, and iteration is essential for earning full marks. This article unpacks the three fundamental control structures, explains their pseudocode forms, and highlights common examination traps.

在几乎所有 A-Level 计算机科学考纲中，程序控制结构都是算法思维的基础。无论你是在卷二上编写伪代码，还是在高级语言中调试程序，掌握顺序、选择和循环都是拿满分的必要条件。本文深度解析三大基本控制结构，说明其伪代码写法，并指出常见考试陷阱。

---

1. The Three Fundamental Structures | 三种基本结构

The famous theorem by Böhm and Jacopini (1966) proved that any computable problem can be solved using only three control structures: sequence, selection, and iteration. No program, however complex, requires anything beyond these three building blocks. This insight allows examiners to test algorithmic logic in a language-independent way through pseudocode.

Böhm 与 Jacopini 于 1966 年提出的著名定理证明：任何可计算问题都仅需顺序、选择、循环三种控制结构即可求解。无论多么复杂的程序，都不需要超出这三种构件。这一结论使考官能够通过伪代码以不依赖具体语言的方式考查算法逻辑。

The table below summarises the three structures and their characteristics:

下表概述了三种结构及其特征：

Structure | 结构	Key feature | 关键特征	Pseudocode keyword | 伪代码关键字
Sequence | 顺序	Executes in order, top to bottom | 自顶向下依次执行	None (implicit) | 无（隐含）
Selection | 选择	Chooses one path among several | 在多个路径中择一执行	IF, CASE
Iteration | 循环	Repeats a block zero or more times | 重复执行代码块多次	FOR, WHILE, REPEAT

---

2. Sequence Structure | 顺序结构

Sequence is the simplest structure: statements are executed one after another in the exact order they appear. Each statement completes before the next begins. In pseudocode, a sequence is simply written as consecutive lines, and examiners expect you to trace through them in order, updating variable values line by line.

顺序结构是最简单的结构：语句严格按照出现的先后次序逐条执行。每条语句完成之后，下一条语句才开始。在伪代码中，顺序结构就是连续的几行代码；考官期望你按顺序逐行追踪，逐句更新变量值。

Consider the following example:

看下面的例子：

x ← 5
y ← x + 2
x ← y × 3
OUTPUT x

The output is 21, not 15. Many students incorrectly treat x as ‘5’ throughout the program. In a sequence, x becomes 7 after the second line, then 21 after the third. Tracing tables remain the best tool for this kind of question.

输出结果是 21，而不是 15。许多学生错误地在整个程序中把 x 当作始终等于 5。在顺序结构中，执行第二行后 x 变为 7，第三行后变为 21。对这种题目，追踪表仍然是最好的工具。

---

3. Selection Structure — IF Statements | 选择结构 — IF 语句

Selection allows a program to make decisions. The simplest form is IF-THEN, where a condition is evaluated as TRUE or FALSE. A condition is a boolean expression using relational operators such as = , ≠ , < , > , ≤ , ≥ , and can be combined with AND, OR, NOT.

选择结构让程序能够做出判断。最简单的形式是 IF-THEN，条件被求值为 TRUE 或 FALSE。条件是用关系运算符（如 =、≠、<、>、≤、≥）构成的布尔表达式，还可以用 AND、OR、NOT 进行组合。

The classic structure in Cambridge-style pseudocode is:

剑桥风格伪代码的经典结构是：

IF <condition> THEN
<statements>
ELSE
<statements>
ENDIF

The ELSE branch is optional; if a condition is FALSE and there is no ELSE, control simply moves to the statement after ENDIF. Every IF must be terminated by ENDIF in formal pseudocode, otherwise marks are lost for missing delimiters.

ELSE 分支是可选的；若条件为 FALSE 且没有 ELSE，控制流将直接跳到 ENDIF 之后的语句。在正式伪代码中，每个 IF 都必须以 ENDIF 结束，否则会因缺少定界符而丢分。

A nested IF occurs when an IF statement appears inside another IF statement. Nesting allows multi-way decisions. However, deeply nested IFs become hard to read; examiners often expect you to replace them with CASE where appropriate.

当一个 IF 语句出现在另一个 IF 语句内部时，就形成了嵌套 IF。嵌套可以实现多路判断。但深层嵌套的 IF 可读性差，考官通常期望你在适当场合改用 CASE 结构。

---

4. Selection Structure — CASE Statements | 选择结构 — CASE 语句

When a variable must be compared against many distinct values, the CASE statement is cleaner than a long chain of IF-ELSE-IF. In formal pseudocode, CASE evaluates an expression once and matches it against a list of values.

当需要把某个变量与多个不同取值比较时，CASE 语句比一长串 IF-ELSE-IF 更清晰。在正式伪代码中，CASE 只对表达式求值一次，并将其与一组取值逐一匹配。

CASE OF grade
‘A’ : OUTPUT ‘Excellent’
‘B’ : OUTPUT ‘Good’
‘C’ : OUTPUT ‘Average’
OTHERWISE : OUTPUT ‘Try harder’
ENDCASE

OTHERWISE acts as a default branch, similar to ELSE in an IF statement. Note that CASE must be closed by ENDCASE, just as IF requires ENDIF. In some programming languages the equivalent keyword is SWITCH, but in CIE/CAIE and Pearson Edexcel pseudocode, CASE OF … ENDCASE is the accepted standard.

OTHERWISE 充当默认分支，类似 IF 语句中的 ELSE。注意，CASE 必须以 ENDCASE 结束，正如 IF 必须以 ENDIF 结束一样。在有些编程语言中对应关键字是 SWITCH，但在 CIE/CAIE 和 Pearson Edexcel 的伪代码中

Published by TutorHao | Computer Science Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

📚 Topological Structures: Basic Concepts and Applications | 拓扑结构基础概念与应用

Topology is a branch of mathematics that studies properties preserved under continuous deformations, such as stretching and bending, but not tearing or gluing. Unlike geometry, which focuses on precise distances and angles, topology asks deeper questions about the “shape” of space itself: what does it mean for two spaces to be equivalent? How can we describe the structure of a set in a way that survives drastic changes?

拓扑学是数学的一个分支，研究在连续变形下保持不变的性质，例如拉伸和弯曲，但不能撕裂或粘合。与关注精确距离和角度的几何学不同，拓扑学提出关于空间本身“形状”的更深刻问题：两个空间等价意味着什么？我们如何用某种方式描述一个集合的结构，使其在剧烈变化中依然保留？

---

1. What Is a Topological Space? | 什么是拓扑空间？

A topological space is a fundamental object in topology. It consists of a set X together with a collection τ of subsets of X, called open sets, satisfying three axioms: the empty set and X itself are open; the union of any family of open sets is open; and the intersection of any finite family of open sets is open.

拓扑空间是拓扑学中的基本对象。它由一个集合 X 和 X 的子集族 τ 组成，这些子集称为开集，并满足三条公理：空集和 X 本身是开集；任意多个开集的并集是开集；有限多个开集的交集是开集。

Formally, we write (X, τ) to denote the topological space. The collection τ is called a topology on X. The open sets encode the notion of nearness or continuity without relying on a distance function.

形式上，我们用 (X, τ) 表示拓扑空间。族 τ 称为 X 上的拓扑。开集编码了邻近性或连续性的概念，而无需依赖距离函数。

Example: On the real line ℝ, the usual topology is generated by all open intervals (a, b) = {x ∈ ℝ : a < x < b}. This topology captures our intuitive idea of continuity for real-valued functions.

例如：在实数线 ℝ 上，通常拓扑由所有开区间 (a, b) = {x ∈ ℝ : a < x < b} 生成。这个拓扑捕捉了实值函数连续性的直观概念。

---

2. Open Sets and Neighborhoods | 开集与邻域

Open sets are the building blocks of a topological space. A subset U of X is open if U belongs to the topology τ. Intuitively, an open set is a region where every point has some “room” around it within the set. This idea is made precise using neighborhoods.

开集是拓扑空间的基石。X 的子集 U 是开集，如果 U 属于拓扑 τ。直观上，开集是一个区域，其中每个点在该集合内都有一些“空间”。这一概念可通过邻域精确化。

A neighborhood of a point x in a topological space (X, τ) is any open set U such that x ∈ U. In the usual topology on ℝ, an open interval containing x is a neighborhood of x. Neighborhoods allow us to define limits, continuity, and convergence without a metric.

在拓扑空间 (X, τ) 中，点 x 的邻域是任意满足 x ∈ U 的开集 U。在 ℝ 的通常拓扑中，包含 x 的开区间就是 x 的邻域。邻域使我们能够在没有度量的情况下定义极限、连续性和收敛。

x ∈ U ⊆ τ 且 U 是开集 ⇒ U 是 x 的邻域

This local perspective turns global questions about space into local questions about points and their surrounding sets, a powerful technique in all areas of topology.

这种局部视角将关于空间的全局问题转化为关于点及其周围集合的局部问题，这是拓扑学各领域中的强大技术。

---

3. Continuous Functions and Homeomorphisms | 连续函数与同胚

In topology, a function f : X → Y between two topological spaces is continuous if the preimage of every open set in Y is an open set in X. This definition generalizes the epsilon-delta definition from calculus and works in any space where open sets are defined.

在拓扑学中，两个拓扑空间之间的函数 f : X → Y 是连续的，如果 Y 中每个开集的原像是 X 中的开集。这个定义推广了微积分中的 ε-δ 定义，适用于任何定义了开集的空间。

A homeomorphism is a bijective continuous function whose inverse is also continuous. If such a function exists between two spaces, they are said to be homeomorphic, meaning they are topologically indistinguishable. For example, a coffee mug and a donut (torus) are homeomorphic because one can be continuously deformed into the other.

同胚是双射且其逆映射也连续的连续函数。如果两个空间之间存在这样的函数，则称它们同胚，意味着它们在拓扑上不可区分。例如，咖啡杯和甜甜圈（环面）是同胚的，因为一个可以连续变形为另一个。

f : X → Y 是同胚 ⇔ f 连续、双射、且 f⁻¹ 连续

Homeomorphism is the central notion of equivalence in topology. When two spaces are homeomorphic, they share all topological properties, such as connectedness, compactness, and number of “holes”.

同胚是拓扑学中等价的核心概念。当两个空间同胚时，它们共享所有拓扑性质，如连通性、紧致性以及“洞”的数量。

---

4. Connectedness | 连通性

A topological space is connected if it cannot be expressed as the union of two disjoint nonempty open sets. In other words, the space is “all in one piece”. The real line ℝ is connected, while the set ℚ of rational numbers with the subspace topology is not.

拓扑空间是连通的，如果它不能表示为两个不相交非空开集的并集。换句话说，空间是“连成一片”的。实数线 ℝ 是连通的，而带有子空间拓扑的有理数集 ℚ 不是连通的。

Connectedness is a topological property: if X is connected and X is homeomorphic to Y, then Y is also connected. This helps us prove that certain spaces are not homeomorphic. For instance, an open interval (0, 1) is connected, but removing a point from it produces two separate pieces, whereas removing a point from a circle leaves a single interval-like piece.

连通性是一个拓扑性质：如果 X 连通且 X 与 Y 同胚，则 Y 也连通。这有助于我们证明某些空间不同胚。例如，开区间 (0, 1) 是连通的，但从中去掉一个点会得到两个分离的部分；而从圆中去掉一个点则留下一个类似区间的单段部分。

A stronger notion is path connectedness: a space is path connected if any two points can be joined by a continuous path within the space. Every path connected space is connected, but the converse is not always true.

更强的概念是道路连通：一个空间是道路连通的，如果任意两点可以通过空间内的连续道路连接。每个道路连通的空间都是连通的，但反之不一定成立。

---

5. Compactness | 紧致性

Compactness is a topological generalization of finiteness. A space is compact if every open cover (a collection of open sets whose union contains the space) has a finite subcover. In ℝ with the usual topology, the closed interval [0, 1] is compact, while the open interval (0, 1) is not.

紧致性是有限性的拓扑推广。一个空间是紧致的，如果每个开覆盖（一组开集，其并集包含该空间）都有有限子覆盖。在具有通常拓扑的 ℝ 中，闭区间 [0, 1] 是紧致的，而开区间 (0, 1) 不是。

Compactness plays a crucial role in analysis and geometry. Continuous images of compact spaces are compact, and continuous real-valued functions on compact spaces attain their maximum and minimum values. This is the topological reason behind the extreme value theorem.

紧致性在分析和几何中起着关键作用。紧致空间的连续像是紧致的，紧致空间上的连续实值函数必取得最大值和最小值。这是极值定理背后的拓扑原因。

X 紧致 ⇒ 每个开覆盖 {Uᵢ}ᵢ∈I 都有有限子覆盖 {Uᵢ₁, …, Uᵢₙ}

For metric spaces, compactness is equivalent to being closed and bounded (Heine-Borel theorem), but in general topology the definition in terms of covers is more powerful and widely applicable.

对于度量空间，紧致性等价于闭且有界（海涅-博雷尔定理），但在一般拓扑学中，基于覆盖的定义更强大且适用范围更广。

---

6. Separation Axioms | 分离公理

Separation axioms describe how well points and closed sets can be “separated” by open sets in a topological space. The most important axioms are T₁, T₂ (Hausdorff), T₃ (regular), and T₄ (normal). A Hausdorff space (T₂) requires that any two distinct points have disjoint neighborhoods.

分离公理描述拓扑空间中点与闭集可以被开集“分离”的程度。最重要的公理是 T₁、T₂（豪斯多夫）、T₃（正则）和 T₄（正规）。豪斯多夫空间（T₂）要求任意两个不同点都有不相交的邻域。

Almost all spaces encountered in analysis, geometry, and physics are Hausdorff. The Hausdorff property ensures that limits of sequences are unique. In non-Hausdorff spaces, a sequence may converge to more than one point, which is typically undesirable for applications.

分析、几何和物理中遇到的大多数空间都是豪斯多夫的。豪斯多夫性质确保序列的极限是唯一的。在非豪斯多夫空间中，序列可能收敛到多个点，这通常在应用中是不可取的。

Normality (T₄) is a stronger condition: any two disjoint closed sets can be separated by disjoint open sets. This property is essential for constructing continuous functions, as in Urysohn’s lemma, a foundational result in topology.

正规性（T₄）是一个更强的条件：任意两个不相交闭集可以被不相交开集分离。这个性质对于构造连续函数至关重要，例如乌雷松引理，这是拓扑学的一个基础结果。

---

7. Applications in Data Analysis | 在数据分析中的应用

Topological data analysis (TDA) uses tools from topology to study the shape of high-dimensional data. Persistent homology, a key technique, tracks how topological features such as connected components, loops, and voids appear and disappear across different scales of a data cloud.

拓扑数据分析（TDA）使用拓扑学工具研究高维数据的形状。持续同调是一种关键技术，它跟踪数据云在不同尺度下拓扑特征（如连通分量、环和空洞）的出现与消失。

In TDA, a point cloud is often converted into a family of simplicial complexes (e.g., Čech or Vietoris-Rips complexes) by increasing a scale parameter ε. As ε grows, new simplices are added, creating or destroying holes. Persistent homology records the birth and death times of these features.

在 TDA 中，点云通常通过增大尺度参数 ε 转化为一族单纯复形（例如 Čech 或 Vietoris-Rips 复形）。随着 ε 增大，新的单形被添加，从而产生或消除空洞。持续同调记录这些特征的出生和死亡时间。

H₀ = 连通分量，H₁ = 环，H₂ = 空洞，Hₖ = k 维“洞”

Persistence diagrams and barcodes summarize the topological signature of the data. These summaries are insensitive to noise and to continuous deformation, making them valuable in fields like biology, materials science, and computer vision.

持续图和条形码总结了数据的拓扑特征。这些摘要对噪声和连续变形不敏感，使其在生物学、材料科学和计算机视觉等领域具有重要价值。

---

8. Applications in Physics and Network Science | 在物理与网络科学中的应用

Topology appears throughout modern physics. In condensed matter physics, topological insulators are materials that behave as insulators in their interior but conduct electricity on their surface. The protected surface states arise from the nontrivial topology of the electronic band structure.

拓扑学贯穿现代物理学。在凝聚态物理中，拓扑绝缘体是内部表现为绝缘体而表面导电的材料。受保护的表面态源于电子能带结构的非平凡拓扑。

In cosmology, the topology of the universe is studied through cosmic microwave background radiation. Observational data place constraints on whether the universe is finite and what its global shape might be, such as a 3-torus or a more exotic manifold.

在宇宙学中，通过宇宙微波背景辐射研究宇宙的拓扑。观测数据对宇宙是否有有限体积以及其整体形状（如三维环面或更奇特的流形）给出了限制。

In network science, topological measures such as connectivity, cycles, and higher-order structures are used to analyze complex networks. Persistent homology can detect meaningful holes or bottlenecks in social, biological, and transportation networks, revealing hidden organizational principles.

在网络科学中，连通性、环路和高阶结构等拓扑度量被用于分析复杂网络。持续同调可以检测社会、生物和交通网络中有意义的空洞或瓶颈，揭示隐藏的组织原理。

---

9. Summary and Further Study | 总结与进一步学习

Topology provides a flexible and powerful language for studying space and continuity. Starting with the simple axioms of a topological space, one can build deep concepts such as connectedness, compactness, and homeomorphism. These ideas not only unify many branches of pure mathematics but also drive modern applications in data science and physics.

拓扑学为研究空间和连续性提供了一种灵活而强大的语言。从拓扑空间的简单公理出发，可以构建连通性、紧致性和同胚等深刻概念。这些思想不仅统一了纯数学的许多分支，也推动了数据科学和物理学中的现代应用。

For students preparing for examinations, the key is to master the definitions and examples, and to practice proving simple topological properties. Visual intuition is helpful, but rigorous arguments based on open sets are essential for success.

对于备考学生，关键是掌握定义和例子，并练习证明简单的拓扑性质。视觉直觉有帮助，但基于开集的严格论证对取得成功至关重要。

Further study leads to algebraic topology, where groups and homology theory classify spaces, and to differential topology, which studies smooth manifolds. These advanced topics continue to reveal the deep connections between geometry, analysis, and topology.

进一步学习将进入代数拓扑（用群和同调论对空间进行分类）以及微分拓扑（研究光滑流形）。这些高级课题继续揭示几何、分析和拓扑之间的深层联系。

---

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

📚 Python Data Structures and Common Operations | Python数据结构与常用操作

Data structures are the backbone of Python programming. In a computer science exam, you must be able to explain how lists, tuples, strings, dictionaries and sets behave, and which operations are efficient. This article reviews the essential Python data structures, their mutability, common methods and typical time complexities, with paired English and Chinese explanations for revision.

数据结构是 Python 编程的核心。在计算机科学考试中，你必须能够解释列表、元组、字符串、字典和集合的行为，并了解哪些操作是高效的。本文复习 Python 的核心数据结构、它们的可变性、常用方法及典型时间复杂度，以中英对照的方式帮助你备考。

1. Why Data Structures Matter | 为什么数据结构重要

A data structure is an organised way of storing and accessing data. Python provides several built-in types, each with different strengths. The choice of structure can affect memory use, speed and readability.

数据结构是有组织地存储和访问数据的方式。Python 提供了多个内置类型，各自有不同优势。选择合适的数据结构会影响内存使用、运行速度和代码可读性。

In exam questions, you may be asked to choose the best structure for a task. For example, use a list when you need ordered access, a dictionary when you need fast lookup by key, a set when you need to remove duplicates, and a tuple when the data should not change.

在考试题目中，你可能会被要求为某个任务选择最佳数据结构。例如，需要有序访问时用列表；需要按键快速查找时用字典；需要去重时用集合；当数据不应改变时用元组。

---

2. Lists: Dynamic Arrays | 列表：动态数组

Lists are ordered, mutable sequences. They may contain elements of different types, and they allow duplicate values. A list stores references to objects, so the same object may appear more than once.

列表是有序、可变的序列。列表可以包含不同类型的元素，也允许重复值。列表存储对象的引用，因此同一个对象可能出现多次。

Common operations include append, insert, remove, pop, index, count, slicing and sorting. Index access by position is O(1), but inserting or deleting from the middle is O(n).

常见操作包括 append、insert、remove、pop、index、count、切片和排序。按位置访问是 O(1)，但从中间插入或删除是 O(n)。

fruits = ['apple', 'banana', 'cherry']
fruits.append('date')          # add to end
fruits.insert(1, 'blueberry')  # insert at index 1
first = fruits[0]              # O(1) index access
removed = fruits.pop()         # remove and return last item
del fruits[0]                  # remove by index
print(fruits)

Because lists are mutable, you can change an element with my_list[i] = value. This is often tested together with the idea of aliasing and copying, which is covered later.

由于列表是可变的，你可以使用 my_list[i] = value 修改元素。这一点常与别名和拷贝的概念一起考查，后文会详细说明。

---

3. Tuples: Immutable Sequences | 元组：不可变序列

Tuples are ordered sequences similar to lists, but they are immutable. Once created, a tuple cannot be changed. This makes tuples hashable when all of their elements are also immutable, so they can be used as dictionary keys.

元组是与列表类似的有序序列，但元组是不可变的。元组一旦创建，就无法修改。当元组中的所有元素都是不可变类型时，元组是可哈希的，因此可以用作字典的键。

Tuples support indexing, slicing, count, index and unpacking. Unpacking is a convenient way to assign separate variables from a tuple.

元组支持索引、切片、count、index 和解包。解包是一种将元组中元素分别赋值给多个变量的便捷方式。

point = (3, 4)
x, y = point
print(point[1])          # 4
# point[1] = 5          # TypeError: 'tuple' object does not support item assignment

Use tuples for fixed records such as coordinates, RGB colours or return values. In advanced use, collections.namedtuple creates tuple-like objects with named fields, which improves readability.

元组适合表示固定记录，例如坐标、RGB 颜色或函数返回值。在进阶用法中，collections.namedtuple 可以创建带字段名的元组类对象，提高代码可读性。

---

4. Strings: Immutable Character Sequences | 字符串：不可变字符序列

Strings are immutable sequences of characters. They behave like tuples for sequences, but they have many text-processing methods. In Python 3, strings are Unicode by default.

字符串是字符的不可变序列。在序列行为上，字符串类似于元组，但它有大量文本处理方法。在 Python 3 中，字符串默认使用 Unicode 编码。

Key methods include split, join, strip, lower, upper, replace, find, count and startswith. Slicing works exactly as with lists.

关键方法包括 split、join、strip、lower、upper、replace、find、count 和 startswith。切片操作与列表完全相同。

word = "algorithm"
print(word[::-1])                 # mhtirogla
parts = "a,b,c".split(",")        # ['a', 'b', 'c']
new = "  hello  ".strip().upper() # 'HELLO'
print(", ".join(["a", "b", "c"])) # a, b, c

Because strings are immutable, repeated concatenation in a loop can be slow. If you must combine many pieces, build a list and call ''.join(list) instead.

因为字符串是不可变的，在循环中反复拼接字符串可能很慢。如果需要合并大量片段，应先把片段放入列表，然后调用 ''.join(list) 进行拼接。

---

5. Dictionaries: Key-Value Maps | 字典：键值映射

Dictionaries store key-value pairs. Keys must be hashable, such as strings, numbers or tuples of immutable objects. Values may be of any type. In Python 3.7 and later, dictionaries preserve insertion order.

字典存储键值对。键必须是可哈希的，例如字符串、数字或包含不可变对象的元组。值可以是任意类型。在 Python 3.7 及更高版本中，字典会保留插入顺序。

The main power of a dictionary is fast lookup by key. Membership testing and retrieval are O(1) on average. Common methods are get, setdefault, update, pop, keys, values and items.

字典的主要优势是按键快速查找。平均情况下，成员检查和取值都是 O(1)。常用方法有 get、setdefault、update、pop、keys、values 和 items。

ages = {"Ali": 17, "Mia": 18}
ages["Bo"] = 19
age = ages.get("Ali", 0)          # returns 17, default if missing
for name, age in ages.items():
    print(name, age)

When counting items, a dictionary or collections.Counter is often useful. For example, you can count the frequency of characters in a string using a dictionary with get.

在统计次数时，字典或 collections.Counter 非常有用。例如，你可以使用带 get 的字典统计字符串中每个字符的出现次数。

---

6. Sets: Unique Elements and Set Operations | 集合：唯一元素与集合运算

A set is an unordered collection of unique hashable elements. Sets are mutable and support O(1) average membership testing. Storing duplicates in a set has no effect because each element can only appear once.

集合是无序且唯一元素的容器，元素必须可哈希。集合是可变的，平均情况下成员检查为 O(1)。由于集合中每个元素只能出现一次，所以重复存储不会生效。

Common set methods include add, remove, discard, union, intersection, difference and symmetric_difference. Operators such as |, &, - and ^ provide a shorter syntax.

常用集合方法包括 add、remove、discard、union、intersection、difference 和 symmetric_difference。运算符 |、&、- 和 ^ 提供了更简短的写法。

s = {1, 2, 3}
s.add(3)               # no duplicate added
t = {2, 3, 4}
print(s & t)           # {2, 3}
print(s | t)           # {1, 2, 3, 4}
print(s - t)           # {1}
print(s ^ t)           # {1, 4}

Use a set whenever you need to remove duplicates from a sequence or test membership quickly. However, sets do not support indexing or ordering, so you cannot access an element by position.

当你需要去除序列中的重复项或快速测试成员身份时，应使用集合。但集合不支持索引和顺序，因此不能按位置访问元素。

---

7. Mutability, Aliasing, and Copying | 可变性、别名与拷贝

In Python, assignment does not copy an object. When you write b = a for a list, both names point to the same object. This is called aliasing. Changing the list through one name affects the other.

在 Python 中，赋值不会复制对象。当你对列表执行 b = a 时，两个名字指向同一个对象，这称为别名。通过其中一个名字修改列表，另一个名字也会受到影响。

For immutable types such as strings and tuples, aliasing is less dangerous because the object cannot change. For mutable types such as lists, dictionaries and sets, you must decide whether you need a copy.

对于字符串和元组等不可变类型，别名问题相对安全，因为对象不能改变。对于列表、字典和集合等可变类型，你必须判断自己是否需要一份拷贝。

a = [1, 2, 3]
b = a                # alias, not a copy
b.append(4)
print(a)             # [1, 2, 3, 4]

c = a.copy()         # shallow copy
c.append(5)
print(a)             # [1, 2, 3, 4]

A shallow copy copies the container but not the nested objects. If the list contains other lists, the inner lists are still shared. Use copy.deepcopy to fully copy nested structures.

浅拷贝会复制外层容器，但不会复制嵌套对象。如果列表中包含其他列表，内部列表仍然共享。使用 copy.deepcopy 可以完整复制嵌套结构。

---

8. List Comprehensions and Generator Expressions | 列表推导式与生成器表达式

Comprehensions give a compact way to build lists, dictionaries and sets. The general form is [expression for item in iterable if condition]. They are common in exam-style code, so you must be able to trace and write them.

推导式提供了一种简洁地构建列表、字典和集合的方式。一般形式是 [expression for item in iterable if condition]。它们在考试代码中很常见，你必须能够追踪和编写它们。

A list comprehension creates a complete list in memory. If you only need to iterate once, use a generator expression with parentheses () instead; it saves memory.

列表推导式会在内存中创建完整列表。如果只需要迭代一次，可以使用带圆括号的生成器表达式 ()，这样更节省内存。

squares = [x * x for x in range(10)]
evens = [x for x in range(20) if x % 2 == 0]
squares_dict = {x: x * x for x in range(5)}
unique_set = {x % 3 for x in range(10)}
total = sum(x * x for x in range(10))  # generator expression

Understanding comprehension order is important. The loop part is written from left to right in the same order as a normal nested loop. Exam questions often ask you to predict the output of a comprehension.

理解推导式的执行顺序非常重要。循环部分从左到右书写，和普通嵌套循环的顺序一致。考试题目经常要求你预测推导式的输出结果。

---

9. Common Operations and Complexity | 常用操作与复杂度

Time complexity is a frequent exam topic. The table below summarises the average complexity of common operations.

时间复杂度是常见的考试主题。下表总结了常见操作的平均复杂度。

Structure	Operation	Average Complexity
List	Index access	O(1)
List	Append	O(1) amortised
List	Insert / delete middle	O(n)
Tuple	Index access	O(1)
String	Find / slice	O(n)
Dict	Get / set / delete by key	O(1) average
Set	Add / remove / membership	O(1) average
List / Tuple	Membership test with in	O(n)

Remember that dictionary and set complexities depend on hash quality. In a good hash table, lookup is O(1), but in the worst case it can become O(n).

请记住，字典和集合的复杂度依赖于哈希质量。在良好的哈希表中，查找是 O(1)，但最坏情况下可能退化为 O(n)。

---

10. Stacks and Queues with Collections | 用 collections 实现栈与队列

Stacks and queues are common abstract data types. In Python, a list can be used as a stack with append and pop, giving LIFO order.

栈和队列是常见的抽象数据类型。在 Python 中，列表可以用 append 和 pop 作为栈使用，遵循后进先出 LIFO 顺序。

For a queue, using pop(0) on a list is slow because shifting elements takes O(n). The collections.deque object supports fast appends and pops from both ends, so it is a better choice for a queue.

对于队列，在列表上使用 pop(0) 较慢，因为移动元素需要 O(n)。collections.deque 支持从两端快速添加和弹出元素，因此它更适合实现队列。

stack = []

stack.append('a')

Published by TutorHao | Computer Science Revision Series | aleveler.com

Find Computer Science Textbooks on eBay UK
New, used and second-hand copies of textbooks and revision guides are often much cheaper than retail — check current listings and prices before you buy.
Browse on eBay UK →
更多咨询请联系16621398022（同微信）
Share this:

				
				Email
			
More

				
				LinkedIn
			

				
				X
			

				
				Facebook
			

				
				Tumblr
			

				
				Reddit
			

				
				Pinterest
			

				
				Telegram
			

				
				WhatsApp
			
Like this:
Like Loading…

📚 Structuring Transactional Writing: Challenges of Form and Expression | 文书类文章的结构与表达难点

Transactional writing appears in letters, emails, reports, proposals, speeches, and articles. It asks you to consider a reader, a context, and a purpose at the same time. Many candidates find it difficult because a good transactional piece must be structurally clear and stylistically convincing. This guide explains the main difficulties and offers practical strategies for planning, writing, and refining your own documents.

文书类文章出现在信件、电子邮件、报告、提案、演讲稿和文章中。它要求你同时考虑读者、语境和写作目的。许多考生觉得它困难，是因为一篇好的文书既要结构清晰，又要在表达上令人信服。本指南将解释其中的主要难点，并为你提供从构思、写作到修改的实用策略。

---

1. What Is Transactional Writing? | 什么是文书类文章

Transactional writing is writing that wants a response from the reader. A letter asking for permission, a report recommending changes, a speech persuading people to recycle, and an article raising awareness about stress are all transactional documents. They are not just descriptions or explanations; they are pieces of writing with a job to do.

文书类文章是一种期望读者作出回应的写作。请求许可的信件、提出改进建议的报告、劝说人们回收利用的演讲稿，以及唤起人们对压力关注的杂志文章，都属于文书类文章。它们不只是描述或解释，而是带有明确任务的写作。

The core challenge is balance. A transactional text must be organised enough for the reader to follow, yet expressive enough to sound natural. Examine questions often reward candidates who can control structure, tone, and rhetorical effect in the same piece of writing.

其核心挑战在于平衡。文书类文章必须有足够的组织结构，让读者能够跟上思路；同时也要有足够的表达力，让语言显得自然。许多考试题目往往同时考察考生对结构、语气和修辞效果的控制能力。

---

2. Audience and Purpose: The Starting Point | 读者与目的：写作的起点

Before planning the structure, you must identify the audience and the purpose. The audience determines how formal the language should be and which arguments will be persuasive. The purpose determines whether you should inform, persuade, request, advise, or complain.

在规划结构之前，你必须明确读者和写作目的。读者决定语言该有多正式、哪些论点更有说服力；目的则决定你应该告知、说服、请求、建议还是投诉。

A useful table of four key questions can guide you:

一个实用的“四问题”框架可以引导你：

Question | 问题	Decision | 影响
Who is the reader? | 读者是谁？	Age, knowledge, attitude, and relationship to you. | 年龄、知识、态度和与你关系。
What is my purpose? | 我的目的是什么？	The action or feeling you want to produce. | 你希望产生的行动或情感。
What text form is required? | 要求写什么文体？	Letter, speech, article, report, or proposal. | 信件、演讲稿、文章、报告或提案。
What tone is appropriate? | 什么语气合适？	Formal, professional, friendly, or passionate. | 正式、专业、友好或富有激情。

If you skip this stage, the writing often becomes generic. A letter to a local newspaper, for example, needs a different voice from a letter to a close friend. Use the text form, audience, and tone to choose vocabulary and sentence types.

如果跳过这一阶段，写出来的文章往往会显得空泛。例如，写给地方报纸的信与写给密友的信，语气和用词应当不同。你要根据文体、读者和语气来选择词汇与句型。

---

3. Structural Conventions by Document Type | 各文书体裁的结构规范

Each transactional form has an expected pattern. Examiners do not require an identical formula every time, but you should show awareness of the conventions that readers expect.

每一种文书文体都有其预期模式。考官并不要求每次都套用完全相同的公式，但你需要表现出对这些阅读惯例的认识。

Form | 文体	Typical Structure | 典型结构	Expression Focus | 表达重点
Formal Letter | 正式信件	Salutation; statement of purpose; main points; polite close. | 称呼；说明目的；主体要点；礼貌结束语。	Polite modal verbs; clear requests; formal linking phrases. | 礼貌的情态动词；清晰的请求；正式连接语。
Speech | 演讲稿	Greeting; engaging introduction; two or three arguments; memorable close. | 问候；吸引人的开头；两三个论点；令人难忘的结尾。	Direct address; rhetorical questions; repetition for impact. | 直接称呼读者；反问；重复以增强效果。
Article | 文章	Headline; lead; developed paragraphs; rounded conclusion. | 标题；导语；展开段落；完整结尾。	Vivid vocabulary; varied sentence length; engagement devices. | 生动的词汇；长短句交替；吸引读者的手法。
Report / Proposal | 报告 / 提案	Title; introduction; findings; recommendations; conclusion. | 标题；引言；调查结果；建议；结论。	Objective tone; numbered points; precise evidence. | 客观语气；分点编号；准确证据。

Notice that a report often uses headings, subheadings, and numbered recommendations. An article, by contrast, should feel less rigid. The best writers vary the pattern according to the reader and the context.

注意报告通常使用标题、小标题和编号建议；而杂志文章则应显得不那么刻板。优秀的作者会根据读者和语境灵活调整模式。

---

4. The Opening: Framing Your Intent | 开头：建立意图与语境

The opening is where many candidates lose marks. A weak opening merely repeats the prompt; a strong opening states the context, identifies the purpose, and establishes a relationship with the reader.

开头是许多考生失分的地方。平庸的开头只会复述题目；优秀的开头会交代背景、说明目的，并与读者建立联系。

For a formal letter, an opening such as “I am writing to express my concern about the planned changes to the school timetable” is clear and direct. For a speech, a rhetorical question may work better: “Have you ever wondered why our lessons feel so crowded?”

对于正式信件，可以用“我写信是为了表达我对学校课表调整计划的担忧”这样清晰直接的句子。对于演讲稿，反问句可能效果更好：“你是否曾想过，为什么我们的课堂显得如此拥挤？”

Context + Purpose + Relationship = Effective Opening

In Chinese, the same principle applies: 开头要先交代语境，再点明意图，最好还能暗示你与读者的关系。例如提议报告中可写“本报告旨在分析食堂运营情况，并提出三项可行建议”。

This pattern is also useful in Chinese: first provide the background, then state your aim, and suggest the relationship with your reader. A proposal can open with “本报告旨在分析食堂运营情况，并提出三项可行建议”。

---

5. Body Paragraphs: Developing Points with Coherence | 主体段：连贯地展开要点

The body of a transactional text should contain a small number of focused points. Each paragraph should present one idea, support it with reason or evidence, and then link it to your overall purpose.

文书的主体部分应当包含少量、紧凑的论点。每个段落应只讨论一个要点，用理由或证据支撑，然后将其与整体目的联系起来。

Point → Reason → Evidence → Link

For example, in a speech supporting longer breaks, a paragraph could work like this: “Longer breaks improve concentration. When students have time to move and talk, they return to class with more energy. In a recent study, schools with twenty-minute breaks reported a rise in classroom participation.”

例如，在一场支持延长课间休息的演讲中，段落可以这样展开：“更长的课间休息能提高专注力。当学生有时间活动和交流，他们返回课堂时更有精力。近期一项研究表明，实行二十分钟课间休息的学校，课堂参与度有所上升。”

The “link” sentence is often forgotten. It reminds the reader how this point connects to the next stage of the argument, such as “This is why cost-free scheduling changes can have an immediate impact.”

“连接句”经常被忽略。它提醒读者这一论点与下一阶段论证的关系，例如：“这就是为什么无需额外费用的时间调整也能立即产生效果。”

---

6. Signposting and Transitional Expressions | 路标与过渡表达

Signposting helps the reader navigate the text. In transactional writing, clear transitions are not optional; they create coherence and make the argument easier to follow.

路标词帮助读者在文中辨明方向。在文书类写作中，清晰的过渡并非可有可无；它们能创造连贯性，让论证更容易被理解。

• To add points: “Furthermore”, “In addition”, “Moreover” | 表示递进：“此外”“而且”“更重要的是”。

• To contrast: “However”, “On the other hand”, “Nevertheless” | 表示转折：“然而”“另一方面”“尽管如此”。

• To show cause: “As a result”, “Consequently”, “Therefore” | 表示因果：“因此”“结果”“所以”。

• To conclude: “In conclusion”, “To sum up”, “Finally” | 表示总结：“总之”“综上所述”“最后”。

Do not use these words mechanically. In a short letter, too many signposts make the writing sound artificial. Use only the ones the reader actually needs.

不要机械地堆砌路标词。在简短的信件中，过多路标会使文章显得做作。你只需使用读者真正需要的过渡表达。

---

7. Tone and Register: Formal vs. Persuasive | 语气与语域：正式与说服

Tone and register are among the most difficult features to control. Formal documents usually avoid contractions, slang, and overly emotional language. Persuasive documents, such as speeches and articles, can use repetition and direct address to create energy.

语气与语域是最难把握的特征之一。正式文书通常不使用缩写、俚语或过度情绪化的语言。而说服性文书，如演讲稿和杂志文章，则可以通过重复和直接称呼来制造感染力。

Compare these two examples:
“The school should do something about the problem.”
“It is essential that the school takes immediate action to address this issue.”

比较以下两个例子：
“学校应该对这个问题做点什么。”
“学校必须立即采取行动解决这一问题。”

The first sentence sounds weak and informal; the second sounds assured and formal. The choice depends on your audience. In a persuasive speech to students, an informal sentence can be powerful. In a proposal to the headteacher, a formal tone is safer.

第一句显得无力且随意；第二句显得坚定而正式。选择取决于读者。如果是在学生面前发表演讲，非正式句子可以很有力量；如果是向校长提交提案，正式语气则更稳妥。

A common difficulty is mixing registers. Avoid writing “The government should totally sort out the issue” in the same report as “It is evident that resources are inadequate.” Choose one consistent voice and maintain it.

常见困难是语域混杂。不要在同一篇报告中既写“政府应该彻底搞定这个问题”，又写“可见资源明显不足”。选择一种一致的语气并贯穿全文。

---

8. Difficulties in Expression: Avoiding Rhetorical Traps | 表达难点：规避修辞陷阱

Many learners struggle because they use vague words, recycle clichés, or string together long sentences without clear control. These habits weaken the force of transactional writing.

许多学习者之所以感到困难，是因为他们使用模糊词汇、套用陈词滥调，或者堆砌缺乏控制的长句。这些习惯会削弱文书的表达力度。

• Vague words: “good”, “bad”, “nice”, “things” have little meaning. Use precise nouns and qualifiers. | 模糊词：“好”“坏”“不错”“东西”缺乏具体含义。请使用精确名词和限定词。

• Overused phrases: “in today’s society”, “every coin has two sides” sound empty in exam writing. | 陈词滥调：“在当今社会”“凡事都有两面性”在考试写作中显得空洞。

• Run-on sentences: joining too many ideas with “and” or “also” loses the reader. Split the ideas. | 连写句：用“和”“也”连接过多想法会让读者迷失。请拆分句子。

• Absolute statements: “Everyone agrees” is easy to attack. Use “Most people would accept”. | 绝对化表述：“所有人都同意”很容易被反驳。可以写“大多数人都会接受”。

To fix these problems, read each sentence and ask: “Is this precise enough? Would this sentence convince someone who disagrees with me?”

要解决这些问题，请逐句自问：“这句话够精确吗？它能否说服一个不同意我的人？”

---

9. Effective Persuasive Devices | 高效说服手段

Persuasion is not about shouting. In transactional writing, it is about choosing the right rhetorical device for the right audience. The following devices are especially useful.

说服不是靠大声疾呼，而是要在文书写作中为合适的读者选择恰当的修辞手段。以下手法尤其有效。

• Rule of three: “slower, cheaper, and fairer” creates rhythm. | 三词结构：“更慢、更便宜、更公平”能产生节奏感。

• Rhetorical question: “Why should we wait?” challenges the reader. | 反问：“我们为何还要等？”向读者提出挑战。

• Direct address: “You will benefit from this change” builds a personal connection. | 直接称呼：“你将受益于这一改变”能建立个人关联。

• Contrast: “Before this proposal, students waited; after it, they will act” highlights change. | 对比：“在该提案之前，学生只能等待；实施后，他们将采取行动。”突显变化。

• Anecdote: a short example from daily life gives the argument a human face. | 轶事：一个日常生活中的短小例子能让论点更有温度。

These devices should support the structure, not replace it. A speech full of rhetorical questions but no clear body is still weak. Use one or two devices per paragraph, and always return to the point.

这些手法应服务于结构，而不是替代结构。一场满是反问却没有清晰主体的演讲仍然是失败的。每个段落使用一两个手法，并始终回到中心论点。

---

10. The Ending: Call to Action or Summary | 结尾：行动号召或总结

The ending must satisfy the reader. A good conclusion restates the original purpose, summarises the main idea, and tells the reader what to do next. Many candidates introduce a new argument in the conclusion; this confuses the message.

结尾必须让读者感到完整。好的结尾会重申最初的目的，概括主要观点，并告诉读者下一步该做什么。许多考生会在结尾引入新论点，这反而扰乱了信息。

For a formal letter, a conventional close is effective: “I hope you will give this matter your careful consideration.” For a speech, a call to action is stronger: “Let us decide today that our school will be a healthier place.” For a report, end with a recommendation or summary of findings.

正式信件可以采用传统结束语：“希望您能慎重考虑此事。”演讲稿可以给出更强有力的行动号召：“让我们今天就决定，把学校变成更健康的地方。”报告则应以建议或结果小结收尾。

Make sure the final sentence sounds intentional. Avoid finishing with “That is all” or “Bye”. The last line should leave a clear impression.

请确保最后一句话听起来是有意为之。避免用“就这样吧”或“再见”收尾。最后一行应当留下清晰的印象。

---

11. Common Mistakes and How to Fix Them | 常见错误与纠正

Below is a table of common mistakes seen in transactional writing, with concrete fixes.

下表列出文书类写作中的常见错误，并给出具体修改建议。

Mistake | 错误	Example | 例子	Fix | 修改
Copying the prompt | 复述题目	“I am writing to talk about school uniform.”	“I am writing to question the school’s current uniform policy.”
Mixed register | 语域混杂	“This is a really massive problem.”	“This issue requires immediate attention.”
No paragraph focus | 中心不明确	A paragraph with three unrelated points.	One point, one paragraph, one link.
Ignoring audience | 忽视读者	Using a joke in an official complaint letter.	Use polite, serious, and measured language.

When you revise your work, check for these errors deliberately. Read as if you were the examiner, not the writer.

修改文章时，请有意识地检查这些错误。你要站在考官而非作者

Published by TutorHao | English Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

📚 Data Structures Core Concepts & Common Exam Questions | 数据结构核心要点与常见题型

Data structures are the backbone of computer science. They define how data is organised, stored, and manipulated in memory, and every algorithm runs on top of some data structure. This article covers the essential data structures you need to master for your A Level or IGCSE Computer Science exam, along with typical exam questions and how to approach them.

数据结构是计算机科学的基石。它定义了数据在内存中如何组织、存储和操作，每一个算法都运行在某种数据结构之上。这篇文章将覆盖你在A Level或IGCSE计算机科学考试中必须掌握的核心数据结构，以及典型题型和解题思路。

---

1. What Is a Data Structure? | 什么是数据结构

A data structure is a systematic way of organising data so that it can be used efficiently. Different structures are suited to different tasks: some optimise for fast searching, others for quick insertion or deletion. Understanding the strengths and trade-offs of each structure is a key exam skill.

数据结构是系统地组织数据以便高效使用的方式。不同的结构适合不同的任务：有些优化快速查找，有些优化快速插入或删除。理解每种结构的优势和权衡是关键的考试技能。

• Linear structures: arrays, linked lists, stacks, queues. | 线性结构：数组、链表、栈、队列。

• Non-linear structures: trees, graphs. | 非线性结构：树、图。

• Abstract data types (ADTs): a theoretical model defined by its behaviour, such as a stack or queue.

抽象数据类型（ADT）是由其行为定义的理论模型，例如栈或队列。ADT 关注”做什么”而不是”怎么做”，而具体的数据结构实现则关心”怎么做”。

---

2. Arrays | 数组

An array is a contiguous block of memory holding elements of the same type, each identified by an index. Random access is constant time O(1) because the memory address is computed directly from the index.

数组是一块连续的内存区域，存放相同类型的元素，每个元素由索引标识。因为内存地址可以直接从索引计算得出，所以随机访问的时间复杂度为常数时间 O(1)。

Address of element i = Base Address + i × Size of element

元素 i 的地址 = 基地址 + i × 元素大小。这一公式在考试中常用于计算数组元素的内存位置。

• Advantages: fast access, cache-friendly, low memory overhead. | 优点：访问快、缓存友好、内存开销低。

• Disadvantages: fixed size, expensive insertion/deletion in the middle. | 缺点：大小固定，中间插入/删除开销大。

Common exam question: A 1D array of 20 integers starts at memory address 1000. Each integer occupies 4 bytes. What is the address of the 15th element?

常见题型：一个包含20个整数的数组从内存地址1000开始，每个整数占4字节。第15个元素的地址是多少？

1000 + 14 × 4 = 1056. (Index 14 because arrays are 0-based.)

1000 + 14 × 4 = 1056。（因为数组从0开始编号，所以是第14个索引。）

---

3. Linked Lists | 链表

A linked list is a linear data structure where each node contains data and a pointer (reference) to the next node. Nodes are not stored contiguously; they are scattered in memory and linked by pointers.

链表是一种线性数据结构，每个节点包含数据和指向下一个节点的指针（引用）。节点不连续存储，而是分布在内存各处，通过指针连接起来。

• Singly linked list: each node has one ‘next’ pointer. | 单向链表：每个节点有一个”下一个”指针。

• Doubly linked list: each node has both ‘next’ and ‘previous’ pointers. | 双向链表：每个节点有”下一个”和”上一个”指针。

• Circular linked list: the last node points back to the first. | 循环链表：最后一个节点指向第一个节点。

Insertion and deletion at the head of a linked list are O(1). Searching and accessing an element by index are O(n) because you must traverse from the head.

在链表头部插入和删除是 O(1) 时间。按索引搜索和访问元素是 O(n) 时间，因为必须从头开始遍历。

Common exam question: Draw a diagram to show the effect of inserting a new node X between nodes A and B in a singly linked list. State which pointers are changed.

常见题型：画出在单向链表中节点 A 和 B 之间插入新节点 X 的示意图，并指出哪些指针发生了改变。

Solution: ① Point X.next to B. ② Point A.next to X. The order matters: if you change A.next first, you lose the reference to B.

解答：① 将 X.next 指向 B。② 将 A.next 指向 X。顺序很重要：如果先改 A.next，就会丢失对 B 的引用。

---

4. Stacks | 栈

A stack is a Last-In-First-Out (LIFO) structure. The last element added is the first to be removed. Core operations are push (add to top) and pop (remove from top), plus peek (view the top without removing).

栈是一种后进先出（LIFO）结构。最后添加的元素最先被移除。核心操作是 push（压入栈顶）和 pop（弹出栈顶），以及 peek（查看栈顶但不移除）。

Time complexity: push O(1), pop O(1), peek O(1)

时间复杂度：push O(1)、pop O(1)、peek O(1)

• Real-world applications: undo/redo in editors, browser back buttons, function call stacks.

• 实际应用：编辑器中的撤销/重做、浏览器的后退按钮、函数调用栈。

Common exam question: A stack initially contains [5, 2, 8] (8 is at the top). Perform: pop, push 3, pop, push 9. What is the final content of the stack?

常见题型：栈初始内容为 [5, 2, 8]（8在栈顶）。执行：pop、push 3、pop、push 9。栈的最终内容是什么？

Solution: After pop → [5, 2]. Push 3 → [5, 2, 3]. Pop → [5, 2]. Push 9 → [5, 2, 9]. Final stack: [5, 2, 9] with 9 at the top.

解答：pop后 → [5, 2]。push 3 → [5, 2, 3]。pop → [5, 2]。push 9 → [5, 2, 9]。最终栈：[5, 2, 9]，9在栈顶。

---

5. Queues | 队列

A queue is a First-In-First-Out (FIFO) structure. Elements are added at the rear (enqueue) and removed from the front (dequeue). This models real-world waiting lines.

队列是一种先进先出（FIFO）结构。元素从队尾加入（入队 enqueue），从队首移除（出队 dequeue）。这模拟了现实世界中的排队场景。

“Static queue” implemented with an array has a problem: as elements are dequeued from the front, the empty spaces cannot be reused unless you shift all elements. A circular queue solves this by wrapping around to the front of the array.

用数组实现的”静态队列”有一个问题：当元素从队首出队后，空出的位置无法复用，除非将所有元素前移。循环队列通过回绕到数组前端来解决这个问题。

• Applications: print job scheduling, CPU process scheduling, buffering in networking.

• 应用：打印任务调度、CPU进程调度、网络中的缓冲处理。

Common exam question (circular queue): A circular queue of size 6 uses two pointers: front = 2 and rear = 4. Enqueue 3 more elements. Show the queue state and state whether the queue is full.

常见题型（循环队列）：大小为6的循环队列，front = 2，rear = 4。再入队3个元素。画出队列状态并判断队列是否已满。

Solution: Enqueue 1st → rear = 5. Enqueue 2nd → rear = 0. Enqueue 3rd → rear = 1. Now front = 2, rear = 1. The queue has 5 elements (indices 2,3,4,5,0). Not yet full. The full condition for a circular queue with one empty slot reserved is rear + 1 = front (mod size), i.e., if another enqueue occurs, rear would become 2 = front, meaning full.

解答：入队第1个 → rear = 5。入队第2个 → rear = 0。入队第3个 → rear = 1。此时 front = 2，rear = 1，队列中有5个元素（索引2,3,4,5,0），尚未满。循环队列预留一个空位的满条件为 rear + 1 = front（对size取模），即如果再入队一个元素，rear 变为 2 = front，即满。

---

6. Binary Trees | 二叉树

A binary tree is a hierarchical structure where each node has at most two children, called left and right. A binary search tree (BST) maintains ordering: the left subtree of a node contains only values less than the node, and the right subtree only values greater than the node.

二叉树是一种层次结构，每个节点最多有两个孩子，称为左孩子和右孩子。二叉搜索树（BST）维持有序性：节点的左子树只包含小于该节点的值，右子树只包含大于该节点的值。

Traversal methods (遍历方法):

Method 方法	Order 访问顺序	Typical use 典型用途
Preorder 前序	Root → Left → Right 根→左→右	Copying a tree 复制树
Inorder 中序	Left → Root → Right 左→根→右	Output values in sorted order 按序输出BST值
Postorder 后序	Left → Right → Root 左→右→根	Deleting a tree 删除树

For a balanced BST, search, insertion and deletion are all O(log n). For a skewed BST (worst case, e.g., inserting sorted data), these degrade to O(n).

对于平衡BST，查找、插入和删除都是 O(log n)。对于倾斜BST（最坏情况，例如插入已排序的数据），这些操作退化为 O(n)。

Common exam question: Show the result of inserting values 40, 20, 60, 10, 30, 50, 70 into an empty BST, then perform an inorder traversal.

常见题型：将值 40, 20, 60, 10, 30, 50, 70 插入空BST中，然后进行中序遍历。

Solution: Inorder traversal of a BST always outputs values in ascending order: 10, 20, 30, 40, 50, 60, 70. No need to draw the tree if you know this property.

解答： BST的中序遍历总是按升序输出值：10, 20, 30, 40, 50, 60, 70。知道这条性质就无需画树。

---

7. Graphs | 图

A graph consists of a set of vertices (nodes) and a set of edges connecting pairs of vertices. Graphs can be directed or undirected, weighted or unweighted. The two main representations are the adjacency matrix and the adjacency list.

图由一组顶点（节点）和一组连接顶点对的边组成。图可以是有向或无向的、加权或无权重的。两种主要表示方式是邻接矩阵和邻接表。

Representation 表示法	Space 空间	Check edge 边查询	Find neighbours 找邻居
Adjacency matrix 邻接矩阵	O(V²)	O(1)	O(V)
Adjacency list 邻接表	O(V+E)	O(degree) O(度)	O(degree) O(度)

The adjacency matrix is better for dense graphs where edge checks are frequent; the adjacency list is more memory-efficient for sparse graphs.

邻接矩阵更适合需要频繁查边的稠密图；邻接表对稀疏图更节省内存。

Common exam question: Draw the adjacency matrix for the directed graph with edges: A→B, A→C, B→C, C→A.

常见题型：画出有向图 A→B、A→C、B→C、C→A 的邻接矩阵。

Solution (rows = source, columns = destination):

	A	B	C
A	0	1	1
B	0	0	1
C	1	0	0

解答：行表示起点，列表示终点，有边则填1，无边填0。

---

8. Sorting Algorithms | 排序算法

Sorting arranges elements in a defined order (ascending or descending). Exam boards commonly test bubble sort, insertion sort, and merge sort. You must know their time complexities and be able to trace them.

排序将元素按特定顺序排列（升序或降序）。考试局常考冒泡排序、插入排序和归并排序。你必须知道它们的时间复杂度并能够手动追踪执行过程。

Algorithm 算法	Best 最优	Average 平均	Worst 最差	Stable 稳定?
Bubble sort 冒泡	O(n)	O(n²)	O(n²)	Yes 是
Insertion sort 插入	O(n)	O(n²)	O(n²)	Yes 是
Merge sort 归并	O(n log n)	O(n log n)	O(n log n)	Yes 是

A stable sort preserves the relative order of equal elements. Merge sort requires O(n) additional space for merging; bubble and insertion sort sort in place.

稳定排序保持相等元素的相对顺序。归并排序需要 O(n) 的额外空间用于合并；冒泡排序和插入排序是原地排序。

Common exam question: Trace two passes of bubble sort on the array [7, 2, 9, 1, 5].

常见题型：对数组 [7, 2, 9, 1, 5] 手动追踪冒泡排序的前两趟。

Solution: Pass 1: compare 7 & 2 → swap [2,7,9,1,5]; 7 & 9 no swap; 9 & 1 → swap [2,7,1,9,5]; 9 & 5 → swap [2,7,1,5,9]. Pass 2: 2 & 7 no swap; 7 & 1 → swap [2,1,7,5,9]; 7 & 5 → swap [2,1,5,7,9].

解答：第1趟：比较7和2 → 交换得 [2,7,9,1,5]；7和9不交换；9和1 → 交换得 [2,7,1,9,5]；9和5 → 交换得 [2,7,1,5,9]。第2趟：2和7不交换；7和1 → 交换得 [2,1,7,5,9]；7和5 → 交换得 [2,1,5,7,9]。

---

9. Searching Algorithms | 查找算法

Linear search and binary search are the two most commonly examined searching algorithms. Linear search scans every element from start to finish; binary search repeatedly halves a sorted array.

线性查找和二分查找是两个最常考的查找算法。线性查找从头到尾逐个扫描元素；二分查找反复对有序数组进行折半。

• Linear search: O(n) time, works on unsorted arrays. | 线性查找：O(n) 时间，适用于未排序数组。

• Binary search: O(log n) time, requires a sorted array. | 二分查找：O(log n) 时间，要求数组已排序。

Binary search pseudocode summary: set low = 0, high = n−1. While low ≤ high: calculate mid = (low + high) ÷ 2. If target = a[mid], return mid. If target < a[mid], set high = mid − 1. Otherwise set low = mid + 1.

二分查找伪代码摘要：令 low = 0，high = n−1。当 low ≤ high 时：计算 mid = (low + high) ÷ 2。如果目标等于 a[mid]，返回 mid。如果目标小于 a[mid]，令 high = mid − 1。否则令 low = mid + 1。

Common exam question: Use binary search to find the value 32 in the sorted array [2, 5, 9, 14, 21, 32, 47, 58]. List the mid indices checked.

常见题型：使用二分查找在有序数组 [2, 5, 9, 14, 21, 32, 47, 58] 中查找值32。列出每次检查的mid索引。

Solution: n = 8, initially low = 0, high = 7. mid = (0+7)÷2 = 3 → a[3] = 14, 32 > 14 → low = 4. mid = (4+7)÷2 = 5 → a[5] = 32, found! Indices checked: 3, 5.

解答： n=8，初始 low=0，high=7。mid=(0+7)÷2=3 → a[3]=14，32大于14 → low=4。mid=(4+7)÷2=5 → a[5]=32，找到了！检查的索引：3, 5。

---

10. Exam Strategy: Common Pitfalls | 应试策略：常见误区

Examiners repeatedly see the same mistakes in data structure questions. Avoiding these will earn you easy marks:

考官反复看到数据机构题目中的相同错误。避免这些错误可以轻松得分：

• Forgetting that arrays are 0-indexed in most programming contexts (list indexing errors). | 忘记大多数编程场景中数组从0开始编号（导致索引错误）。

• Using a queue algorithm on a stack problem, or vice versa (confusing FIFO with LIFO). | 在栈问题中使用队列算法，或反之（混淆FIFO与LIFO）。

• Applying binary search to an unsorted array. | 对未排序数组使用二分查找。

• Omitting the base case when tracing recursion on a tree. | 在树上做递归追踪时遗漏基本情况。

• Confusing the time complexity of insertion for arrays vs linked lists: arrays O(n) for middle insertion due to shifting; linked lists O(1) if the position pointer is known.

混淆数组和链表插入的时间复杂度：数组在中间插入需要移动元素所以是O(n)；链表如果已知位置指针则是O(1)。

Always state the time complexity of the operation you implement, and justify your choice of data structure for a given scenario by mentioning the time complexity of key operations.

始终说明你实现的操作的时间复杂度，并通过提及关键操作的时间复杂度来证明在给定场景下选择某个数据结构的合理性。

---

Published by TutorHao | Computer Science Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

📚 The Evolution and Impact of Great Power Relations in the 21st Century | 21世纪大国关系的演变与影响

The dawn of the 21st century fundamentally reshaped the landscape of international relations. The unipolar moment after the Cold War gave way to a more complex, multipolar system in which established powers and emerging states negotiated a new order.

21世纪的开启从根本上重塑了国际关系格局。冷战结束后的单极时刻让位于更加复杂的多极体系，既有大国与新兴国家在这一体系中对新秩序进行着博弈与协调。

---

1. The Transition from Unipolarity to Multipolarity | 从单极向多极的转型

The early 2000s appeared to confirm American dominance, yet the Iraq War and the 2008 financial crisis exposed the limits of unilateral action. Power became more diffuse, with China, the European Union, Russia, and India each commanding significant influence in their respective regions.

21世纪初，美国的优势地位似乎得到了确认，但伊拉克战争和2008年金融危机暴露了单边行动的局限。权力变得更加分散，中国、欧盟、俄罗斯和印度各自在其所在区域拥有重要影响力。

Multipolarity does not mean equality, but it marks the end of a single hegemonic narrative. The 2010s witnessed the rise of regional groupings such as BRICS and the Shanghai Cooperation Organisation, reflecting a deliberate effort to balance Western-centric institutions.

多极化并不意味着各国平等，但它标志着单一霸权叙事的终结。2010年代，金砖国家和上海合作组织等区域集团兴起，反映出各国刻意努力平衡以西方为中心的机构。

This structural shift influenced every major crisis of the century, from the 2011 Libya intervention to the 2014 Ukraine crisis, where the alignment of great powers determined diplomatic outcomes.

这一结构性转变影响了本世纪的每场重大危机，从2011年利比亚干预到2014年乌克兰危机，大国之间的立场协调决定了外交结局。

---

2. US-China Relations: From Engagement to Strategic Competition | 中美关系：从接触到战略竞争

At the start of the century, the United States welcomed China’s accession to the WTO in 2001, expecting economic integration to foster political liberalisation. This period of engagement produced deep trade interdependence but also growing strategic distrust.

世纪之初，美国欢迎中国于2001年加入世贸组织，期望经济一体化能促进政治自由化。这一接触时期带来了深度的贸易相互依存，同时也滋生了日益加深的战略不信任。

The 2010s saw a decisive shift. Washington’s “Pivot to Asia”, the Trump-era trade war, and the Biden administration’s “competition” framework reflected a bipartisan consensus that China was a revisionist power.

2010年代出现了决定性转向。华盛顿的“重返亚洲”、特朗普时代的贸易战以及拜登政府的“竞争”框架，反映出两党一致认为中国是一个修正主义大国。

Technology became the key battleground. Export controls on semiconductors, disputes over 5G networks, and restrictions on AI investment highlighted how technological supremacy is now central to great power rivalry.

技术成为关键战场。半导体出口管制、5G网络争端以及人工智能投资限制，凸显出技术霸权如今已成为大国竞争的核心。

---

3. Sino-Russian Partnership: Strategic Alignment or Marriage of Convenience? | 中俄伙伴关系：战略结盟还是权宜联姻？

After the Cold War, Russia and China gradually built a “comprehensive strategic partnership of coordination” based on shared opposition to Western interventionism. The 2001 Treaty of Good-Neighbourliness set the legal foundation for border stability and mutual trust.

冷战后，俄中两国逐步建立起基于共同反对西方干涉主义的“全面战略协作伙伴关系”。2001年《睦邻友好合作条约》为边境稳定和互信奠定了法律基础。

The partnership deepened after the 2014 Ukraine crisis. Russia turned to China for economic support, while China gained a reliable energy supplier and a partner in balancing NATO’s expansion.

2014年乌克兰危机后，这一伙伴关系进一步深化。俄罗斯转向中国寻求经济支持，而中国则由此获得了可靠的能源供应国以及制衡北约东扩的伙伴。

However, the relationship is not without limits. China has avoided providing lethal military aid to Russia in the Ukraine war, and both states retain distinct geopolitical interests in Central Asia and the Arctic.

然而，这种关系并非毫无限度。在乌克兰战争中，中国避免向俄罗斯提供致命性军事援助，双方在中亚和北极地区仍保留着各自不同的地缘政治利益。

---

4. US-European Relations: Alliance Under Stress | 美欧关系：压力下的同盟

The transatlantic alliance endured significant shocks in the twenty-first century. The 2003 Iraq War divided “Old Europe” from the United States, while the 2016 Brexit referendum weakened the EU’s global voice and complicated Washington’s coordination with Brussels.

跨大西洋联盟在21世纪经受了巨大冲击。2003年伊拉克战争使“老欧洲”与美国产生裂痕，而2016年英国脱欧公投削弱了欧盟在全球的声音，并使华盛顿与布鲁塞尔的协调更加复杂。

Donald Trump’s presidency fundamentally tested the alliance. He questioned NATO’s relevance, demanded greater European defence spending, and imposed tariffs on European goods. The Biden administration restored conventional alliances but continued to pressure Europe on China policy.

唐纳德·特朗普的总统任期从根本上考验了该联盟。他质疑北约的意义，要求欧洲增加防务开支，并对欧洲商品加征关税。拜登政府虽恢复了传统联盟，但在对华政策上继续向欧洲施压。

Europe responded by advancing “strategic autonomy”, culminating in the 2022 Strategic Compass and increased defence investment after the Ukraine invasion. Yet Europe remains divided between those who fear abandonment and those who fear entanglement.

欧洲则以推进“战略自主”作为回应，最终促成2022年《战略指南针》的出台以及乌克兰被入侵后防务投入的增加。然而，欧洲仍分裂为担忧被抛弃与担忧被卷入两派。

---

5. The Rise of Emerging Powers: India, Brazil, and Beyond | 新兴大国的崛起：印度、巴西及其他

India’s economic liberalisation since 1991 and its rapid growth made it a pivotal swing state in great power politics. New Delhi joined the Quadrilateral Security Dialogue (Quad) with the US, Japan, and Australia, yet maintained close energy and defence ties with Russia.

印度自1991年以来的经济自由化和快速增长，使其成为大国政治中的关键摇摆力量。新德里加入了美日印澳四边安全对话，同时与俄罗斯保持着密切的能源与防务关系。

Brazil, as the leader of South America, pushed for reformed international institutions. Its influence in BRICS and at climate negotiations demonstrated that middle powers could shape agendas even without full great power status.

巴西作为南美领导者，推动改革国际机构。它在金砖国家和气候谈判中的影响力表明，中等强国即使不具备完整的大国地位，也能塑造议程。

These emerging powers do not seek to overturn the system entirely but demand a redistribution of voice and representation. Their rise has made the G20 more relevant than the G7, illustrating the shift from Western exclusivity to global inclusion.

这些新兴大国并不寻求彻底推翻现有体系，而是要求重新分配话语权和代表性。它们的崛起使二十国集团比七国集团更具现实意义，体现出从西方排他性转向全球包容性的变化。

---

6. Economic Interdependence and Weaponised Interdependence | 经济相互依存与武器化的相互依存

Globalisation in the 1990s and 2000s created unprecedented economic integration. Supply chains stretched across continents, and trade volumes exploded. However, interdependence also created vulnerabilities, as states discovered that rivals could be coerced through economic means.

1990年代和2000年代的全球化创造了前所未有的经济一体化。供应链延伸至各大洲，贸易量急剧增长。然而，相互依存也带来了脆弱性，各国发现可以通过经济手段胁迫对手。

The concept of “weaponised interdependence” became prominent in the 2020s. The United States deployed financial sanctions and export controls, while China countered with export bans on rare earths and critical minerals.

“武器化的相互依存”概念在2020年代变得突出。美国运用金融制裁和出口管制，中国则以稀土和关键矿产出口禁令作为反制。

This trend has accelerated regionalisation and “friend-shoring”, as major powers seek to reduce dependence on hostile economies. The COVID-19 pandemic further exposed the fragility of global supply chains, prompting strategic stockpiles and domestic production subsidies.

这一趋势加速了区域化和“友岸外包”，各大国纷纷寻求减少对敌意经济体的依赖。新冠疫情进一步暴露了全球供应链的脆弱性，促使各国建立战略储备并实施国内生产补贴。

---

7. Global Governance: Cooperation and Deadlock | 全球治理：合作与僵局

Great power relations profoundly affect the workability of global institutions. The UN Security Council, designed for great power consensus, frequently fell into paralysis. Vetoes blocked resolutions on Syria and Ukraine, raising questions about the Council’s legitimacy.

大国关系深刻影响着全球机构的可运行性。以大国协商一致为设计基础的联合国安理会频频陷入瘫痪。否决权阻挡了关于叙利亚和乌克兰的决议，引发对安理会合法性的质疑。

Climate change emerged as the rare domain of continued cooperation. The Paris Agreement of 2015 represented a fragile compromise between the United States, China, and Europe, though subsequent withdrawals and re-entries undermined its stability.

气候变化成为少有的持续合作领域。2015年《巴黎协定》代表了美中欧三方之间脆弱的妥协，尽管后来的退出与重新加入削弱了其稳定性。

The pandemic and the Ukraine war revealed the absence of a functional global crisis management mechanism. The failure to coordinate vaccine distribution and economic sanctions illustrated both the need for and the impossibility of unified governance under bipolar rivalry.

疫情和乌克兰战争暴露出全球危机管理机制的缺失。疫苗分配和经济制裁协调的失败，既说明在两极竞争下统一治理的必要，也说明了其不可能性。

---

8. Technology, Information, and the New Frontier of Power | 科技、信息与权力的新前沿

Digital technology became the primary arena of great power competition. From artificial intelligence to quantum computing, states raced to secure strategic advantage. The United States and China dominate the field, creating a “tech bipolarity” that shapes alliances and supply chains.

数字技术成为大国竞争的主要舞台。从人工智能到量子计算，各国争相获取战略优势。美国和中国主导了这一领域，形成了塑造联盟和供应链的“科技两极化”。

Cyberspace has also become a zone of conflict. Cyber attacks on critical infrastructure, disinformation campaigns, and election interference emerged as tools of statecraft, operating below the threshold of armed conflict.

网络空间也成为冲突区域。对关键基础设施的网络攻击、虚假信息传播以及选举干预，逐渐成为低于武装冲突门槛的国家行为工具。

Information warfare accompanied every major geopolitical event, from the 2016 US presidential election to the 2022 Russian invasion of Ukraine. The control over platforms, satellite communications, and data flows now rivals traditional military power in importance.

信息战伴随了每一场重大地缘政治事件，从2016年美国总统选举到2022年俄罗斯入侵乌克兰。对平台、卫星通信和数据流的控制，其重要性如今可与传统军事力量相匹敌。

---

9. Regional Conflicts and the Great Power Chessboard | 区域冲突与大国博弈棋盘

Regional conflicts increasingly became proxy arenas for great power rivalry. In the Middle East, the Syrian civil war drew in Russia, the United States, Iran, and Turkey, complicating peace efforts and reshaping alliances.

区域冲突日益成为大国较量的代理人战场。在中东，叙利亚内战将俄罗斯、美国、伊朗和土耳其卷入其中，使和平努力更加复杂，并重塑了联盟格局。

Europe’s deadliest conflict since 1945, the Ukraine war, became the defining crisis of 21st-century great power relations. It united NATO and the EU against Russia, pushed Finland and Sweden into the alliance, and weaponised energy supplies.

作为自1945年以来欧洲最致命的冲突，乌克兰战争成为21世纪大国关系的标志性危机。它使北约和欧盟联合对抗俄罗斯，推动芬兰和瑞典加入联盟，并将能源供应武器化。

In the Indo-Pacific, territorial disputes in the South China Sea and the Taiwan Strait brought China into confrontation with the United States and its regional partners. Crisis management mechanisms, such as military hotlines, remain fragile but essential.

在印太地区，南海和台海的领土争端使中国与美国及其区域伙伴发生对抗。军事热线等危机管理机制虽仍显脆弱，却至关重要。

---

10. The Shifting Balance and Its Global Impact | 力量对比的变迁及其全球影响

The evolution of great power relations has produced a world of multiple overlapping crises: geopolitical, economic, and technological. The traditional hierarchy of states remains, but the line between “great” and “middle” powers has blurred.

大国关系的演变造就了一个多重危机叠加的世界：地缘政治、经济和技术层面均是如此。传统上国家的等级结构依然存在，但“大国”与“中等强国”之间的界限已经模糊。

For smaller states, this means both risk and opportunity. They can play off rivals to gain leverage, but they also risk being crushed in the gaps. The rise of supply-chain security and bloc politics threatens the open global order.

对较小国家而言，这既意味着风险也意味着机遇。它们可以在竞争者之间左右逢源以获取筹码，但也可能在大国夹缝中被碾碎。供应链安全和集团政治的兴起，正威胁着开放的国际秩序。

Ultimately, the 21st century has not returned to 19th-century balance-of-power politics. Instead, it has created a hybrid system where economic interdependence coexists with strategic rivalry, where global problems require joint action, and where no single power can impose its will.

最终，21世纪并未回到19世纪的均势政治。相反，它创造了一个混合体系：经济相互依存与战略竞争并存，全球问题需要联合行动，而任何单一强权都无法强加其意志。

The future of international order depends on whether today’s great powers can establish effective rules for competition, avoid catastrophic miscalculation, and manage the shared challenges of climate change, pandemics, and technological disruption.

国际秩序的未来取决于今天的大国能否为竞争建立有效规则，避免灾难性误判，并共同应对气候变化、大流行病和技术颠覆等共同挑战。

---

Published by TutorHao | History Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

📚 Capital Structure and Its Determinants | 公司资本结构及其影响因素

Capital structure refers to the way a company finances its overall operations and growth by using different sources of funds: debt, equity, or hybrid securities. It is one of the most critical decisions in corporate finance because it directly affects the firm’s risk, cost of capital, and value.

资本结构是指公司通过不同资金来源（债务、股权或混合证券）为其整体运营和发展融资的方式。它是公司金融中最关键的决策之一，因为它直接影响企业的风险、资本成本和企业价值。

---

1. What Is Capital Structure? | 什么是资本结构

In simple terms, capital structure is the mix of long-term sources of finance used by a company. A firm can raise funds by issuing shares (equity financing) or by borrowing money (debt financing). The proportion of debt to equity is the essence of capital structure.

简单来说，资本结构是公司所使用的长期融资来源的组合。企业可以通过发行股票（股权融资）或借入资金（债务融资）筹集资金。债务与股权的比例就是资本结构的核心。

A company with more debt is said to be highly geared (leveraged), while one with more equity is lowly geared. The target capital structure is the mix that minimises the cost of capital and maximises shareholder wealth.

债务较多的公司被称为高杠杆公司，而股权较多的公司为低杠杆公司。目标资本结构是使资本成本最小化并使股东财富最大化的组合。

资本结构通常以资产负债表的右侧表示，包含长期负债、优先股和普通股等项目。

---

2. Measuring Capital Structure | 资本结构的衡量指标

Financial analysts use several ratios to measure and compare a firm’s capital structure. The most common ones include the debt-to-equity ratio, the debt ratio, and the equity ratio.

财务分析师使用多种比率来衡量和比较企业的资本结构。最常用的包括债务权益比率、债务比率和权益比率。

• Debt-to-Equity Ratio = Total Debt / Total Equity
• 债务权益比率 = 总债务 / 总权益
• Debt Ratio = Total Debt / Total Assets
• 债务比率 = 总债务 / 总资产
• Equity Ratio = Total Equity / Total Assets
• 权益比率 = 总权益 / 总资产

These ratios indicate the financial risk of a company. A high debt-to-equity ratio signals high financial leverage, which may boost returns but also increases the probability of bankruptcy.

这些比率反映了公司的财务风险。高债务权益比率意味着高财务杠杆，这可能会提高回报，但也会增加破产的可能性。

---

3. Equity vs. Debt: Key Differences | 股权融资与债务融资的主要区别

Equity and debt are the two fundamental building blocks of capital structure. They differ in terms of ownership, cost, risk, and treatment of taxes.

股权和债务是资本结构的两个基本组成部分。它们在所有权、成本、风险和税务处理方面存在差异。

Feature | 特征	Debt | 债务	Equity | 股权
Ownership | 所有权	Creditors have no ownership | 债权人无所有权	Shareholders own the company | 股东拥有公司
Return | 回报	Fixed interest payment | 固定利息	Dividends, variable | 股息，不固定
Tax treatment | 税务处理	Interest is tax-deductible | 利息税前扣除	Dividends paid from after-tax profits | 股息从税后利润中支付
Risk | 风险	Lower risk for investors, but mandatory payments | 投资者风险较低，但还款是强制的	Higher risk, residual claim | 风险较高，剩余求偿权

---

4. Cost of Capital and WACC | 资本成本与加权平均资本成本

The cost of capital is the minimum return a company must earn on its investments to satisfy its investors. The weighted average cost of capital (WACC) reflects the blended cost of all sources of finance.

资本成本是公司为满足其投资者而必须从投资中获得的最低回报率。加权平均资本成本（WACC）反映了所有融资来源的混合成本。

WACC = (E/V) × Re + (D/V) × Rd × (1 – Tc)

Where E is the market value of equity, D is the market value of debt, V = E + D, Re is the cost of equity, Rd is the cost of debt, and Tc is the corporate tax rate.

其中E是权益的市场价值，D是债务的市场价值，V = E + D，Re是股权成本，Rd是债务成本，Tc是公司税率。

As debt is cheaper than equity due to tax shields and lower risk, adding debt can reduce WACC initially. However, beyond a certain point, financial distress costs push WACC upward.

由于税收抵免和较低的风险，债务成本低于股权成本，因此适当地增加债务可以降低WACC。然而，超过某一点后，财务困境成本会使WACC上升。

---

5. Modigliani-Miller Theorem | 莫迪利亚尼-米勒定理

The Modigliani-Miller (MM) theorem is a cornerstone of capital structure theory. In a world without taxes, transaction costs, or bankruptcy costs, MM argued that capital structure does not affect a firm’s value.

莫迪利亚尼-米勒（MM）定理是资本结构理论的基石。在没有税收、交易成本或破产成本的世界中，MM认为资本结构不影响企业价值。

This is because investors can replicate any capital structure by borrowing or lending on their own accounts. Thus, the total value of the firm depends only on its operating cash flows, not on the debt-equity split.

这是因为投资者可以通过自己的借贷行为复制任何资本结构。因此，企业的总价值仅取决于其经营现金流，而不是债务与权益的划分。

Once taxes are introduced, MM’s second proposition states that the value of a levered firm equals the value of an unlevered firm plus the present value of the interest tax shield.

一旦引入税收，MM第二定理指出，有负债企业的价值等于无负债企业的价值加上利息税盾的现值。

VL = VU + Tc × D

This suggests that firms should borrow as much as possible to maximise value, which is unrealistic in practice due to bankruptcy risks.

这表明企业应尽可能多地借款以使价值最大化，但这在实践中因破产风险而并不现实。

---

6. Trade-off Theory | 权衡理论

The trade-off theory extends MM by recognising real-world frictions. It argues that firms balance the tax benefits of debt against the costs of financial distress and potential bankruptcy.

权衡理论通过承认现实世界的摩擦对MM理论进行了扩展。该理论认为企业会在债务的税收利益与财务困境及潜在破产成本之间进行权衡。

As debt levels increase, the marginal benefit of each additional unit of debt falls, while the marginal cost rises. The optimal capital structure occurs where these two margins are equal.

随着债务水平上升，每增加一单位债务的边际收益下降，而边际成本上升。最优资本结构出现在这两个边际量相等之处。

Financial distress costs include direct costs like legal and administrative expenses, and indirect costs like loss of customers, suppliers, and employee morale.

财务困境成本包括直接成本（如法律和行政费用）和间接成本（如客户流失、供应商流失和员工士气下降）。

因此，企业不会无限制地使用债务，而会选择一个目标杠杆比率，这通常因行业和企业特性而异。

---

7. Pecking Order Theory | 优序融资理论

The pecking order theory, proposed by Myers and Majluf, suggests that firms prefer internal financing first, then debt, and finally equity as a last resort.

优序融资理论由迈尔斯和马吉卢夫提出，认为企业优先选择内部融资，其次是债务，最后才考虑股权融资。

This preference arises from information asymmetry: managers know more about the firm’s prospects than outside investors. Issuing equity is often seen as a signal that shares are overvalued, which can depress the stock price.

这种偏好源于信息不对称：管理者比外部投资者更了解公司的前景。发行股票通常被视为股价被高估的信号，可能会压低股价。

Therefore, profitable firms tend to accumulate retained earnings and borrow only when necessary, while issuing equity only when their debt capacity is exhausted.

因此，盈利能力强的公司倾向于积累留存收益，只在必要时负债，并且只在债务能力耗尽时才发行股票。

这一理论解释了为什么现实中许多盈利企业杠杆率反而较低，这与权衡理论的预测并不完全一致。

---

8. Internal Determinants of Capital Structure | 资本结构的内部影响因素

Several firm-specific factors influence the choice of capital structure. Profitability is one of the most significant: firms with high returns on assets tend to use less debt because they can finance growth from internal funds.

影响资本结构选择的公司特定因素有很多。盈利能力是最重要的因素之一：资产回报率高的公司使用的债务较少，因为它们可以利用内部资金为增长提供资金。

• Asset tangibility: firms with more tangible assets can pledge them as collateral, allowing more debt.
• 资产有形性：拥有更多有形资产的公司可以将其作为抵押品，从而允许更多债务。
• Growth opportunities: firms with strong growth prospects face higher costs of financial distress, so they tend to use more equity.
• 成长机会：具有强劲增长前景的公司面临更高的财务困境成本，因此倾向于使用更多股权。
• Firm size: larger firms are more diversified and have better access to credit markets, enabling higher leverage.
• 公司规模：大公司更加多元化，且更容易进入信贷市场，因此杠杆率可以更高。
• Managerial attitude: conservative managers prefer lower leverage to reduce financial risk.
• 管理层态度：保守的管理者倾向于降低杠杆以减少财务风险。

---

9. External Determinants of Capital Structure | 资本结构的外部影响因素

External factors also play a vital role in shaping financing decisions. Interest rates, tax policy, and market conditions can all push a company toward debt or equity.

外部因素在塑造融资决策中也起着重要作用。利率、税收政策和市场条件都可能推动公司选择债务或股权融资。

• Interest rate levels: low interest rates reduce the cost of debt and encourage borrowing.
• 利率水平：低利率降低了债务成本，鼓励企业借款。
• Tax rate: higher corporate tax rates increase the benefit of the debt tax shield, making debt more attractive.
• 税率：较高的公司税率提高了债务税盾的收益，使债务更具吸引力。
• Stock market conditions: a bull market may ease equity issuance, while a bear market makes debt more favourable.
• 股票市场状况：牛市可能使股票发行更容易，而熊市则使债务融资更为有利。
• Industry norms: firms often imitate the capital structures of their competitors to avoid losing market credibility.
• 行业惯例：公司往往会模仿竞争对手的资本结构，以免失去市场信誉。

---

10. Financial Leverage and Risk | 财务杠杆与风险

Financial leverage refers to the use of debt to amplify returns to shareholders. A rise in operating profits will cause a more than proportional increase in earnings per share when a firm is leveraged.

财务杠杆是指利用债务来放大股东回报。当公司使用杠杆时，营业利润的上升将导致每股收益更大比例的上升。

The degree of financial leverage (DFL) measures this sensitivity.

财务杠杆系数（DFL）衡量这种敏感性。

DFL = Percentage Change in EPS / Percentage Change in EBIT

High leverage magnifies profits in good times but also amplifies losses during downturns. This trade-off between higher returns and higher risk is fundamental to capital structure decisions.

高杠杆在景气时期会放大利润，但在经济下行时期也会放大损失。高回报与高风险之间的权衡是资本结构决策的基础。

此外，财务杠杆过高可能导致债务评级下调，增加再融资成本，甚至引发流动性危机。

---

11. EBIT-EPS Analysis | EBIT-EPS分析

The EBIT-EPS analysis is a practical tool used to compare different financing plans. It examines how alternative capital structures affect earnings per share at various levels of operating profit (EBIT).

EBIT-EPS分析是比较不同融资方案的实际工具。它考察在不同营业利润（EBIT）水平下，各种资本结构如何影响每股收益。

To find the indifference point, set the EPS formulas of two plans equal to each other and solve for EBIT.

为找到无差异点，将两个方案的EPS公式相等，并求解EBIT。

EPS = ((EBIT – Interest) × (1 – Tax rate) – Preference dividends) / Number of shares

If expected EBIT exceeds the indifference point, the higher-leverage plan yields higher EPS; otherwise, the equity-financed plan is superior.

如果预期EBIT高于无差异点，高杠杆方案将产生更高的每股收益；否则，股权融资方案更优。

---

12. Conclusion: Choosing the Optimal Capital Structure | 结论：选择最优资本结构

There is no single formula for the optimal capital structure. It depends on the unique circumstances of each firm, including industry, asset base, growth stage, and market conditions.

最优资本结构没有唯一的公式。它取决于每家企业的具体情况，包括行业、资产基础、发展阶段和市场条件。

In practice, firms combine theory with managerial judgement. They consider risk tolerance, investor expectations, and strategic flexibility when setting their target capital structure.

在实践中，企业将理论与管理判断相结合。在设定目标资本结构时，它们会考虑风险承受能力、投资者预期和战略灵活性。

Ultimately, a well-chosen capital structure lowers financing costs, enhances shareholder value, and provides the financial stability needed to weather economic cycles.

最终，合理的资本结构能够降低融资成本，提升股东价值，并提供抵御经济周期所需的财务稳定性。

---

Published by TutorHao | Business Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

📚 Survival Models: Concepts and Applications | 生存模型的概念与应用

Survival analysis is a branch of statistics that focuses on the time until an event of interest occurs. Whether the event is death, machine failure, or patient recovery, survival models provide a powerful framework for analyzing time-to-event data that standard statistical methods cannot handle. In this article, we explore the core concepts of survival models, the key functions that define them, and their real-world applications in actuarial science, medicine, and engineering.

生存分析是统计学的一个分支，专门研究从起点到某一感兴趣事件发生所经历的时间。无论该事件是死亡、机器故障还是患者康复，生存模型都为分析”时间到事件”数据提供了强大的框架，而这些数据是标准统计方法无法处理的。在本文中，我们将深入探讨生存模型的核心概念、定义它们的关键函数，以及它们在精算科学、医学和工程学中的实际应用。

---

1. What Is a Survival Model? | 什么是生存模型

A survival model is a statistical model that describes the probability distribution of the time T until an event occurs. The time variable T is always non-negative, and the event of interest is often called a “failure” in engineering contexts or a “terminal event” in medical contexts. Unlike ordinary regression models, survival models must account for censoring — a situation where the event has not yet been observed for some individuals by the end of the study.

生存模型是一种描述事件发生时间 T 的概率分布的统计模型。时间变量 T 始终为非负数，感兴趣的事件在工程领域常被称为”失效”，在医学领域常被称为”终点事件”。与普通回归模型不同，生存模型必须处理删失——即某些研究对象在研究结束时尚未观察到事件发生的情况。

The fundamental quantity in any survival model is the random variable T, which has a probability density function f(t) and a cumulative distribution function F(t) = P(T ≤ t). From these basic building blocks, we derive two specialized functions that form the heart of survival analysis: the survival function and the hazard function.

任何生存模型的基本量都是随机变量 T，它具有概率密度函数 f(t) 和累积分布函数 F(t) = P(T ≤ t)。从这些基本构件出发，我们推导出构成生存分析核心的两个专门函数：生存函数和风险函数。

---

2. Survival Function S(t) | 生存函数 S(t)

The survival function, denoted S(t), gives the probability that an individual survives beyond time t. Formally, it is defined as the complement of the cumulative distribution function:

生存函数（记作 S(t)）表示个体存活超过时间 t 的概率。形式上，它被定义为累积分布函数的补函数：

S(t) = P(T > t) = 1 − F(t) = ∫ₜ^∞ f(u)du

The survival function has three key properties. First, S(0) = 1, meaning at the start of observation, all individuals are alive. Second, S(t) is a non-increasing function of t, reflecting that survival probability cannot increase over time. Third, as t → ∞, S(t) → 0, meaning eventually the event will occur for everyone.

生存函数具有三个关键性质。第一，S(0) = 1，意味着在观察开始时所有个体都存活。第二，S(t) 是 t 的非增函数，反映生存概率不会随时间而增加。第三，当 t → ∞ 时，S(t) → 0，意味着最终每个人都会发生事件。

In actuarial science, the survival function is intimately related to life tables. For a life aged x, the survival probability is often written as ₜpₓ = S(t), which denotes the probability that a person age x survives t more years. This quantity is the cornerstone of premium calculation and pension liability valuation.

在精算科学中，生存函数与生命表密切相关。对于年龄为 x 的个体，生存概率通常记为 ₜpₓ = S(t)，表示年龄为 x 的人再多存活 t 年的概率。这一量是保费计算和养老金负债评估的基石。

---

3. Hazard Function h(t) | 风险函数 h(t)

The hazard function, also called the hazard rate or force of mortality, measures the instantaneous rate at which the event occurs at time t, conditional on survival up to time t. It is defined as:

风险函数（也称为危险率或死亡力）衡量在存活到时间 t 的条件下，事件在 t 时刻发生的瞬时速率。其定义为：

h(t) = lim(Δt→0) P(t ≤ T < t + Δt | T ≥ t) / Δt

Equivalently, the hazard function can be expressed in terms of the density and survival functions:

等价地，风险函数可以用密度函数和生存函数来表达：

h(t) = f(t) / S(t)

Unlike the survival function, which decreases over time, the hazard function can take any shape. It may be constant (as in the exponential distribution), increasing (as in aging populations where mortality rises with age), decreasing (as in the “burn-in” phase of manufactured components), or bathtub-shaped (common in human mortality across the lifespan).

与随时间递减的生存函数不同，风险函数可以呈现各种形态。它可能是常数（如指数分布）、递增（如死亡率随年龄上升的老龄化人群）、递减（如制造部件的”磨合期”），或浴缸形（人类一生中的死亡率常见此形态）。

---

4. Relationship between Survival and Hazard | 生存函数与风险函数的关系

The survival and hazard functions are two sides of the same coin. From the definition h(t) = f(t)/S(t) and the fact that f(t) = −dS(t)/dt, we can derive a fundamental differential equation:

生存函数和风险函数是同一枚硬币的两面。根据 h(t) = f(t)/S(t) 的定义以及 f(t) = −dS(t)/dt 的事实，我们可以推导出一个基本的微分方程：

h(t) = −d[ln S(t)]/dt

Integrating both sides from 0 to t yields the crucial connection between cumulative hazard and survival:

两边从 0 到 t 积分，就得到累积风险与生存之间的关键联系：

S(t) = exp(−∫₀^t h(u)du) = exp(−H(t))

Here, H(t) = ∫₀^t h(u)du is called the cumulative hazard function. This elegant exponential relationship means that if we can estimate the hazard function, we automatically obtain the survival function, and vice versa. Furthermore, the density function can be recovered as f(t) = h(t) · S(t).

这里的 H(t) = ∫₀^t h(u)du 称为累积风险函数。这一优雅的指数关系意味着，如果我们能估计风险函数，就能自动获得生存函数，反之亦然。此外，密度函数也可以通过 f(t) = h(t) · S(t) 恢复。

For actuarial students, it is particularly useful to memorize the discrete analogue: for integer-age life tables, the force of mortality μₓ is approximately equal to −ln(pₓ), where pₓ is the one-year survival probability. This approximation is excellent when mortality rates are low.

对于精算专业的学生来说，记住离散形式的对应关系特别有用：对于整数年龄生命表，死亡力 μₓ 近似等于 −ln(pₓ)，其中 pₓ 是一年生存概率。当死亡率较低时，这种近似非常精确。

---

5. Censoring | 删失

Censoring is the defining feature that distinguishes survival analysis from ordinary statistical analysis. Right censoring occurs when a subject is lost to follow-up or survives beyond the end of the study — we know only that their event time exceeds a certain value. Left censoring occurs when the event has already happened before observation begins. Interval censoring arises when we only know the event occurred within a specific time interval.

删失是生存分析与普通统计分析相区别的标志性特征。右删失发生在受试者失访或在研究结束时仍存活的情况——我们只知道他们的事件时间超过了某个值。左删失发生在观察开始之前事件已经发生的情况。区间删失出现在我们只知道事件在某个特定时间区间内发生的情况。

Ignoring censoring leads to severe bias. If we simply discard censored observations, we underestimate survival because we ignore the information that these individuals survived for a substantial period. If we treat censored times as actual event times, we overestimate the hazard. Survival models are specifically designed to incorporate censored data without bias.

忽略删失会导致严重的偏差。如果我们简单地丢弃删失观测值，就会低估生存率，因为我们忽略了这些个体存活了相当长一段时间的信息。如果我们把删失时间当作实际事件时间，又会高估风险。生存模型正是专门为无偏地纳入删失数据而设计的。

A common assumption in survival analysis is non-informative censoring, meaning that the censoring mechanism is independent of the event time. Under this assumption, censored observations contribute their survival information up to the censoring time through the survival function S(cᵢ), where cᵢ is the censoring time for subject i.

生存分析中一个常见的假设是非信息性删失，即删失机制与事件时间独立。在此假设下，删失观测通过生存函数 S(cᵢ) 贡献其截至删失时间的生存信息，其中 cᵢ 是第 i 个受试者的删失时间。

---

6. Kaplan-Meier Estimator | Kaplan-Meier 估计量

The Kaplan-Meier (KM) estimator, also known as the product-limit estimator, is a non-parametric method for estimating the survival function from observed data. It handles right-censored data naturally and produces a step-function estimate of S(t). The idea is straightforward: at each distinct event time t(j), we multiply the current survival estimate by the fraction of individuals at risk who survive that instant.

Kaplan-Meier（KM）估计量，也称为乘积极限估计量，是一种从观测数据估计生存函数的非参数方法。它自然地处理右删失数据，并产生 S(t) 的阶梯函数估计。其思想非常直接：在每个不同的事件时间 t(j) 处，将当前生存估计乘以在风险集中能存活过那一时刻的个体比例。

Ŝ(t) = ∏_{t(j) ≤ t} [1 − d(j)/n(j)]

where d(j) is the number of events at time t(j) and n(j) is the number of individuals at risk just before t(j). Censored observations are removed from the risk set at their censoring time but do not otherwise contribute a multiplicative factor.

其中 d(j) 是在时间 t(j) 发生的事件数，n(j) 是恰好在 t(j) 之前处于风险集中的个体数。删失观测在其删失时间从风险集中移除，但不另外贡献乘性因子。

The KM estimator is widely used to compare survival between groups — for example, comparing the survival of patients receiving two different treatments. A log-rank test can then be used to test whether the survival curves differ significantly. In actuarial work, KM methods are applied to analyse policy lapse rates and mortality improvement trends from portfolio experience data.

KM 估计量被广泛用于比较不同组间的生存——例如，比较接受两种不同治疗的患者生存率。随后可以使用对数秩检验来检验生存曲线是否存在显著差异。在精算工作中，KM 方法用于分析保单失效率和来自投资组合经验数据的死亡率改善趋势。

---

7. Parametric Models: Exponential and Weibull | 参数模型：指数分布与威布尔分布

While non-parametric methods are flexible, parametric models offer smoother estimates and allow extrapolation beyond the observation period. The simplest parametric model is the exponential distribution, which assumes a constant hazard:

虽然非参数方法灵活，但参数模型能够提供更平滑的估计，并允许超出观察期进行外推。最简单的参数模型是指数分布，它假设风险为常数：

h(t) = λ, S(t) = e^(−λt), 其中 t ≥ 0, λ > 0

The exponential model has the memoryless property: the probability of surviving an additional interval does not depend on how long one has already survived. This is unrealistic for human lifetimes but adequate for certain electronic components where failure is purely random.

指数模型具有无记忆性：再存活一个额外区间的概率不取决于已经存活了多久。这对于人类寿命来说是不现实的，但对于某些故障纯属随机的电子元件来说是合适的。

The Weibull distribution generalizes the exponential model by allowing the hazard to be monotonic. Its hazard function is:

威布尔分布通过允许风险为单调函数来推广指数模型。其风险函数为：

h(t) = λκ(λt)^(κ−1), κ > 0

When the shape parameter κ = 1, the Weibull reduces to the exponential distribution. When κ > 1, the hazard increases over time (appropriate for aging populations); when κ < 1, the hazard decreases (appropriate for infant mortality or burn-in failures). The Weibull survival function is S(t) = exp(−(λt)^κ).

当形状参数 κ = 1 时，威布尔分布退化为指数分布。当 κ > 1 时，风险随时间递增（适用于老龄化人群）；当 κ < 1 时，风险递减（适用于婴儿死亡率或磨合期故障）。威布尔生存函数为 S(t) = exp(−(λt)^κ)。

---

8. Cox Proportional Hazards Model | Cox 比例风险模型

The Cox proportional hazards model is the most widely used regression model in survival analysis. It relates the hazard function to a set of covariates without assuming a particular form for the baseline hazard:

Cox 比例风险模型是生存分析中应用最广泛的回归模型。它将风险函数与一组协变量联系起来，而无需假设基线风险的具体形式：

h(t | X) = h₀(t) · exp(β₁X₁ + β₂X₂ + … + β_pX_p)

Here, h₀(t) is an unspecified baseline hazard, and the covariates X₁, …, X_p enter through a multiplicative exponential term. The key assumption is proportionality: the hazard ratio between two individuals with different covariates is constant over time, since the ratio exp(βX₁)/exp(βX₂) does not depend on t.

其中 h₀(t) 是未指定的基线风险，协变量 X₁, …, X_p 通过乘性指数项进入模型。关键假设是比例性：具有不同协变量的两个个体之间的风险比随时间保持恒定，因为比值 exp(βX₁)/exp(βX₂) 不依赖于 t。

The coefficients β are estimated by maximizing the partial likelihood — a clever construction that cancels out the baseline hazard h₀(t). Interpretation of coefficients is in terms of log hazard ratios. For example, a coefficient β = 0.5 for a binary treatment indicator means that treatment multiplies the hazard by e^0.5 ≈ 1.65, so the treatment group has 65% higher instantaneous risk.

系数 β 通过最大化偏似然来估计——这是一种巧妙的构造，能够消去基线风险 h₀(t)。系数的解释以对数风险比的形式进行。例如，对于二元治疗指标，系数 β = 0.5 意味着治疗使风险乘以 e^0.5 ≈ 1.65，即治疗组的瞬时风险高出 65%。

---

9. Applications in Actuarial Science and Medicine | 在精算科学与医学中的应用

Survival models are indispensable in actuarial science. Life insurance pricing relies on survival functions from life tables to calculate net premiums. Pension actuaries use survival models to project the duration of benefit payments and to value defined-benefit pension plan liabilities. In health insurance, survival models help forecast the incidence of chronic diseases and the cost trajectory of long-term care policies.

生存模型在精算科学中不可或缺。人寿保险定价依赖生命表中的生存函数来计算净保费。养老金精算师使用生存模型来预测福利支付的持续时间，并对固定收益养老金计划的负债进行评估。在健康保险中，生存模型帮助预测慢性病的发病率和长期护理保单的成本轨迹。

In medical statistics, survival models are used in clinical trials to compare the efficacy of new drugs versus placebos, using endpoints such as overall survival or disease-free survival. The Kaplan-Meier curve, often accompanied by a log-rank test, is the standard output of any oncology trial. Cox regression adjusts for confounding factors like age, sex, and disease stage, enabling researchers to isolate the true treatment effect.

在医学统计中，生存模型在临床试验中用于比较新药与安慰剂的疗效，终点指标包括总生存期或无病生存期。Kaplan-Meier 曲线通常伴随对数秩检验，是任何肿瘤学试验的标准输出。Cox 回归调整年龄、性别和疾病分期等混杂因素，使研究人员能够分离出真正的治疗效果。

Beyond these fields, survival models power reliability engineering — predicting the lifetime of industrial components and scheduling preventive maintenance. In economics, they model the duration of unemployment spells; in sociology, the time to marriage or divorce; in customer analytics, the time until subscription cancellation. Anywhere “time until an event” matters, survival analysis provides the rigorous statistical toolkit.

除了这些领域，生存模型还驱动着可靠性工程——预测工业部件的寿命并安排预防性维护。在经济学中，它们用于建模失业持续期；在社会学中，用于分析结婚或离婚的时间；在客户分析中，用于研究订阅取消的时间。凡是”到事件发生的时间”重要的地方，生存分析都能提供严谨的统计学工具。

---

10. Conclusion | 结语

Survival models offer a coherent framework for analysing time-to-event data that ordinary regression cannot accommodate. The survival function S(t) describes the probability of surviving beyond time t; the hazard function h(t) captures the instantaneous risk; and the relationship S(t) = exp(−H(t)) unifies them. Censoring is handled naturally through likelihood-based inference, and tools such as the Kaplan-Meier estimator, Weibull models, and the Cox proportional hazards model provide a spectrum from fully non-parametric to fully parametric approaches.

生存模型为分析”时间到事件”数据提供了一个连贯的框架，而普通回归无法处理这类数据。生存函数 S(t) 描述存活超过时间 t 的概率；风险函数 h(t) 捕捉瞬时风险；关系式 S(t) = exp(−H(t)) 将二者统一起来。删失通过基于似然的推断被自然处理，Kaplan-Meier 估计量、威布尔模型和 Cox 比例风险模型等工具提供了从完全非参数到完全参数方法的光谱。

For students preparing for mathematics and actuarial examinations, mastering the definitions, interrelationships, and estimation techniques of survival models is not merely an academic exercise — it is a directly applicable skill in insurance, healthcare research, and data science. We encourage learners to practice deriving the key formulas from first principles and to apply them to real datasets. Understanding survival models will serve you well in any career where timing and risk intersect.

对于准备数学和精算考试的学生而言，掌握生存模型的定义、相互关系和估计技术不仅仅是学术练习——它在保险、医疗研究和数据科学中都是直接可用的技能。我们鼓励学习者从基本原理出发练习推导关键公式，并将其应用于真实数据集。理解生存模型将有助于你从事任何时间与风险交汇的职业。

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

📚 Four Practical Solution Methods for Partial Differential Equations | 偏微分方程的四种常用求解方法

Partial differential equations (PDEs) form the backbone of mathematical physics, engineering, and advanced applied mathematics. At A-Level and early university level, students are expected to recognise standard types of PDEs and apply systematic techniques to solve them. This article presents four essential methods: separation of variables, the method of characteristics, Fourier transforms, and Laplace transforms—each illustrated with typical examples and clear procedural steps.

偏微分方程（PDE）是数学物理、工程学以及高等应用数学的基石。在 A-Level 和大学初级阶段，学生需要能够识别标准类型的 PDE，并运用系统的方法求解。本文将介绍四种必备方法：分离变量法、特征线法、傅里叶变换法和拉普拉斯变换法——每种方法都配有典型例题和清晰的解题步骤。

---

1. Understanding the PDE Landscape | 认识偏微分方程的基本类型

Before selecting a solution method, you must classify the PDE. The three classical second-order linear PDEs are the wave equation ∂²u/∂t² = c²∂²u/∂x², the heat equation ∂u/∂t = k∂²u/∂x², and Laplace’s equation ∂²u/∂x² + ∂²u/∂y² = 0. Each has distinct physical meaning and demands a different approach.

在选择求解方法之前，必须先对 PDE 进行分类。三类经典的二阶线性 PDE 分别是：波动方程 ∂²u/∂t² = c²∂²u/∂x²、热传导方程 ∂u/∂t = k∂²u/∂x² 以及拉普拉斯方程 ∂²u/∂x² + ∂²u/∂y² = 0。每一种方程都有独特的物理含义，需要采用不同的求解策略。

The order, linearity, and boundary conditions determine which method is most efficient. A first-order PDE like uₓ + uᵧ = 0 can often be solved by characteristics, while a second-order PDE on a finite interval may suit separation of variables. Transform methods excel on infinite domains.

方程的阶数、线性性质以及边界条件决定了哪种方法最有效。一阶 PDE（如 uₓ + uᵧ = 0）通常可用特征线法求解；有限区间上的二阶 PDE 适合分离变量法；而变换方法在半无界或无界区域上表现尤为出色。

---

2. Separation of Variables: Principle | 分离变量法：基本原理

The core idea of separation of variables is to assume the solution can be written as a product of single-variable functions. For example, in the heat equation, we set u(x,t) = X(x)T(t). Substituting this product into the PDE allows us to separate the variables into two ordinary differential equations connected by a separation constant.

分离变量法的核心思想是假设解可以写成单变量函数的乘积形式。例如，对于热传导方程，我们令 u(x,t) = X(x)T(t)。将这个乘积代入 PDE 后，可以将变量分离为两个由分离常数联系的常微分方程。

Because each side depends on a different variable, both sides must equal a constant, conventionally written as −λ. This produces an eigenvalue problem for X(x) subject to the homogeneous boundary conditions, yielding discrete allowed values of λ.

由于等式两边分别依赖于不同的变量，两边必须都等于一个常数，通常记为 −λ。这便产生了一个关于 X(x) 的特征值问题，结合齐次边界条件，可以得到离散的允许值 λ。

---

3. Applying Separation of Variables to the Heat Equation | 用分离变量法求解热传导方程

Consider the heat equation uₜ = k uₓₓ for 0 < x < L, with boundary conditions u(0,t) = u(L,t) = 0 and initial condition u(x,0) = f(x). Setting u(x,t) = X(x)T(t) gives X(x)T'(t) = k X''(x)T(t), which rearranges to T'/kT = X''/X = −λ.

考虑热传导方程 uₜ = k uₓₓ，其中 0 < x < L，边界条件为 u(0,t) = u(L,t) = 0，初始条件为 u(x,0) = f(x)。令 u(x,t) = X(x)T(t) 可得 X(x)T'(t) = k X''(x)T(t)，整理后得到 T'/kT = X''/X = −λ。

The spatial equation X” + λX = 0 with X(0) = X(L) = 0 has non-trivial solutions only when λ = n²π²/L², giving Xₙ(x) = sin(nπx/L). The temporal equation gives Tₙ(t) = Cₙ exp(−n²π²kt/L²). By the superposition principle, u(x,t) = Σₙ₌₁^∞ Cₙ sin(nπx/L) exp(−n²π²kt/L²).

空间方程 X” + λX = 0 在 X(0) = X(L) = 0 的条件下，仅当 λ = n²π²/L² 时才有非平凡解，此时 Xₙ(x) = sin(nπx/L)。时间方程给出 Tₙ(t) = Cₙ exp(−n²π²kt/L²)。由叠加原理，u(x,t) = Σₙ₌₁^∞ Cₙ sin(nπx/L) exp(−n²π²kt/L²)。

Finally, apply the initial condition. Set t = 0, so f(x) = Σ Cₙ sin(nπx/L). Using the orthogonality of sine functions ∫₀ᴸ sin(nπx/L) sin(mπx/L) dx = L/2 δₙₘ, we obtain Cₙ = (2/L) ∫₀ᴸ f(x) sin(nπx/L) dx.

最后应用初始条件。令 t = 0，得到 f(x) = Σ Cₙ sin(nπx/L)。利用正弦函数的正交性 ∫₀ᴸ sin(nπx/L) sin(mπx/L) dx = L/2 δₙₘ，可得 Cₙ = (2/L) ∫₀ᴸ f(x) sin(nπx/L) dx。

---

4. The Method of Characteristics: First-Order PDEs | 特征线法：一阶偏微分方程

The method of characteristics reduces a first-order PDE to a system of ordinary differential equations along curves called characteristic lines. For a quasi-linear PDE of the form a(x,t)uₓ + b(x,t)uₜ = c(x,t,u), the characteristic equations are dx/ds = a, dt/ds = b, and du/ds = c.

特征线法将一阶 PDE 沿着称为“特征线”的曲线化简为一组常微分方程。对于形如 a(x,t)uₓ + b(x,t)uₜ = c(x,t,u) 的拟线性 PDE，其特征方程为 dx/ds = a，dt/ds = b，du/ds = c。

Along a characteristic curve, the PDE becomes an ODE for u. Once the family of characteristic curves is found, the solution is determined by prescribing initial data on a non-characteristic curve.

沿着特征曲线，PDE 变成了关于 u 的 ODE。一旦找到特征曲线族，就可以通过非特征曲线上的初始条件来确定解。

---

5. Characteristics in Action: The Advection Equation | 特征线法实战：平流方程

Solve the advection equation uₜ + c uₓ = 0, where c is a constant, with initial condition u(x,0) = f(x). The characteristic equations are dx/ds = c, dt/ds = 1, du/ds = 0. From dx/dt = c, we obtain x − ct = constant along each characteristic.

求解平流方程 uₜ + c uₓ = 0，其中 c 为常数，初始条件为 u(x,0) = f(x)。特征方程为 dx/ds = c，dt/ds = 1，du/ds = 0。由 dx/dt = c 可得，沿每一条特征线 x − ct = 常数。

Since du/ds = 0, u is constant along this family of straight lines. Therefore the solution is simply u(x,t) = f(x − ct), which represents a wave profile travelling to the right with speed c.

因为 du/ds = 0，所以 u 在整族直线上保持恒定。因此解就是 u(x,t) = f(x − ct)，表示一个以速度 c 向右传播的波形。

This method extends naturally to systems of first-order equations and to some second-order equations that can be factorised into first-order operators. It is especially useful in wave propagation and traffic flow problems.

这一方法可以自然推广到一阶方程组，也可用于一些可分解为一阶算子的二阶方程。它在波传播和交通流问题中尤其有用。

---

6. Fourier Transform: Concept and Key Properties | 傅里叶变换：概念与关键性质

For PDEs on infinite or semi-infinite domains, integral transforms often provide the most direct route. The Fourier transform of a function f(x) is defined as F(k) = ∫₋∞^∞ f(x) e^(−ikx) dx, with the inverse f(x) = (1/2π) ∫₋∞^∞ F(k) e^(ikx) dk.

对于无界或半无界区域上的 PDE，积分变换往往是最直接的路径。函数 f(x) 的傅里叶变换定义为 F(k) = ∫₋∞^∞ f(x) e^(−ikx) dx，其逆变换为 f(x) = (1/2π) ∫₋∞^∞ F(k) e^(ikx) dk。

The crucial property is that differentiation becomes multiplication: if F(k) is the transform of f(x), then the transform of f'(x) is ikF(k), and for the second derivative, f”(x) transforms to −k²F(k). This converts PDEs into ODEs (or even algebraic equations) in the transform variable.

关键性质在于微分变为乘法：若 F(k) 是 f(x) 的变换，则 f'(x) 的变换为 ikF(k)，二阶导数 f”(x) 的变换为 −k²F(k)。这可以将 PDE 转化为关于变换变量的 ODE（甚至代数方程）。

---

7. Fourier Transform Solution of Laplace’s Equation | 傅里叶变换求解拉普拉斯方程

Consider Laplace’s equation on the upper half-plane: uₓₓ + uᵧᵧ = 0, for −∞ < x < ∞, y > 0, with boundary condition u(x,0) = f(x) and the requirement that u → 0 as y → ∞.

考虑上半平面上的拉普拉斯方程：uₓₓ + uᵧᵧ = 0，其中 −∞ < x < ∞，y > 0，边界条件为 u(x,0) = f(x)，并要求当 y → ∞ 时 u → 0。

Applying the Fourier transform in x, we denote û(k,y) = ∫₋∞^∞ u(x,y) e^(−ikx) dx. The PDE becomes −k²û + ∂²û/∂y² = 0, which is an ODE in y. Its general solution is û(k,y) = A(k)e^(−|k|y) + B(k)e^(|k|y). To ensure boundedness as y → ∞, we set B(k) = 0.

对 x 变量应用傅里叶变换，记 û(k,y) = ∫₋∞^∞ u(x,y) e^(−ikx) dx。原 PDE 变为 −k²û + ∂²û/∂y² = 0，这是关于 y 的 ODE。其通解为 û(k,y) = A(k)e^(−|k|y) + B(k)e^(|k|y)。为保证当 y → ∞ 时有界，令 B(k) = 0。

Using the boundary condition, A(k) = F(k). Thus û(k,y) = F(k)e^(−|k|y). Inverting gives u(x,y) = (1/2π) ∫₋∞^∞ F(k)e^(−|k|y)e^(ikx) dk, which can be recognised as a convolution with the Poisson kernel.

利用边界条件可得 A(k) = F(k)。因此 û(k,y) = F(k)e^(−|k|y)。作逆变换得到 u(x,y) = (1/2π) ∫₋∞^∞ F(k)e^(−|k|y)e^(ikx) dk，这可以看作是与泊松核的卷积。

---

8. Laplace Transform: Tailored for Initial-Value Problems | 拉普拉斯变换：专为初值问题设计

The Laplace transform is defined as U(s) = ∫₀^∞ u(t)e^(−st) dt. It is particularly suited to PDEs where time t appears with first or second derivatives and where initial conditions at t = 0 are specified.

拉普拉斯变换定义为 U(s) = ∫₀^∞ u(t)e^(−st) dt。它特别适合处理时间变量 t 出现一阶或二阶导数且给定 t = 0 处初始条件的 PDE。

The key identities are: transform of u'(t) is sU(s) − u(0), and transform of u”(t) is s²U(s) − su(0) − u'(0). These formulas build the initial conditions directly into the transformed equation.

关键恒等式为：u'(t) 的变换是 sU(s) − u(0)，u”(t) 的变换是 s²U(s) − su(0) − u'(0)。这些公式将初始条件直接内嵌到变换后的方程中。

---

9. Laplace Transform Solution of the Wave Equation | 拉普拉斯变换求解波动方程

Solve the semi-infinite wave equation uₜₜ = c²uₓₓ for x > 0, t > 0, with u(x,0) = 0, uₜ(x,0) = 0, and the boundary condition u(0,t) = g(t). Taking the Laplace transform with respect to t, let U(x,s) = ℒ{u(x,t)}.

求解半无界波动方程 uₜₜ = c²uₓₓ，其中 x > 0，t > 0，初始条件为 u(x,0) = 0，uₜ(x,0) = 0，边界条件为 u(0,t) = g(t)。对时间变量 t 取拉普拉斯变换，令 U(x,s) = ℒ{u(x,t)}。

s²U(x,s) = c² ∂²U/∂x²

This ODE has the general solution U(x,s) = A(s)e^(−sx/c) + B(s)e^(sx/c). Requiring U to be bounded as x → ∞ forces B(s) = 0. The boundary condition gives U(0,s) = G(s), so A(s) = G(s).

该 ODE 的通解为 U(x,s) = A(s)e^(−sx/c) + B(s)e^(sx/c)。要求当 x → ∞ 时 U 有界可得 B(s) = 0。边界条件给出 U(0,s) = G(s)，因此 A(s) = G(s)。

We now have U(x,s) = G(s)e^(−sx/c). By the shift theorem, multiplying by e^(−sx/c) corresponds to a time delay. Hence u(x,t) = g(t − x/c) for t ≥ x/c and u(x,t) = 0 for t < x/c.

于是得到 U(x,s) = G(s)e^(−sx/c)。由位移定理，乘以 e^(−sx/c) 对应于时间延迟。因此 u(x,t) = g(t − x/c)（当 t ≥ x/c 时），而当 t < x/c 时 u(x,t) = 0。

---

10. Choosing the Right Method: A Strategic Guide | 选择正确的方法：策略指南

Selecting the appropriate method is a skill in itself. The domain geometry, the type of PDE, and the form of the boundary or initial conditions all influence this choice. The table below summarises the typical scenarios.

选择合适的方法本身是一项技能。区域几何形状、PDE 的类型以及边界或初始条件的形式都会影响选择。下表总结了典型适用场景。

Method	Best Used For	Key Advantage
Separation of Variables	Second-order linear PDEs on finite intervals with homogeneous boundary conditions	Produces eigenfunction series; transparent physical interpretation
Characteristics	First-order PDEs, especially advection and wave problems	Reduces PDE to ODE along curves
Fourier Transform	PDEs on infinite domains with spatial derivatives	Converts x-derivatives into algebraic factors
Laplace Transform	Initial-value problems in time on semi-infinite domains	Builds initial conditions directly into the solution

Transform methods often overlap; for instance, Laplace transforms can serve as an alternative to Fourier transforms for non-periodic functions on (0, ∞). The choice also depends on whether you prefer solving ODEs in a transform variable or summing infinite series.

变换方法常有重叠；例如，对于 (0, ∞) 上的非周期函数，拉普拉斯变换可以替代傅里叶变换。选择还取决于你更喜欢在变换变量中解 ODE，还是更喜欢对无穷级数求和。

---

11. Common Pitfalls and Working Accuracy | 常见陷阱与运算准确性

Students frequently make mistakes when applying separation of variables by forgetting to check whether the boundary conditions are homogeneous. If they are not, you must first homogenise them or handle the steady-state part separately.

学生在使用分离变量法时常见的错误是忘记检查边界条件是否为齐次。如果不是齐次边界条件，必须先进行齐次化处理，或者将稳态部分单独处理。

Another common pitfall is losing the absolute value in the Fourier solution of Laplace’s equation: writing e^(−ky) instead of e^(−|k|y) leads to divergent integrals for even initial data. Always check the decay requirement as y → ∞ before discarding the positive exponent term.

另一个常见陷阱是在拉普拉斯方程的傅里叶解中丢掉绝对值：写成 e^(−ky) 而不是 e^(−|k|y) 会导致对于偶初始数据积分发散。在舍弃正指数项之前，务必检查当 y → ∞ 时的衰减条件。

When using Laplace transforms, remember that the inverse transform must be considered piecewise, especially when delays such as e^(−sx/c) appear. The shift theorem applies only to functions that are zero before the shift time.

使用拉普拉斯变换时，记住逆变换必须分段考虑，特别是出现形如 e^(−sx/c) 的延迟项时。位移定理仅适用于在延迟时刻之前为零的函数。

---

12. Revision Checklist and Final Remarks | 复习清单与结语

To master the four methods, practise identifying the equation type first. For each problem, ask: Is the domain finite or infinite? Is the equation first or second order? Are the initial conditions or boundary conditions more prominent?

要掌握这四种方法，请先练习判断方程类型。每遇到一个问题，问自己：区域是有限还是无限？方程是一阶还是二阶？初始条件和边界条件哪一个更突出？

Separation of Variables suits finite intervals with two homogeneous boundary conditions. Method of Characteristics is the natural choice for first-order PDEs. Fourier Transform is ideal for infinite spatial domains, while Laplace Transform is preferred when time-domain initial conditions dominate.

分离变量法适用于具有两个齐次边界条件的有限区间；特征线法是处理一阶 PDE 的自然选择；傅里叶变换适合无界空间区域；而拉普拉斯变换在时间域初始条件占主导时更为优选。

Work through at least two full examples of each method from the A-Level specification. Focus on clean algebraic manipulation, correct application of orthogonality relations, and careful handling of inverse transforms. With structured practice, PDEs become a rewarding topic rather than an intimidating one.

请完成 A-Level 考纲中每种方法至少两道完整例题。重点关注简洁的代数变形、正交关系的正确应用以及逆变换的细致处理。通过系统训练，偏微分方程将从一个令人畏惧的课题变为一个收获满满的板块。

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

📚 Implicit Differentiation and Its Applications | 隐函数微分法及其应用

In A-Level Mathematics, differentiation is one of the most powerful tools you will master. While explicit functions of the form y = f(x) are straightforward to differentiate, many real-world relationships are expressed implicitly, where x and y are interwoven in a single equation. This article explores the implicit function theorem at the level required for advanced study, focusing on the technique of implicit differentiation and its wide-ranging applications.

在A-Level数学中，微分是你将掌握的最强大的工具之一。形如 y = f(x) 的显函数可以直接求导，但现实世界中的许多关系是以隐式表达的，x 与 y 交织在同一个方程之中。本文将探讨隐函数定理在高级学习阶段所需的核心内容，重点聚焦隐函数微分法及其广泛的应用。

---

1. What Is an Implicit Function? | 什么是隐函数？

An explicit function is written as y = f(x), where y is expressed solely in terms of x. For example, y = x² + 3x − 1 is explicit. An implicit function, by contrast, is one in which y is not isolated on one side of the equation. Instead, both x and y appear together, such as in x² + y² = 25 or x³ + y³ = 6xy.

显函数写作 y = f(x)，即 y 仅用 x 表达。例如，y = x² + 3x − 1 就是一个显函数。相反，隐函数中 y 并未被单独分离到等式的一侧，x 与 y 同时出现，例如 x² + y² = 25 或 x³ + y³ = 6xy。

Not every implicit equation can be rearranged into explicit form. The equation y⁴ + xy = sin x, for instance, cannot be solved for y using elementary operations. Nevertheless, we can still find dy/dx at any point where the curve is well-behaved. This is where implicit differentiation comes to our rescue.

并非每个隐式方程都能重排为显式形式。例如，方程 y⁴ + xy = sin x 就无法通过初等运算解出 y。然而，我们仍然可以在曲线表现良好的任意点处求出 dy/dx。这正是隐函数微分法的用武之地。

---

2. The Core Principle: The Chain Rule | 核心原理：链式法则

The fundamental idea behind implicit differentiation is the chain rule. When we differentiate a term involving y with respect to x, we must multiply by dy/dx. For example, the derivative of y² with respect to x is 2y·dy/dx, because y is itself a function of x. Similarly, the derivative of sin y is cos y · dy/dx.

隐函数微分法背后的基本思想是链式法则。当我们对包含 y 的项关于 x 求导时，必须乘以 dy/dx。例如，y² 关于 x 的导数是 2y·dy/dx，因为 y 本身是 x 的函数。类似地，sin y 的导数是 cos y · dy/dx。

d/dx [f(y)] = f′(y) · dy/dx

This simple rule allows us to differentiate equations that cannot be solved explicitly for y. The process mirrors standard differentiation, but every time we differentiate a y-term, we append dy/dx as a factor.

这条简单规则使我们能够对无法显式解出 y 的方程求导。整个过程的每一步都与标准微分一致，但每次对含 y 的项求导时，都要额外乘以 dy/dx 这个因子。

---

3. Step-by-Step Implicit Differentiation | 隐函数微分法的分步操作

Consider the equation x² + y² = 25, a circle of radius 5. To find dy/dx, we differentiate both sides of the equation term by term with respect to x:

考虑方程 x² + y² = 25，这是一个半径为 5 的圆。为了求 dy/dx，我们对等式两边逐项关于 x 求导：

d/dx (x²) + d/dx (y²) = d/dx (25)

This gives 2x + 2y·dy/dx = 0. Solving for dy/dx yields dy/dx = −x/y. Notice that the derivative is expressed in terms of both x and y, which is perfectly acceptable for implicit functions.

由此得到 2x + 2y·dy/dx = 0。解出 dy/dx，得到 dy/dx = −x/y。注意到导数同时包含 x 与 y，这对隐函数而言是完全合理的。

dy/dx = −x/y

At the point (3, 4) on the circle, the gradient is −3/4. At (0, 5), the gradient is 0, and at (5, 0), dy/dx is undefined because the tangent is vertical. This captures the geometry beautifully.

在圆上的点 (3, 4) 处，斜率为 −3/4。在 (0, 5) 处，斜率为 0；在 (5, 0) 处，由于切线为竖直方向，dy/dx 无定义。这完美地体现了曲线的几何性质。

---

4. Worked Example: A Classic A-Level Problem | 典型例题：一道经典的A-Level题目

Find the equation of the tangent to the curve x³ + y³ = 6xy (the folium of Descartes) at the point (3, 3).

求曲线 x³ + y³ = 6xy（笛卡尔叶形线）在点 (3, 3) 处的切线方程。

First, differentiate both sides with respect to x:

首先，对等式两边关于 x 求导：

3x² + 3y²·dy/dx = 6y + 6x·dy/dx

Group the terms containing dy/dx on one side:

将含有 dy/dx 的项归到等式一侧：

3y²·dy/dx − 6x·dy/dx = 6y − 3x²

Factor out dy/dx and solve:

提取公因式 dy/dx 并求解：

dy/dx = (6y − 3x²) / (3y² − 6x)

Substituting x = 3 and y = 3 gives dy/dx = (18 − 27) / (27 − 18) = −9/9 = −1. The tangent line is therefore y − 3 = −1(x − 3), which simplifies to y = −x + 6.

代入 x = 3 与 y = 3，得到 dy/dx = (18 − 27) / (27 − 18) = −9/9 = −1。因此切线方程为 y − 3 = −1(x − 3)，化简得 y = −x + 6。

---

5. Implicit Differentiation of Trigonometric and Exponential Terms | 三角与指数项的隐函数微分

Implicit differentiation extends naturally to functions involving trigonometric, exponential, and logarithmic terms. Consider the equation sin(xy) = x. To differentiate, we first use the chain rule on sin(xy), then the product rule on xy:

隐函数微分法自然可以推广到包含三角、指数和对数项的函数。考虑方程 sin(xy) = x。要对其求导，我们先用链式法则处理 sin(xy)，再用乘积法则处理 xy：

cos(xy) · (y + x·dy/dx) = 1

Expanding and rearranging:

展开并整理：

y·cos(xy) + x·cos(xy)·dy/dx = 1

dy/dx = [1 − y·cos(xy)] / [x·cos(xy)]

Similarly, for an equation like eʸ = x² + y, differentiating gives eʸ·dy/dx = 2x + dy/dx, from which we solve dy/dx = 2x / (eʸ − 1).

类似地，对于 eʸ = x² + y 这样的方程，求导得到 eʸ·dy/dx = 2x + dy/dx，从中解出 dy/dx = 2x / (eʸ − 1)。

---

6. Parametric Equations and Implicit Functions | 参数方程与隐函数的联系

There is a close relationship between parametric equations and implicit functions. A curve defined parametrically by x = f(t) and y = g(t) can often be converted into an implicit equation by eliminating the parameter t. Conversely, some implicit curves are best studied using a parameterisation.

参数方程与隐函数之间有密切的联系。由 x = f(t) 和 y = g(t) 参数化定义的曲线，通常可以通过消去参数 t 转化为隐式方程。反过来，某些隐式曲线用参数化方法来研究最为方便。

When working with parametric equations, we use dy/dx = (dy/dt) / (dx/dt), provided dx/dt ≠ 0. This formula is itself derived from the chain rule and complements implicit differentiation in the toolbox of advanced differentiation techniques.

在处理参数方程时，我们使用 dy/dx = (dy/dt) / (dx/dt)，前提是 dx/dt ≠ 0。该公式本身也是由链式法则推导而来，与隐函数微分法共同构成高等微分技巧工具箱中的互补工具。

---

7. Second Derivatives of Implicit Functions | 隐函数的二阶导数

Sometimes we need the second derivative, d²y/dx², of an implicit function. This requires differentiating dy/dx again with respect to x. Since dy/dx is typically expressed in terms of both x and y, we must use implicit differentiation on the derivative itself.

有时我们需要隐函数的二阶导数 d²y/dx²。这需要对 dy/dx 再次关于 x 求导。由于 dy/dx 通常同时包含 x 与 y，我们必须对导数本身再次使用隐函数微分法。

Take the circle x² + y² = 25 again. We know dy/dx = −x/y. Differentiating with respect to x using the quotient rule:

再次以圆 x² + y² = 25 为例。已知 dy/dx = −x/y。使用商法则关于 x 求导：

d²y/dx² = [−y − (−x)·dy/dx] / y² = (−y + x·dy/dx) / y²

Substituting dy/dx = −x/y gives d²y/dx² = −(x² + y²)/y³ = −25/y³. This elegant result tells us about the concavity of the circle at any point.

代入 dy/dx = −x/y，得到 d²y/dx² = −(x² + y²)/y³ = −25/y³。这一简洁的结果告诉我们圆在任意一点的凹凸性。

---

8. Applications to Rates of Change | 相关变化率中的应用

One of the most practical applications of implicit differentiation is in related rates problems, where two or more quantities change simultaneously. These problems are a staple of A-Level applications of differentiation.

隐函数微分法最实际的应用之一出现在相关变化率问题中，即两个或多个量同时变化的情形。这类问题是A-Level微分应用部分的常考内容。

For example, consider a ladder of length 5 m sliding down a wall. If the top of the ladder is slipping down at a rate of 0.5 m/s when the bottom is 3 m from the wall, how fast is the bottom moving? Let x be the distance from the wall and y the height of the top. Then x² + y² = 25, and differentiating with respect to time t gives:

例如，考虑一把 5 米长的梯子沿墙滑下。当梯子底端距墙 3 米时，若梯子顶端以 0.5 m/s 的速度下滑，底端移动的速度是多少？设 x 为底端到墙的距离，y 为顶端的高度。则 x² + y² = 25，关于时间 t 求导得：

2x·dx/dt + 2y·dy/dt = 0

When x = 3, y = √(25 − 9) = 4. Given dy/dt = −0.5, we solve:

当 x = 3 时，y = √(25 − 9) = 4。已知 dy/dt = −0.5，我们解出：

2·3·dx/dt + 2·4·(−0.5) = 0 → dx/dt = 2/3 m/s

The bottom of the ladder is moving away from the wall at a speed of 2/3 m/s.

梯子底端正以 2/3 m/s 的速度远离墙面。

---

9. Tangents and Normals on Implicit Curves | 隐式曲线的切线与法线

Implicit differentiation is indispensable for finding tangents and normals to curves that cannot be expressed explicitly. The gradient at a given point (x₁, y₁) is found by substituting the coordinates into the expression for dy/dx. The normal gradient is the negative reciprocal of the tangent gradient.

在求无法显式表达之曲线的切线与法线时，隐函数微分法不可或缺。只需将给定点 (x₁, y₁) 的坐标代入 dy/dx 的表达式，即可得到该点的斜率。法线斜率则是切线斜率的负倒数。

For instance, on the curve x² + xy + y² = 7, at the point (1, 2), differentiating implicitly gives 2x + y + x·dy/dx + 2y·dy/dx = 0. Substituting x = 1, y = 2 gives 2 + 2 + dy/dx + 4·dy/dx = 0, so dy/dx = −4/5. The normal at this point has slope 5/4.

例如，在曲线 x² + xy + y² = 7 上，于点 (1, 2) 处，隐式求导得到 2x + y + x·dy/dx + 2y·dy/dx = 0。代入 x = 1，y = 2，得 2 + 2 + dy/dx + 4·dy/dx = 0，因此 dy/dx = −4/5。该点的法线斜率为 5/4。

---

10. The General Implicit Function Theorem | 一般形式的隐函数定理

At a more advanced level, the implicit function theorem guarantees the existence of a differentiable function y = g(x) defined implicitly by F(x, y) = 0, provided that at a point (a, b) where F(a, b) = 0, the partial derivative ∂F/∂y is non-zero. In this case:

在更高级的层面，隐函数定理保证：若 F(x, y) = 0 在点 (a, b) 处满足 F(a, b) = 0，且偏导数 ∂F/∂y 在该点非零，则存在一个可微函数 y = g(x) 由 F(x, y) = 0 隐式定义。此时：

dy/dx = −(∂F/∂x) / (∂F/∂y)

This powerful formula provides a direct method for computing dy/dx without rearranging the equation. For F(x, y) = x² + y² − 25 = 0, we have ∂F/∂x = 2x and ∂F/∂y = 2y, giving dy/dx = −x/y, exactly as before.

这一强有力的公式提供了一种无需重排方程即可直接计算 dy/dx 的方法。对于 F(x, y) = x² + y² − 25 = 0，有 ∂F/∂x = 2x，∂F/∂y = 2y，得到 dy/dx = −x/y，与此前结果完全一致。

---

11. Common Pitfalls and Exam Tips | 常见误区与应试技巧

Students frequently make avoidable errors when applying implicit differentiation. The most common mistake is forgetting to multiply by dy/dx when differentiating y-terms. Another frequent error is mishandling the product rule when both x and y appear in the same term, such as xy.

学生在应用隐函数微分法时常犯可以避免的错误。最常见的错误是忘记在对含 y 的项求导时乘以 dy/dx。另一个常见错误是在同一项中同时包含 x 与 y 时（如 xy），未能正确处理乘积法则。

• Always check that every y-term produces a dy/dx factor after differentiation.
• Use the product rule carefully for terms like xy, x²y, or sin(xy).
• Substitute coordinates only after differentiating, not before.
• If a question asks for the equation of a tangent or normal, leave the final answer in a clear, simplified form such as y = mx + c or ax + by + c = 0.

在中文环境中，许多学生亦会在熟练度上有所欠缺。应对方法包括：多加练习不同类型的隐式方程，尤其是包含三角和指数函数的类型；在检验答案时，可将所得 dy/dx 与显式求解（如果可行）的结果进行对比验证。

• 务必检查每个含 y 的项求导后是否产生 dy/dx 因子。
• 对 xy、x²y 或 sin(xy) 等乘积项，仔细使用乘积法则。
• 先完成求导，再代入坐标，切勿颠倒顺序。
• 若题目要求切线或法线方程，最终答案应以清晰的简化形式给出，如 y = mx + c 或 ax + by + c = 0。

---

12. Summary and Further Reading | 总结与延伸阅读

Implicit differentiation is an essential technique that extends the power of calculus to equations that resist explicit rearrangement. By applying the chain rule to y-terms, we can compute dy/dx for a vast class of curves, find tangents and normals, solve related rates problems, and even compute second derivatives. The implicit function theorem provides a rigorous foundation for these operations and connects to the broader theory of multivariable calculus.

隐函数微分法是一项至关重要的技术，它将微积分的威力扩展到那些无法显式重排的方程。通过对含 y 的项应用链式法则，我们可以计算一大类曲线的 dy/dx，求切线与法线，解决相关变化率问题，甚至计算二阶导数。隐函数定理为这些操作提供了严谨的理论基础，并连接到多元微积分的更广阔理论。

To master this topic, practise with a variety of curves: circles, ellipses, folia, and implicitly defined trigonometric equations. Work through past-paper questions systematically, and verify each step with the chain rule in mind. With consistent practice, implicit differentiation becomes a reliable and rewarding part of your mathematical toolkit.

要掌握这一主题，请对各种曲线多加练习：圆、椭圆、叶形线以及隐式定义的三角方程。系统地完成历年真题，并在每一步中牢记链式法则。坚持不懈地练习之后，隐函数微分法将成为你数学工具箱中可靠而令人得心应手的部分。

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）