📚 A-Level CIE Mathematics: Numerical Methods Key Points Review | A-Level CIE 数学：数值方法 考点精讲
In pure mathematics, many equations and definite integrals cannot be solved analytically using elementary functions. Numerical methods provide powerful tools to approximate roots of equations and evaluate integrals to any desired accuracy. For CIE A-Level Mathematics (9709), the key techniques are locating roots by sign change, the bisection method, fixed-point iteration, the Newton-Raphson method, and the trapezium rule for numerical integration. This article presents a thorough revision of these topics, focusing on exam-relevant concepts, formulas, error analysis, and convergence conditions.
在纯数学中，许多方程和定积分无法用初等函数解析求解。数值方法为我们提供了强大的工具，能够以任意精度逼近方程的根和计算积分。针对 CIE A-Level 数学（9709），重点数值方法包括：通过符号变化定位根、二分法、不动点迭代、牛顿-拉夫森方法，以及用于数值积分的梯形法则。本文对这些考点进行了系统梳理，重点讲解核心概念、公式、误差分析以及迭代收敛条件，帮助考生高效备考。
1. Introduction to Numerical Methods | 数值方法简介
When an equation f(x) = 0 cannot be solved exactly, or when the antiderivative of a function is not expressible in elementary terms, numerical approximations become essential. In A-Level, you are expected to apply iterative and bracketing methods to locate roots, and to use the trapezium rule to estimate definite integrals. Understanding the strengths, limitations, and error behaviour of each method is crucial for both accuracy and exam technique.
当方程 f(x) = 0 无法精确求解，或者被积函数的原函数不能用初等函数表示时，数值近似就变得必不可少。在 A-Level 考试中，要求你能够应用迭代法和区间夹逼法来寻找根，并使用梯形法则估算定积分。深入理解每种方法的优点、局限性和误差特性，对于确保计算精度和掌握解题技巧都至关重要。
2. Locating Roots by Sign Change | 用符号变化定位根
If a continuous function f(x) changes sign between x = a and x = b, i.e. f(a) × f(b) < 0, then there must be at least one root in the interval (a, b). This is a direct consequence of the Intermediate Value Theorem. Always verify continuity first – discontinuities invalidate the argument. In an exam, you should evaluate f(a) and f(b) and show they have opposite signs.
如果连续函数 f(x) 在 x = a 和 x = b 处的函数值符号相反，即 f(a) × f(b) < 0，那么在区间 (a, b) 内至少存在一个根。这是介值定理的直接推论。务必先确认函数是否连续——间断点会使这一推理失效。考试中，你需要计算 f(a) 和 f(b)，并展示二者异号。
This sign-change principle is the foundation of the bisection method and is often used to determine a suitable starting interval for iterative methods. Remember to write a conclusion like “A root lies between a and b since f is continuous and f(a) and f(b) have opposite signs.”
这种符号变化原理是二分法的基础，也常用来为迭代法确定合适的初始区间。记得写出类似“因为 f 连续且 f(a) 与 f(b) 异号，所以在 a 和 b 之间存在一个根”的结论。
3. Bisection Method (Interval Halving) | 二分法（区间分半法）
The bisection method systematically narrows down the interval containing a root. Given an initial interval [a₀, b₀] with f(a₀)f(b₀) < 0, the midpoint c = (a + b)/2 is evaluated. Then check the sign of f(c). If f(a)f(c) < 0, the root lies in [a, c]; otherwise it lies in [c, b]. The interval length halves each step, guaranteeing convergence, albeit relatively slowly.
二分法逐步缩小包含根的区间。给定初始区间 [a₀, b₀] 且满足 f(a₀)f(b₀) < 0，计算中点 c = (a + b)/2，然后检查 f(c) 的符号。若 f(a)f(c) < 0，则根在 [a, c] 中；否则根在 [c, b] 中。每步区间长度减半，保证了方法的收敛性，但收敛速度相对较慢。
The error after n iterations is at most (b₀ – a₀)/2ⁿ. To achieve a specified accuracy ε, you need n ≥ log₂((b₀ – a₀)/ε). Examiners may ask you to perform two or three iterations and state the final interval and its width.
经过 n 次迭代后，误差最大为 (b₀ – a₀)/2ⁿ。为了达到指定的准确度 ε，需要 n ≥ log₂((b₀ – a₀)/ε)。考官可能会让你进行两到三次二分迭代，并要求给出最终的区间及其宽度。
c = (a + b) / 2
The midpoint is the new estimate. Because the interval is always halved, the method is robust and will never fail on a continuous function with a sign change. However, it does not use the actual function shape, making it slower than Newton-Raphson when precision is high.
中点是新的近似根。由于区间总是被对半分，该方法非常稳健，对于有符号变化的连续函数绝不会失败。但是，它没有利用函数的实际形状，因此在需要高精度时比牛顿-拉夫森方法慢。
4. Iterative Methods and Rearrangement | 迭代法与方程重排
An iterative method involves rewriting f(x) = 0 into the form x = g(x). Starting with an initial guess x₀, a sequence is generated via xₙ₊₁ = g(xₙ). If the sequence converges to a limit α, then α = g(α), so α is a solution of the original equation. The success of iteration hinges on the choice of rearrangement and the initial value.
迭代法需要将 f(x) = 0 改写为 x = g(x) 的形式。给定初始猜测值 x₀，通过 xₙ₊₁ = g(xₙ) 生成迭代序列。如果序列收敛到极限 α，则 α = g(α)，从而 α 是原方程的解。迭代的成败取决于重排式子的选择和初始值。
xₙ₊₁ = g(xₙ)
Common exam tasks: (i) show that an equation can be rearranged into a given iterative form, (ii) use a given x₀ to find x₁, x₂, x₃, (iii) determine whether the iteration is converging by checking if values are getting closer, (iv) demonstrate that a specific arrangement leads to divergence. The iterative formula must be used exactly as specified, and values should be given to the required accuracy at each step.
常见的考试任务包括：(i) 证明某方程可以重排为给定的迭代式；(ii) 使用给定的 x₀ 计算 x₁, x₂, x₃；(iii) 通过观察值是否靠近来判断迭代是否收敛；(iv) 说明某种特定重排会导致发散。题目给出的迭代公式必须原样使用，每一步的结果都要保留到规定的精度。
5. Newton-Raphson Method | 牛顿-拉夫森方法
The Newton-Raphson method is a powerful iterative technique that uses tangents to approximate the root. Starting from an initial estimate xₙ, the next estimate is the x-intercept of the tangent line at (xₙ, f(xₙ)). The iteration formula is derived from the tangent equation, yielding a quadratic convergence rate when the method works.
牛顿-拉夫森方法是一种强大的迭代技术，利用切线来逼近方程的根。从初始近似值 xₙ 出发，下一个近似值即为曲线在 (xₙ, f(xₙ)) 处的切线与 x 轴的交点。该迭代公式由切线方程导出，当方法成功时，可达到二次收敛速度。
xₙ₊₁ = xₙ – f(xₙ) / f'(xₙ)
The derivation: tangent at xₙ is y – f(xₙ) = f'(xₙ)(x – xₙ). Setting y = 0 gives 0 – f(xₙ) = f'(xₙ)(x – xₙ) → x = xₙ – f(xₙ)/f'(xₙ). This becomes the iterative formula. It requires that f'(xₙ) ≠ 0; otherwise the method fails.
推导过程：在 xₙ 处的切线方程为 y – f(xₙ) = f'(xₙ)(x – xₙ)。令 y = 0 得 0 – f(xₙ) = f'(xₙ)(x – xₙ) → x = xₙ – f(xₙ)/f'(xₙ)，这就是迭代公式。该方法要求 f'(xₙ) ≠ 0，否则方法失效。
In CIE exams, you may be asked to apply the Newton-Raphson method once or twice to refine a root. Choose an initial guess close to the root, and ensure the derivative is straightforward to compute. The method can fail if the initial guess is near a stationary point or an inflection point with a very small derivative, causing the iteration to diverge.
在 CIE 考试中，你可能需要应用牛顿法一至两次以改进根的精度。初始猜测值应选在根附近，并确保导数易于计算。如果初始点靠近驻点或导数很小的拐点，该方法可能会失效并导致发散。
6. Convergence of Iterative Methods | 迭代法的收敛性
For a rearrangement x = g(x), the iteration xₙ₊₁ = g(xₙ) converges to α if |g'(x)| < 1 in an interval containing α and the initial guess. This is the essential convergence condition. If |g'(α)| > 1, the iteration will likely diverge. In an exam, you might be asked to explain why a particular rearrangement fails, or to check that |g'(x)| < 1 near a given root.
对于重排式 x = g(x)，迭代 xₙ₊₁ = g(xₙ) 收敛到 α 的条件是：在包含 α 和初始猜测值的区间内，|g'(x)| < 1。这是基本的收敛条件。如果 |g'(α)| > 1，迭代很可能发散。考试中可能会要求你解释为什么某种重排会失败，或验证在给定根附近 |g'(x)| < 1。
For Newton-Raphson, convergence is generally very rapid if the initial guess is sufficiently close and f'(α) ≠ 0. The condition for convergence is more subtle but still relies on g(x) = x – f(x)/f'(x) and the behaviour of its derivative near the root. However, CIE focuses more on applying the formula and analysing failures, rather than rigorous convergence proofs.
对于牛顿-拉夫森方法，如果初始猜测足够接近且 f'(α) ≠ 0，收敛通常非常快。收敛条件更为复杂，但依旧依赖于 g(x) = x – f(x)/f'(x) 及其在根附近的导数特性。不过，CIE 更注重公式的应用和失败原因的分析，而不是严格的收敛性证明。
|g'(α)| < 1 ⇒ convergence
If you suspect divergence, show that |g'(x₀)| > 1 or that the iterates are moving further apart. Always refer to the gradient condition in your justification.
如果你怀疑发散，可以证明 |g'(x₀)| > 1 或迭代值在逐渐远离。在阐述理由时一定要提及梯度条件。
7. Numerical Integration – Trapezium Rule | 数值积分——梯形法则
The trapezium rule approximates the area under a curve y = f(x) from x = a to x = b by dividing the interval into n strips of equal width h = (b – a)/n, and replacing each strip with a trapezium. The area of each trapezium is the average of the two parallel sides times the width. The combined estimate is a weighted sum of the ordinates.
梯形法则将曲线 y = f(x) 下从 x = a 到 x = b 的面积近似为若干窄梯形的面积之和。将区间等分为 n 个宽度为 h = (b – a)/n 的小条，每个小条用梯形代替。每个梯形的面积等于两条平行边（函数值）的平均值乘以宽度。整体的积分估计值就是一系列纵坐标的加权和。
∫ₐᵇ f(x) dx ≈ ½h [y₀ + 2(y₁ + y₂ + … + yₙ₋₁) + yₙ]
Here yᵢ = f(a + i h). The formula uses ‘half the first and last ordinates plus all the intermediate ordinates’ all multiplied by h. In words: h/2 × (first + last + 2 × sum of the rest). Always use a table of x and y values to avoid arithmetic errors.
其中 yᵢ = f(a + i h)。该公式可表述为“第一个和最后一个纵坐标的一半，加上中间所有纵坐标，再乘以 h”。口头记忆法：h/2 × (首项 + 末项 + 2 × 其余各项之和)。应试时最好列一个 x 和 y 的数据表，以减少算术错误。
CIE often asks you to use a specified number of strips (e.g., n = 4) and to state whether the estimate is an overestimate or underestimate. This is determined by the concavity of the curve: if f”(x) > 0 (convex), the trapezia overestimate; if f”(x) < 0 (concave), they underestimate. Checking the sign of the second derivative over the interval confirms the direction of the error.
CIE 常要求用指定的条数（例如 n = 4）进行计算，并判断结果是高估还是低估。这取决于曲线的凹向：若 f”(x) > 0（下凸），梯形会高估；若 f”(x) < 0（上凸），则低估。通过检查区间上二阶导数的符号，即可确认误差方向。
8. Error Analysis and Bounds | 误差分析与界限
For the trapezium rule, the error E is bounded by |E| ≤ (b – a)³ / (12 n²) × M₂, where M₂ is the maximum value of |f”(x)| on [a, b]. This formula is useful to determine the number of strips needed for a given tolerance. However, CIE questions typically focus on practical error estimation rather than rigorous bounds, e.g., by comparing with the exact value or improving the approximation by doubling n.
对于梯形法则，误差 E 的界限为 |E| ≤ (b – a)³ / (12 n²) × M₂，其中 M₂ 是 |f”(x)| 在 [a, b] 上的最大值。该公式可用于确定满足给定精度所需的条数。不过，CIE 题目通常更注重实际误差的估算，而不是严格的误差界限，例如通过与精确值比较，或通过加倍 n 来改进近似值。
With bisection, the error bound is half the interval width: |root − c| ≤ (b – a)/2ⁿ⁺¹ after n iterations. For Newton-Raphson, error decreases approximately quadratically, meaning the number of correct decimal places roughly doubles each step once close to the root. When working with iterative methods, always round carefully and keep enough decimal places to observe convergence or oscillation.
二分法的误差界限是区间宽度的一半：经过 n 次迭代后，|根 − c| ≤ (b – a)/2ⁿ⁺¹。对于牛顿-拉夫森方法，误差大约以二次速度减小，即一旦接近根，每步正确的有效位数大致翻倍。使用迭代法时，始终要谨慎舍入，并保留足够的小数位数以观察收敛或振荡的趋势。
9. Practical Tips for Exam Questions | 考试解题实用技巧
Always show all substitutions and intermediate working. For the trapezium rule, build a table of x and y values, and then apply the formula. For iteration, record each xₙ to at least 5 decimal places, even if the final answer requires 3. When asked to demonstrate that a root is correct to a specified accuracy, check that f(x) changes sign over an interval smaller than the required tolerance – for example, for a root to 2 decimal places, test an interval of width 0.01.
始终展示所有的代换和中间步骤。对于梯形法则，列出 x 和 y 的数值表，再套用公式。对于迭代法，每个 xₙ 至少保留 5 位小数，即使最终答案只要求 3 位。当题目要求证明根的精度达到某一位时，应检查 f(x) 在一个比要求精度更小的区间内发生符号变化——例如，要证明根精确到两位小数，可以测试宽度为 0.01 的区间。
For rearrangements, double-check that the iterative form matches the given equation exactly. A common pitfall is misplacing a sign or forgetting to isolate x. If the iteration diverges, sketch a cobweb or staircase diagram to illustrate, and refer to the magnitude of g'(x). When determining over/underestimation in numerical integration, calculate f”(x) and note its sign over the full interval – not just at one point.
对于重排题，要再次确认迭代式与给定方程完全一致。常见错误是符号放错位置，或忘记正确分离出 x。如果迭代发散，可以画出蛛网图或阶梯图加以说明，并提及 g'(x) 的大小。在判断数值积分是高估还是低估时，应计算 f”(x) 并观察其在全区间上的符号，而非仅看某一点。
10. Summary and Comparison of Methods | 各方法总结与比较
Each numerical method has distinct characteristics. Bisection is guaranteed to converge for continuous functions with a sign change but is slow. Fixed-point iteration can be fast if |g'(x)| < 1, but requires a suitable rearrangement and may diverge. Newton-Raphson converges very quickly near the root, but needs the derivative and can fail if the initial guess is poor or f'(x) is zero. The trapezium rule gives good integral estimates with easy computation and a predictable error direction based on concavity.
每种数值方法都有其鲜明的特点。二分法对于有符号变化的连续函数保证收敛，但速度较慢。不动点迭代在 |g'(x)| < 1 时可以很快，但需要合适的重排，且可能发散。牛顿-拉夫森方法在根附近收敛极快，但需要导数，并且若初始猜测不佳或 f'(x) 为零则可能失败。梯形法则计算简便，同时可以根据凹凸性预判误差方向，给出较好的积分估计。
| Method | Pros | Cons |
|---|---|---|
| Bisection | Always converges if root bracketed | Slow, requires continuous function |
| Iteration (x=g(x)) | Simple, can be fast | May diverge; need |g’|<1 |
| Newton-Raphson | Quadratic convergence | Requires derivative; can fail near stationary points |
| Trapezium rule | Easy to compute; known error sign | Approximation only; accuracy improves with more strips |
方法优点和缺点总结表（英文）：Bisection 总是收敛如果包含根区间；迭代法简单可能快速；牛顿法二次收敛；梯形法则易于计算已知误差符号。
By understanding these methods and practising past paper questions, you can confidently tackle any numerical methods problem in the CIE A-Level Mathematics examination. Remember to carefully check sign changes, stability of iterations, and concavity for integration. Good luck!
深入理解这些方法，并结合往年真题进行练习，你就能自信地应对 CIE A-Level 数学考试中的任何数值方法问题。牢记仔细检查符号变化、迭代的稳定性以及积分的凹凸性。祝你取得好成绩！
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