📚 Numerical Methods for IB & WJEC Mathematics | IB 与 WJEC 数学：数值方法考点精讲
Numerical methods provide powerful techniques for solving mathematical problems that are difficult or impossible to handle analytically. From approximating roots of equations to estimating definite integrals and solving differential equations, these methods form a core part of the IB Mathematics: Analysis and Approaches (AA) and Applications and Interpretation (AI) syllabuses, as well as the WJEC A-level Mathematics specification. Mastering numerical methods not only helps you tackle exam questions but also builds a deeper understanding of how mathematics is applied in real-world situations.
数值方法为求解那些难以或无法用解析方法处理的数学问题提供了强有力的工具。无论是逼近方程的根、估算定积分还是求解微分方程，这些方法都是 IB 数学 AA 和 AI 课程以及 WJEC A-level 数学考试的核心内容。掌握数值方法不仅有助于应对考试中的相关题目，也能加深对数学在现实世界中应用方式的理解。
1. What Are Numerical Methods? | 什么是数值方法？
Numerical methods are algorithms that use arithmetic operations to produce approximate solutions to mathematical problems. Unlike analytical methods that give exact answers in terms of symbols, numerical methods often involve iteration – repeatedly applying a formula to get closer to the true solution.
数值方法是利用算术运算来生成数学问题近似解的算法。与用符号给出精确答案的解析方法不同，数值方法通常涉及迭代——即重复应用某个公式，使其结果越来越接近真实解。
They are used when an equation cannot be solved algebraically, such as eˣ = 3 – x, or when an integral has no elementary antiderivative, like ∫ e^(–x²) dx.
当方程无法用代数方法求解时（例如 eˣ = 3 – x），或者当积分没有初等原函数时（例如 ∫ e^(–x²) dx），就会用到数值方法。
In examinations, you may be asked to perform a few iterations of a method, analyse the accuracy of an approximation, or explain the conditions under which a method converges.
在考试中，你可能需要执行某个方法的几次迭代、分析近似值的精确度，或解释某种方法收敛的条件。
2. Errors in Numerical Analysis | 数值分析中的误差
Understanding error is crucial in numerical work. Absolute error is defined as |approximate value – true value|, while relative error is the absolute error divided by the true value. In most exam contexts, the true value is unknown, so we monitor how successive approximations change.
理解误差在数值计算中至关重要。绝对误差定义为 |近似值 – 真实值|，而相对误差则是绝对误差除以真实值。在大多数考试情境中，真实值是未知的，因此我们通过观察相邻两次近似值之间的变化来评估误差。
Truncation error occurs when we cut off an infinite process after a finite number of steps, e.g. stopping a Taylor series after a few terms. Rounding error arises because computers and calculators work with a fixed number of decimal places.
截断误差是指我们在有限步数后终止一个无限过程而产生的误差，例如在保留几项后就停止泰勒级数展开。舍入误差则是由于计算机和计算器只能处理固定小数位数而产生的误差。
In IB and WJEC exams, you should always give answers to a specified degree of accuracy, often stated as “3 significant figures” or “2 decimal places”, and you must avoid premature rounding in intermediate steps.
在 IB 和 WJEC 考试中，你应始终按指定精度给出答案，通常要求“3位有效数字”或“2位小数”，并且必须避免在中间步骤中过早四舍五入。
3. Bisection Method | 二分法
The bisection method is the simplest root-finding technique. It requires a continuous function f(x) and an interval [a, b] where f(a) and f(b) have opposite signs. By the Intermediate Value Theorem, there must be at least one root in that interval.
二分法是最简单的求根方法。它要求函数 f(x) 连续，并且区间 [a, b] 满足 f(a) 与 f(b) 异号。根据介值定理，该区间内至少存在一个根。
The algorithm: calculate the midpoint c = (a + b)/2. If f(c) = 0, c is the root. Otherwise, replace either a or b with c so that the sign change is preserved, and repeat until the interval is sufficiently small.
算法如下：计算中点 c = (a + b)/2。若 f(c) = 0，则 c 即为根。否则，用 c 替换 a 或 b 中使得函数值异号的那一个，重复该过程直至区间足够小。
Advantages: it always converges if the initial conditions are met. Disadvantages: it is slow (linear convergence), and finding an initial bracket can be challenging.
优点：只要满足初始条件总能收敛。缺点：收敛速度较慢（线性收敛），并且寻找初始区间有时并不容易。
You may be asked to perform two or three iterations and to state how many more iterations are needed to achieve a given tolerance. The error after n iterations is at most (b – a)/2ⁿ.
考题可能要求你执行两到三次迭代，并说明还需多少次迭代才能达到给定的容差。n 次迭代后误差最多为 (b – a)/2ⁿ。
4. Newton-Raphson Method | 牛顿-拉夫森法
The Newton-Raphson method uses the tangent line to a curve to find successively better approximations to a root. The iteration formula is:
牛顿-拉夫森法利用曲线上的切线来生成越来越精确的根的近似值。其迭代公式为：
xₙ₊₁ = xₙ – f(xₙ)/f'(xₙ)
This formula is derived by setting the x-intercept of the tangent at (xₙ, f(xₙ)) as the next approximation. The method converges quadratically if the initial guess is sufficiently close to the root and f'(root) ≠ 0.
该公式的推导基于将点 (xₙ, f(xₙ)) 处切线的 x 轴截距作为下一个近似值。若初始猜测值足够接近根且 f'(root) ≠ 0，该方法具有二次收敛性。
Common exam tasks: perform a given number of iterations, explain why the method fails when f'(x) is close to zero, and apply it to contextual problems like finding interest rates or population growth parameters.
常见考题：执行指定次数的迭代；解释当 f'(x) 接近零时该方法为何失效；并将该方法应用于情境问题，如求利率或人口增长参数等。
Warning: Newton-Raphson can diverge or cycle if the starting point is poorly chosen or if there is a stationary point near the root. Always check f'(x) is not zero at the starting value.
注意：若初始点选择不当或根附近存在驻点，牛顿-拉夫森法可能会发散或陷入循环。务必检查在起始点处 f'(x) 不为零。
5. Fixed-Point Iteration | 不动点迭代
A fixed-point iteration rearranges the equation f(x) = 0 into the form x = g(x). Starting with an initial guess x₀, the sequence xₙ₊₁ = g(xₙ) is generated. If the sequence converges, it converges to a point where x = g(x), which is a root of the original equation.
不动点迭代将方程 f(x) = 0 改写为 x = g(x) 的形式。从初始猜测值 x₀ 开始，生成序列 xₙ₊₁ = g(xₙ)。若该序列收敛，则收敛于满足 x = g(x) 的点，即原方程的根。
The convergence condition is |g'(x)| < 1 in a neighbourhood of the root. If |g'(root)| > 1, the iteration will diverge. This can be illustrated using cobweb or staircase diagrams.
收敛条件是：在根的邻域内有 |g'(x)| < 1。若 |g'(root)| > 1，迭代将发散。这一点可以用蛛网图或阶梯图来示意。
In exams, you might be asked to rewrite a given equation in a form suitable for iteration, to demonstrate convergence or divergence by calculating a few terms, or to explain why a particular rearrangement fails.
考试中，你可能需要将给定方程改写为适合迭代的形式，通过计算几项数值来演示收敛或发散，或者解释某种特定变形为何失败。
6. Trapezium Rule | 梯形法则
The trapezium rule estimates the area under a curve by dividing it into a number of trapeziums. For n strips of equal width h = (b – a)/n, the approximate integral is:
梯形法则通过将曲线下方区域划分为一系列梯形来估算面积。假设有 n 个等宽条带，宽度为 h = (b – a)/n，则近似积分为：
∫ₐᵇ f(x) dx ≈ h/2 [y₀ + 2(y₁ + y₂ + … + yₙ₋₁) + yₙ]
where yᵢ = f(xᵢ) and xᵢ = a + i h. The more strips you use, the better the approximation generally becomes.
其中 yᵢ = f(xᵢ)，xᵢ = a + i h。通常条带越多，近似效果越好。
This method is particularly useful when the function is given as a table of values or when an exact integral is impossible to find. In WJEC and IB exams, you may also have to calculate the percentage error of the approximation compared to an exact value.
当函数以表格数值形式给出或积分无法精确求出时，该方法格外有用。在 WJEC 和 IB 考试中，你可能还需要计算近似值与精确值之间的百分比误差。
The trapezium rule overestimates when the curve is concave up and underestimates when concave down, facts that can be explored in graphical interpretation questions.
当曲线凹向上时梯形法则会高估，凹向下时则会低估，这一性质常在图形解释题中考查。
7. Simpson’s Rule | 辛普森法则
Simpson’s rule gives a more accurate approximation of a definite integral by fitting quadratic curves through sets of three points. It requires an even number of strips (n must be even). The formula is:
辛普森法则通过每三个点拟合二次曲线来获得更精确的定积分近似值。它要求条带数为偶数（n 必须为偶数）。其公式为：
∫ₐᵇ f(x) dx ≈ h/3 [y₀ + 4(y₁ + y₃ + … + yₙ₋₁) + 2(y₂ + y₄ + … + yₙ₋₂) + yₙ]
The pattern of coefficients is 1, 4, 2, 4, 2, …, 4, 1. Simpson’s rule is exact for polynomials up to degree 3, giving it much higher accuracy than the trapezium rule for smooth functions.
系数模式为 1, 4, 2, 4, 2, …, 4, 1。辛普森法则对于次数不超过 3 的多项式是精确成立的，因此对于光滑函数，其精度远高于梯形法则。
Questions often ask you to apply Simpson’s rule with a given number of ordinates or to compare its accuracy with the trapezium rule. Make sure you correctly identify the number of intervals and that n is even.
题目常要求你使用给定数量的纵坐标应用辛普森法则，或比较其与梯形法则的精确度。务必正确识别区间数并确保 n 为偶数。
8. Euler’s Method | 欧拉方法
Euler’s method is used to find numerical solutions to first-order differential equations of the form dy/dx = f(x, y) with a given initial condition. Starting from (x₀, y₀), the next point is calculated using a small step size h:
欧拉方法用于求形如 dy/dx = f(x, y) 且带有给定初始条件的一阶微分方程的数值解。从点 (x₀, y₀) 出发，采用小步长 h 计算下一点：
yₙ₊₁ = yₙ + h f(xₙ, yₙ)
This process approximates the solution curve by a sequence of short line segments. The smaller the step size h, the more accurate the approximation, but the more calculations are required.
该过程通过一系列短线段来近似解曲线。步长 h 越小，近似越精确，但计算量也越大。
Typical exam questions provide a differential equation, an initial condition, and a specific step size, and you must fill in a table of x and y values, often using corrector-improver steps (like the improved Euler method) if specified.
典型考题给出一个微分方程、一个初始条件和一个特定的步长，你需要填写 x 和 y 值的表格。若题目要求，还可能涉及改进欧拉法（如预报-校正法）。
9. Convergence and Stability | 收敛性与稳定性
Convergence refers to whether an iterative method approaches the true solution as the number of iterations increases. For root-finding, we want the error to shrink steadily. Linear convergence means the error reduces by roughly a constant factor each step, while quadratic convergence means the number of correct digits doubles with each iteration (like Newton-Raphson).
收敛性指的是随着迭代次数的增加，迭代法是否趋近于真实解。在求根中，我们希望误差稳步缩小。线性收敛意味着每一步误差大致按固定比例减少，而二次收敛则意味着每一步正确位数加倍（如牛顿-拉夫森法）。
Stability is particularly important for differential equation solvers. Euler’s method can become unstable if h is too large, producing wildly oscillating values. The concept of stability regions is explored in further study but may appear in IB HL questions conceptually.
稳定性对于微分方程求解器尤为重要。若步长 h 太大，欧拉方法可能变得不稳定，产生剧烈振荡的数值。稳定域的概念会在后续学习中深入探讨，但在 IB HL 题目中可能以概念形式出现。
Exam strategies: always check the behaviour of |g'(x)| near the root for fixed-point iteration. For Euler’s method, a smaller h improves accuracy but increases rounding errors due to more steps, which is a classic trade-off.
考试策略：对于不动点迭代，始终检查在根附近的 |g'(x)| 行为。对于欧拉方法，更小的 h 可提高精度，但因步数增多会导致舍入误差累积，这是一个典型的权衡考量。
10. Common Mistakes & Exam Tips | 常见错误与考试技巧
Mistake 1: Using the wrong number of strips in the trapezium or Simpson’s rule. Remember: with n strips, there are n+1 ordinates. For Simpson’s rule, n must be even; many students lose marks by using an odd number of strips.
错误一：梯形法则或辛普森法则中条带数用错。记住：n 个条带对应 n+1 个纵坐标。辛普森法则要求 n 为偶数，许多学生因使用奇数个条带而失分。
Mistake 2: Incorrect rounding. Always work with full calculator precision throughout intermediate steps and only round the final answer. Premature rounding can cause the final answer to be outside the accepted tolerance.
错误二：舍入不当。所有中间步骤都应使用计算器的全部精度，只在最后答案处进行舍入。过早舍入可能导致最终答案超出允许误差范围。
Mistake 3: Not checking convergence conditions for fixed-point iteration. An iteration that diverges will lead to nonsensical answers; knowing how to test |g'(x)| < 1 can save time and clarify the mark scheme.
错误三：未检查不动点迭代的收敛条件。发散的迭代会导致荒谬的结果；知道如何检验 |g'(x)| < 1 可以节省时间并更好地把握评分要点。
Mistake 4: Forgetting to give the final answer in the required form. If the question asks for a root to 3 decimal places, state the full decimal, not just a truncated version. Similarly, for errors, give both the approximate value and the error bound if requested.
错误四：忘记按题目要求形式给出最终答案。如果题目要求给出保留 3 位小数的根，必须写出完整小数，而非截断值。同样，对于误差，若题目要求，应同时给出近似值和误差范围。
Exam tip: showing your iterations clearly in a table, with columns labelled xₙ, f(xₙ), f'(xₙ) etc., makes it easier to gain method marks and avoid arithmetic slip-ups.
考试技巧：用表格清晰地展示迭代过程，标注 xₙ、f(xₙ)、f'(xₙ) 等列，这样更容易拿到方法分并避免计算失误。
11. Summary and Revision Checklist | 总结与复习清单
Here is a quick revision checklist to ensure you are ready for questions on numerical methods:
以下是快速复习清单，可确保你已做好应对数值方法问题的准备：
| Topic / 主题 | Key Concept / 关键概念 | Check / 核查 |
|---|---|---|
| Bisection Method / 二分法 | Sign change, midpoint, linear convergence | ☐ |
| Newton-Raphson / 牛顿-拉夫森法 | xₙ₊₁ = xₙ – f(xₙ)/f'(xₙ), quadratic convergence | ☐ |
| Fixed-Point Iteration / 不动点迭代 | x = g(x), |g'(x)| < 1 for convergence | ☐ |
| Trapezium Rule / 梯形法则 | h/2[y₀+2(sum)+ yₙ], n strips, over/under estimate | ☐ |
| Simpson’s Rule / 辛普森法则 | h/3[1,4,2,4,…,1], n even, exact up to cubic | ☐ |
| Euler’s Method / 欧拉方法 | yₙ₊₁ = yₙ + h f(xₙ,yₙ), step size h | ☐ |
| Error Analysis / 误差分析 | Absolute, relative, truncation, rounding errors | ☐ |
Regular practice with past paper questions will help you internalise these algorithms and recognise the subtle differences between them. In addition, learning to use your calculator efficiently—especially the TABLE and RECURSION functions—can significantly speed up your work during the exam.
定期练习历年真题有助于你内化这些算法并辨别它们之间的细微差别。此外，学会高效使用计算器——特别是 TABLE 和 RECURSION 功能——可显著提高考试时的答题速度。
Numerical methods are not just a collection of formulas; they are a way of thinking about approximation, error, and the power of iterative processes. By understanding the underlying ideas, you will be able to answer even unfamiliar problems with confidence.
数值方法不仅仅是公式的集合，它们还代表着一种关于近似、误差以及迭代过程力量的思维方式。通过理解其背后思想，你将能够自信地解答哪怕是陌生的题目。
Published by TutorHao | Mathematics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
屏轩国际教育cambridge primary/secondary checkpoint, cat4, ukiset,ukcat,igcse,alevel,PAT,STEP,MAT, ibdp,ap,ssat,sat,sat2课程辅导，国外大学本科硕士研究生博士课程论文辅导