Numerical Methods for IB OCR Mathematics | IB OCR 数学：数值方法考点精讲

📚 Numerical Methods for IB OCR Mathematics | IB OCR 数学：数值方法考点精讲

Numerical methods provide powerful techniques for obtaining approximate solutions to problems that cannot be solved analytically. From finding roots of equations to evaluating definite integrals and solving differential equations, these algorithms are essential in both the IB Diploma Programme (Analysis & Approaches and Applications & Interpretation) and OCR A Level Mathematics. Mastering the underlying principles, error analysis, and practical implementation is crucial for success in the exam. This revision guide systematically covers all key concepts, complete with worked examples and examiner-friendly tips.

数值方法为解决无法解析求解的问题提供了强有力的近似计算技术。无论是求方程根、计算定积分还是求解微分方程，这些算法在 IB 文凭课程（分析与方法和应用与解释）和 OCR A Level 数学中都至关重要。掌握基本原理、误差分析以及实际应用是考试成功的关键。本考点精讲系统梳理核心概念，并配有例题和应考技巧。

1. What Are Numerical Methods? | 什么是数值方法？

Numerical methods refer to iterative algorithms that produce approximate numerical solutions. Unlike algebraic manipulation that yields an exact answer, these techniques converge towards a solution by repeating a set of calculations. They are particularly useful when dealing with non-linear equations, irregular integrals, or differential equations lacking closed-form solutions.

数值方法指通过迭代算法产生近似数值解。与通过代数变形获得精确解不同，这些技术通过重复一组计算逐步逼近真实解。在处理非线性方程、不规则积分或无解析解的微分方程时，数值方法尤其有用。

In the IB and OCR curricula, you will encounter methods for locating roots (Bisection, Newton-Raphson), numerical integration (Trapezoidal rule, Simpson’s rule), and first-order differential equations (Euler’s method). Emphasis is placed on understanding rate of convergence, error bounds, and the ability to implement these algorithms with a calculator or spreadsheet.

在 IB 和 OCR 课程中，你将接触到求根法（二分法、牛顿-拉夫森法）、数值积分法（梯形法则、辛普森法则）以及一阶微分方程数值解法（欧拉方法）。重点在于理解收敛速度、误差界限，以及使用计算器或电子表格实现这些算法的能力。

2. Error Analysis: Absolute & Relative Error | 误差分析：绝对误差与相对误差

Any numerical approximation carries an inherent error. The absolute error is the absolute difference between the true value x_true and the approximate value x_approx:

任何数值近似都带有固有误差。绝对误差是真实值 x_true 与近似值 x_approx 之差的绝对值：

Absolute error = | x_true – x_approx |

Relative error normalises this difference by the true value, often expressed as a percentage:

相对误差将这个差值除以真实值，通常以百分比表示：

Relative error = | (x_true – x_approx) / x_true | × 100%

In practice, x_true is unknown, so we use the approximation itself for relative error or rely on error bounds derived from the method. For iterative methods, checking the difference between successive approximations gives a practical stopping criterion:

实际中 x_true 未知，因此我们使用近似值本身来计算相对误差，或依赖方法给出的误差界。对于迭代方法，检查连续近似值之差提供了一个实用的停止准则：

| x_n+1 – x_n | < tolerance

Understanding error helps justify when an approximation is ‘good enough’ and is a frequent exam question, especially when asked to compare methods.

理解误差有助于评判近似何时“足够好”，也是常见的考题，尤其要求比较不同方法时。

3. Bisection Method | 二分法

The bisection method is a bracketing method that repeatedly halves an interval [a, b] where f(a) and f(b) have opposite signs, guaranteeing a root lies in between (Intermediate Value Theorem). At each step, compute the midpoint c = (a+b)/2, evaluate f(c), and replace either a or b with c depending on the sign of f(c). The process continues until the interval width is less than the desired tolerance.

二分法是一种区间划分法，不断对分区间 [a, b]，其中 f(a) 与 f(b) 异号，根据介值定理确保区间内存在根。每一步计算中点 c = (a+b)/2，求 f(c)，并根据 f(c) 的符号将 a 或 b 替换为 c。重复直到区间宽度小于预定容差。

Key properties:

关键性质：

Always converges if the function is continuous and sign change exists.

若函数连续且存在符号改变，则总收敛。
Linear convergence: the error is roughly halved each iteration.

线性收敛：误差大约每次迭代减半。
Number of iterations required to achieve a tolerance ε: n ≥ log₂((b-a)/ε).

达到容差 ε 所需迭代次数：n ≥ log₂((b-a)/ε)。

Example: Find √2 using bisection on f(x)=x²-2 in [1,2]. After 3 iterations, the interval is [1.375, 1.4375], giving an approximation of root ≈1.40625.

例如：用二分法求 √2，令 f(x)=x²-2，区间 [1,2]。迭代 3 次后区间为 [1.375, 1.4375]，近似根为 1.40625。

4. Newton-Raphson Method | 牛顿-拉夫森法

The Newton-Raphson method uses the tangent line at a current guess x_n to generate a better approximation. The iterative formula is:

牛顿-拉夫森法利用当前猜测值 x_n 处的切线来生成更好的近似。迭代公式为：

x_n+1 = x_n – f(x_n) / f'(x_n)

Geometrically, the next estimate is the x-intercept of the tangent. The method converges quadratically near a simple root, meaning the number of correct decimal places roughly doubles each step. However, convergence is not guaranteed; poor initial guesses or stationary points can cause divergence.

几何上，下一个估计值是切线与 x 轴的交点。在单根附近该方法具有二次收敛性，即小数点后正确位数每步大约翻倍。但收敛并不保证；不良的初值或驻点可能导致发散。

OCR and IB papers often ask for one or two iterations, or to show that the formula simplifies for a given function. For instance, to find ³√10, set f(x)=x³-10, then x_n+1 = (2x_n + 10/x_n²)/3. Always state the derivative clearly.

OCR 和 IB 考试常要求进行一两次迭代，或证明公式对给定函数简化。例如求 ³√10，设 f(x)=x³-10，则 x_n+1 = (2x_n + 10/x_n²)/3。必须清晰地写出导数。

5. Numerical Integration: Trapezoidal Rule | 数值积分：梯形法则

The trapezoidal rule approximates the definite integral ∫_a^b f(x) dx by dividing [a,b] into n strips of equal width h = (b-a)/n and summing the areas of trapezoids:

梯形法则通过将区间 [a,b] 等分为 n 个宽度 h = (b-a)/n 的条形，并计算梯形面积之和来近似定积分 ∫_a^b f(x) dx：

∫_a^b f(x) dx ≈ (h/2)[f(x₀) + 2f(x₁) + 2f(x₂) + … + 2f(x_n-1) + f(x_n)]

where x_i = a + i·h. The error is approximately bounded by (b-a)h²|f”(ξ)|/12 for some ξ in (a,b), so doubling the number of strips reduces error roughly by a factor of 4 (second-order accuracy). In exams, you will typically be given a table of values and asked to apply the rule with a specific number of ordinates.

其中 x_i = a + i·h。误差大约限制在 (b-a)h²|f”(ξ)|/12，ξ 为 (a,b) 内某点。因此增加条数加倍可使误差约为原来的 1/4（二阶精度）。考试常给出数值表，要求用特定纵坐标数应用该法则。

Caution: When calculating by hand, organise work with a clear table showing x and f(x) values. The multipliers 1,2,2,…,1 are critical; a common mistake is misplacing the 2’s.

注意：手算时请用清晰表格列出 x 与 f(x) 值。系数 1,2,2,…,1 至关重要；常见错误是放错 2 的位置。

6. Simpson’s Rule | 辛普森法则

Simpson’s rule provides a more accurate approximation by fitting quadratic polynomials through sets of three points. It requires an even number of strips (n even) and uses the formula:

辛普森法则通过三点拟合二次抛物线给出更精确的近似。它要求有偶数个条形（n 为偶数），公式如下：

∫_a^b f(x) dx ≈ (h/3)[f(x₀) + 4f(x₁) + 2f(x₂) + 4f(x₃) + 2f(x₄) + … + f(x_n)]

The pattern of coefficients is 1,4,2,4,2,…,4,1. Error term is proportional to h⁴, giving fourth-order accuracy: halving h reduces error by a factor of about 16. For functions with continuous fourth derivatives, Simpson’s rule is vastly superior to the trapezoidal rule with the same number of points.

系数模式为 1,4,2,4,2,…,4,1。误差项与 h⁴ 成正比，为四阶精度：h 减半，误差约降为 1/16。对于具有连续四阶导数的函数，辛普森法则在相同点数下远优于梯形法则。

An exam question may ask you to compare results from both rules, or to estimate the maximum error using the formula: |error| ≤ (b-a)h⁴|f⁽⁴⁾(ξ)|/180. However, IB and OCR often focus on applying the rule and understanding why n must be even.

考题可能要求比较两种法则的结果，或用公式 |error| ≤ (b-a)h⁴|f⁽⁴⁾(ξ)|/180 估计最大误差。但 IB 和 OCR 通常侧重于应用该法则，以及理解为何 n 必须为偶数。

7. Euler’s Method for Differential Equations | 欧拉方法求解微分方程

Euler’s method solves first-order ordinary differential equations dy/dx = f(x,y) with an initial condition y(x₀) = y₀. Starting at (x₀, y₀), the algorithm steps forward with a fixed step size h:

欧拉方法用于求解一阶常微分方程 dy/dx = f(x,y)，给定初始条件 y(x₀) = y₀。从 (x₀, y₀) 出发，算法以固定步长 h 向前推进：

x_n+1 = x_n + h

y_n+1 = y_n + h × f(x_n, y_n)

It uses the slope at the beginning of each interval to extrapolate linearly. This method is first-order accurate, meaning the global error is proportional to h. Small h gives more accuracy but increases computational steps. A typical exam task provides a table to complete for several steps and then sketches the approximate solution curve.

该方法利用每个区间起点的斜率进行线性外推。欧拉方法具有一阶精度，全局误差与 h 成正比。h 越小精度越高，但计算量增加。典型考题提供表格，要求完成若干步骤的填写，并绘制近似解曲线。

Don’t confuse Euler’s method with the improved Euler (Heun’s) method, which IB HL may mention, but standard Euler is the core competency. Be meticulous with decimal places; rounding prematurely can propagate error.

勿将欧拉方法与改进的欧拉方法（Heun 法）混淆，后者 IB HL 可能提及，但标准欧拉法是核心要求。谨防小数舍入过早导致误差累积。

8. Convergence and Choice of Method | 收敛性与方法选择

Choosing the right numerical tool depends on the problem and required accuracy. For equation solving, bisection is robust but slow; Newton-Raphson is fast if the derivative is available and initial guess is close. For integration, Simpson is preferred for smooth integrands, while trapezoidal is simpler but less accurate. Euler is the simplest ODE solver, but its error accumulates quickly over larger intervals.

选择合适的数值工具取决于问题类型和精度要求。对于方程求解，二分法稳健但缓慢；牛顿-拉夫森法在导数易求且初值合适时快速。对于积分，光滑被积函数优先使用辛普森法则，梯形法则简单但精度较低。欧拉方法是最简单的 ODE 求解器，但误差在大区间上累积较快。

Understanding convergence criteria is essential. A method converges if successive approximations get closer to the true solution. For iterative formulas like x_n+1=g(x_n), convergence requires |g'(x)| < 1 near the fixed point. Examiners may ask to show that an iteration converges, or to find an interval where it works.

理解收敛条件是基本要求。若连续近似值越来越接近真实解，则方法收敛。对于迭代公式 x_n+1=g(x_n)，在不动点附近需满足 |g'(x)| < 1 才收敛。考官可能要求证明某迭代收敛，或找出收敛的区间。

Table comparing methods:

各方法对比表：

Method	Application	Accuracy	Pitfalls
Bisection	Root-finding	Linear (halving interval)	Requires sign change; slow
Newton-Raphson	Root-finding	Quadratic (fast)	May diverge; needs derivative
Trapezoidal	Integration	O(h²)	Low accuracy for rough functions
Simpson	Integration	O(h⁴)	Even n required
Euler	ODEs	O(h)	Accumulated error; not stable for stiff ODEs

（表格：方法对比）

9. Exam Strategy & Common Mistakes | 考试策略与常见错误

Show full working for iterative methods, clearly storing intermediate values to at least one more decimal place than the required answer accuracy. Label your columns neatly. For numerical integration, copy ordinates accurately from the question; misreading a single f(x) value penalises the entire calculation. When asked to ‘show that’ an iteration formula arises from a given equation, rearrange the equation stepwise and explain the logic.

对于迭代法，写出完整过程，中间值至少多保留一位小数。整齐标注表格列。数值积分中要准确抄录纵坐标值；一个 f(x) 值读错会导致全题失分。当要求‘证明’迭代公式由给定方程导出时，逐步移项并解释逻辑。

Rounding errors: carrying out many iterations with prematurely rounded values can lead to drift. Maintain sufficient precision until the final answer, then round as instructed. Also, remember that an approximate solution obtained from, say, Newton-Raphson, does not prove that the root is exact—use appropriate language such as ‘x ≈ 1.618 (3 d.p.)’.

舍入误差：过早舍入会导致多次迭代后漂移。保留充分精度直至得出最后答案，再按要求舍入。同时注意，牛顿-拉夫森法等得出的近似解并不证明根是精确的——应使用适当表述如 ‘x ≈ 1.618（3 位小数）’。

Finally, manage time: some numerical methods questions involve many repetitive steps; if a question says ‘or otherwise find to 3 s.f.’, you might use a solver on a calculator to check, but always present the required working. IB Paper 3 (HL) may include deeper analysis of error and convergence, so be prepared to discuss order of convergence.

最后，合理分配时间：一些数值方法题包含大量重复步骤；如果题目说‘或用其他方法求至三位有效数字’，可用计算器求解进行验证，但务必给出所需过程。IB HL 试卷三可能包含更深的误差和收敛分析，需准备讨论收敛阶。

10. Real-World Connections & Practice | 实际联系与练习建议

Numerical methods are the backbone of computational science. Weather forecasting, structural engineering, and machine learning all rely on such algorithms. Understanding the trade-off between computational cost and accuracy prepares you for further study in applied mathematics and STEM fields.

数值方法是计算科学的基石。天气预报、结构工程和机器学习均依赖此类算法。理解计算成本与精度的权衡可为后续应用数学及 STEM 领域的学习做好准备。

To master this topic, practice past paper questions under timed conditions. After solving an iteration manually, check with a GDC (graphical display calculator) or Python script to verify results. Focus on explaining why a particular method is suitable, and always draw diagrams (for Newton-Raphson, sketch the tangent; for integration, sketch strips) to solidify understanding.

为掌握本专题，请在限时条件下练习历年真题。手动迭代完成后，用图形计算器或 Python 脚本验证结果。重点关注解释某方法为何合适，并始终画出示意图（牛顿-拉夫森法画切线，数值积分画条形），以巩固理解。

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

Numerical Methods for IB OCR Mathematics | IB OCR 数学：数值方法 考点精讲