📚 IB & Edexcel Maths: Numerical Methods | IB 与 Edexcel 数学：数值方法考点精讲
Numerical methods provide powerful techniques for solving equations and evaluating integrals when exact algebraic solutions are impossible or impractical. In both the IB (Analysis & Approaches and Applications & Interpretation) and Edexcel A Level Mathematics curricula, numerical methods appear as an essential bridge between pure theory and real‑world computation. Understanding these algorithms, their convergence, and their error behaviour will not only prepare you for exam questions but also give you tools used in science, engineering and finance. This article unpacks the key numerical methods tested, explains how to apply them accurately, and highlights common mistakes to avoid.
数值方法为求解无法获得精确代数解的方程和积分提供了强大的技术手段。在 IB（分析与方法、应用与解释）和 Edexcel A Level 数学课程中，数值方法连接了纯理论与实际计算。理解这些算法、它们的收敛性以及误差行为，不仅能帮你在考试中取胜，还能为你提供科学、工程和金融领域常用的工具。本文将拆解必考的数值方法，讲解如何准确应用，并指出常见错误。
1. What Are Numerical Methods? | 什么是数值方法？
Numerical methods are algorithms that produce approximate solutions to mathematical problems. Unlike exact algebraic manipulation, which yields closed‑form answers such as x = √2, numerical methods generate a sequence of improving estimates. They are particularly indispensable when dealing with transcendental equations like eˣ + x = 0 or when evaluating definite integrals that lack elementary antiderivatives, for example ∫₀¹ e^(−x²) dx.
数值方法是产生数学问题近似解的算法。与得到封闭形式答案（如 x = √2）的精确代数运算不同，数值方法生成一系列逐步改进的估计值。当处理超越方程（如 eˣ + x = 0）或计算没有初等原函数的定积分（例如 ∫₀¹ e^(−x²) dx）时，数值方法尤为重要。
Both IB and Edexcel expect you to understand when to apply a method, how to set up the iteration or formula, and how to interpret the precision of the result. You will encounter them mainly in the contexts of root finding and numerical integration.
IB 和 Edexcel 都要求你掌握何时应用某种方法、如何构造迭代或公式，以及如何解读结果的精度。考试中主要涉及求根和数值积分两大类。
2. Locating Roots: Change of Sign | 定位根：符号变化
If a continuous function f(x) changes sign over an interval [a, b], i.e. f(a) × f(b) < 0, then by the Intermediate Value Theorem there exists at least one root r ∈ (a, b) such that f(r) = 0. This simple principle is the foundation of most root‑finding techniques.
若连续函数 f(x) 在区间 [a, b] 上符号改变，即 f(a) × f(b) < 0，根据介值定理，在 (a, b) 内至少存在一个根 r 使 f(r) = 0。这一简单原理是大多数求根方法的基础。
Examiners often ask you to verify a root exists in a given interval by evaluating f at the endpoints and showing the sign change. Always state that f is continuous over the interval and conclude that a root lies within.
考官常要求通过计算端点函数值并展示符号变化来验证给定区间内存在根。一定要说明函数在该区间连续，并得出存在根的结论。
3. The Bisection Method | 二分法
The bisection method repeatedly halves the interval [a, b] and selects the subinterval where the sign change occurs. Starting with a₁ = a and b₁ = b, calculate the midpoint m = (aₙ + bₙ)/2. If f(m) has the same sign as f(aₙ), replace aₙ with m; otherwise replace bₙ with m. The root is bracketed within a shrinking interval, guaranteeing linear convergence — roughly one additional decimal place every three or four iterations.
二分法反复将区间 [a, b] 减半，并选择发生符号变化的子区间。从 a₁ = a、b₁ = b 开始，计算中点 m = (aₙ + bₙ)/2。若 f(m) 与 f(aₙ) 同号，则用 m 替换 aₙ；否则用 m 替换 bₙ。根始终被框定在逐渐缩小的区间内，保证了线性收敛——大约每迭代三到四次，精确到一位额外小数。
Exam tip: When approaching a required accuracy, say 10⁻³, stop when the interval width (bₙ − aₙ) is less than twice the tolerance, and take the midpoint as the estimate.
考试建议：当需达到 10⁻³ 精度时，当区间宽度 (bₙ − aₙ) 小于两倍容差时停止，并取中点作为估计值。
4. Newton‑Raphson Method | 牛顿‑拉夫森法
The Newton‑Raphson method uses the tangent line at an initial guess x₀ to rapidly approach a root. The iterative formula is:
xₙ₊₁ = xₙ − f(xₙ) / f'(xₙ)
牛顿‑拉夫森法利用初始猜测值 x₀ 处的切线快速逼近根。迭代公式为：
xₙ₊₁ = xₙ − f(xₙ) / f'(xₙ)
This method exhibits quadratic convergence near a simple root — the number of correct digits roughly doubles with each iteration. However, it may fail if f'(xₙ) ≈ 0, if the initial guess is far from the root, or if the function oscillates. In the exam, you need to differentiate accurately and carry out successive iterations, often recording your answers to a specified number of significant figures.
该方法在单根附近具有二次收敛性——每迭代一次，正确位数大约翻倍。然而，若 f'(xₙ) ≈ 0、初始猜测离根太远或函数振荡，则可能失败。考试中需准确求导并进行迭代，常需按指定有效数字记录答案。
For example, to solve 2x³ − 7x − 4 = 0 starting with x₀ = 2, compute f(2) = 3, f'(2) = 17, giving x₁ = 2 − 3/17 ≈ 1.8235. Continue until the desired precision is reached.
例如，求解 2x³ − 7x − 4 = 0，从 x₀ = 2 开始，计算 f(2) = 3，f'(2) = 17，得 x₁ = 2 − 3/17 ≈ 1.8235。继续迭代直至达到所需精度。
5. Fixed‑Point Iteration | 不动点迭代法
A rearrangement of f(x) = 0 into the form x = g(x) leads to the iteration xₙ₊₁ = g(xₙ). If |g'(x)| < 1 near the root, the iteration converges; this is the condition for a "cobweb" or "staircase" diagram. The method is simple but often slower than Newton‑Raphson.
将 f(x) = 0 改写为 x = g(x) 的形式，即可得到迭代公式 xₙ₊₁ = g(xₙ)。若在根附近有 |g'(x)| < 1，迭代收敛；这是出现“蛛网”或“阶梯”图的判别条件。该方法简单，但通常比牛顿法慢。
In exam problems, you might be given g(x) and asked to perform a few iterations or to investigate convergence by evaluating g'(x) near the initial guess. Always verify that the gradient condition holds if you are required to justify convergence.
考试中可能给定 g(x)，要求进行几次迭代，或通过在初始猜测附近求 g'(x) 来考察收敛性。若需证明收敛，务必验证梯度条件成立。
6. Numerical Integration: Trapezium Rule | 数值积分：梯形法
When an integral cannot be evaluated exactly, the trapezium rule approximates the area under y = f(x) by dividing the interval [a, b] into n strips of equal width h = (b − a)/n. The approximate integral is:
∫ₐᵇ f(x) dx ≈ h/2 [y₀ + 2(y₁ + y₂ + … + yₙ₋₁) + yₙ]
当积分无法精确计算时，梯形法通过将区间 [a, b] 等分为 n 个宽度 h = (b − a)/n 的条带，来逼近 y = f(x) 下的面积。近似积分公式为：
∫ₐᵇ f(x) dx ≈ h/2 [y₀ + 2(y₁ + y₂ + … + yₙ₋₁) + yₙ]
The error is of order h² for a single strip but accumulates to roughly proportional to 1/n² overall. The rule overestimates when the curve is concave down, otherwise underestimates. Both IB and Edexcel ask you to apply the formula, estimate errors, and comment on how doubling n improves accuracy.
单个条带的误差是 h² 量级，总体误差大约正比于 1/n²。曲线下凹时梯形法会高估，反之低估。IB 与 Edexcel 均要求应用公式、估计误差，并讨论 n 加倍对精度的改善。
7. Simpson’s Rule | 辛普森法
Simpson’s rule fits quadratic curves through sets of three points, requiring an even number of strips (n even). With h = (b − a)/n, the formula is:
∫ₐᵇ f(x) dx ≈ h/3 [y₀ + 4(y₁ + y₃ + …) + 2(y₂ + y₄ + …) + yₙ]
辛普森法通过每组三个点拟合二次曲线，要求条带数 n 为偶数。取 h = (b − a)/n，公式为：
∫ₐᵇ f(x) dx ≈ h/3 [y₀ + 4(y₁ + y₃ + …) + 2(y₂ + y₄ + …) + yₙ]
This method is remarkably accurate for polynomials up to degree three and gives an error proportional to 1/n⁴. In exams, you must carefully count the ordinates and multiply by the correct coefficients (1, 4, 2, 4, … , 2, 4, 1). The IB syllabus explicitly includes Simpson’s rule, while Edexcel may present it as an extension or in further maths options.
该方法对三次及以下多项式异常精确，误差与 1/n⁴ 成正比。考试中须仔细计数纵坐标，并乘以正确系数（1, 4, 2, 4, … , 2, 4, 1）。IB 课程明确包含辛普森法，Edexcel 可能在进阶纯数中作为拓展出现。
8. Error Analysis and Estimation | 误差分析与估计
Numerical answers are meaningless without a statement of accuracy. For root finding, the absolute error |xₙ − r| can be bounded by the interval width (bisection) or by the difference between successive iterates (Newton‑Raphson and fixed‑point). For integration, the error is often estimated by comparing two approximations with different step sizes, e.g. using the difference between trapezium rule and Simpson’s rule or halving h.
数值答案若不注明精度便毫无意义。求根时，绝对误差 |xₙ − r| 可由区间宽度（二分法）或连续迭代之差（牛顿法与不动点法）界定。积分时，误差通常通过比较不同步长的两次近似值来估计，比如利用梯形法与辛普森法之间的差异，或将 h 减半。
The following table summarizes error behaviour:
| Method / 方法 | Convergence / 收敛速度 | Error estimate / 误差估计 |
|---|---|---|
| Bisection / 二分法 | Linear / 线性 | (b − a)/2ⁿ |
| Newton‑Raphson / 牛顿法 | Quadratic / 二次 | |xₙ₊₁ − xₙ| < tolerance |
| Trapezium Rule / 梯形法 | O(1/n²) | Compare with n and 2n strips |
| Simpson’s Rule / 辛普森法 | O(1/n⁴) | Compare with trapezium or halve h |
Examiners value your ability to interpret the significance of these error bounds rather than simply stating a number.
考官看重你解读误差界限意义的能力，而非仅仅报出数字。
9. Common Exam Pitfalls | 常见考试陷阱
Misapplying the change of sign test: A sign change guarantees a root only if the function is continuous. Discontinuities like vertical asymptotes can produce false positives. Always check continuity.
误用符号变化检验：只有函数连续时符号变化才能保证有根。垂直渐近线等间断点可能产生假阳性。务必检查连续性。
Newton‑Raphson divergence: If the first derivative is near zero or the guess is poor, the next iterate can shoot far away. In such cases, the method may fail; recognize when divergence occurs.
牛顿法发散：若一阶导数为零附近或猜测糟糕，下一个迭代值可能飞远。此时方法可能失效；要能识别发散情况。
Rounding errors: Carry calculations with more significant figures than required and round only at the final answer. Premature rounding can cascade into large inaccuracies.
舍入误差：计算时保留比要求更多的有效数字，仅在最终答案处舍入。过早舍入会累积成较大的不准确。
Confusing ordinates in Simpson’s rule: Forgetting the 4 and 2 multipliers or mislabelling y₀, y₁, … is a common arithmetic slip. Write out the ordinates explicitly before summing.
辛普森法纵坐标混淆：忘记 4 和 2 的乘数或标错 y₀, y₁, … 是常见算术错误。在求和前先明确写出纵坐标。
10. Worked Exam‑style Examples | 考试风格例题详解
Example 1 – Root finding: The equation 3ˣ − x² = 0 has a root in [1, 2]. Use Newton‑Raphson with x₀ = 1.5 to find the root correct to 3 decimal places. Let f(x) = 3ˣ − x², so f'(x) = 3ˣ ln 3 − 2x. x₁ = 1.5 − (3^1.5 − 2.25)/(3^1.5 ln 3 − 3) ≈ 1.400. Continue until convergence: x ≈ 1.445.
例题 1 – 求根：方程 3ˣ − x² = 0 在 [1, 2] 内有根。使用牛顿‑拉夫森法，x₀ = 1.5，求根至 3 位小数。令 f(x) = 3ˣ − x²，f'(x) = 3ˣ ln 3 − 2x。x₁ = 1.5 − (3^1.5 − 2.25)/(3^1.5 ln 3 − 3) ≈ 1.400。继续迭代直至收敛：x ≈ 1.445。
Example 2 – Numerical integration: Estimate ∫₀² ln(1 + eˣ) dx using the trapezium rule with n = 4 strips. h = 0.5. x₀ = 0, y₀ = ln 2 ≈ 0.6931; x₁ = 0.5, y₁ = ln(1 + e⁰·⁵) ≈ 1.3741; x₂ = 1, y₂ ≈ 1.8773; x₃ = 1.5, y₃ ≈ 2.3588; x₄ = 2, y₄ ≈ 2.8322. Approx = 0.5/2 × [0.6931 + 2(1.3741+1.8773+2.3588) + 2.8322] ≈ 3.488.
例题 2 – 数值积分：使用 n = 4 条带的梯形法估计 ∫₀² ln(1 + eˣ) dx。h = 0.5。x₀ = 0, y₀ = ln 2 ≈ 0.6931；x₁ = 0.5, y₁ ≈ 1.3741；x₂ = 1, y₂ ≈ 1.8773；x₃ = 1.5, y₃ ≈ 2.3588；x₄ = 2, y₄ ≈ 2.8322。近似 = 0.5/2 × [0.6931 + 2(1.3741+1.8773+2.3588) + 2.8322] ≈ 3.488。
Always state the number of strips, the strip width, and show each evaluation. Examiners reward clear working as much as the final number.
务必标明条带数、条带宽度，并展示每步求值。考官对清晰过程的评分不下于最终结果。
11. IB vs Edexcel: Syllabus Nuances | IB 与 Edexcel：考纲差异
While the core techniques overlap, there are slight emphases to note. IB Analysis & Approaches (SL and HL) includes the bisection method, Newton‑Raphson, trapezium rule, and error analysis within the “Number and Algebra” and “Calculus” strands. HL students may also encounter order of convergence and Simpson’s rule. IB Applications & Interpretation places stronger emphasis on using technology (GDC) but still requires manual iterations and interpretation. Edexcel A Level Pure Mathematics 3 covers change of sign, iteration, Newton‑Raphson, and trapezium rule; Simpson’s rule appears in Further Mathematics options. Edexcel often asks for graphical illustration of staircase/cobweb diagrams and uses structured questions that build up from locating a root to applying an iteration and commenting on accuracy.
核心技巧虽有重叠，但侧重点有细微差别。IB 分析与方法的 SL 和 HL 在“数与代数”及“微积分”部分包含二分法、牛顿‑拉夫森法、梯形法和误差分析；HL 还涉及收敛阶和辛普森法。IB 应用与解释更强调使用图形计算器，但仍要求手动迭代和解释。Edexcel A Level 纯数 3 涵盖符号变化、迭代、牛顿‑拉夫森法和梯形法；辛普森法出现于进阶纯数选项。Edexcel 常要求图示阶梯/蛛网图，采用结构化问题，从定位根逐步到应用迭代并评价精度。
12. Final Tips for Mastery | 掌握数值方法的终极建议
Numerical methods reward systematic, well‑documented working. Whatever the exam board, practise setting out iterations in clear tables with columns for n, xₙ, f(xₙ), f'(xₙ), and xₙ₊₁. For integration, organise your ordinates before plugging into the formula. Always relate your final answer back to the context — whether it is “the root correct to 3 decimal places” or “the approximate area, which is an overestimate because…”
数值方法青睐系统化、有清晰记录的过程。不论哪个考试局，都要练习用清晰的表格列示迭代，包含 n、xₙ、f(xₙ)、f'(xₙ) 和 xₙ₊₁。对于积分，先将纵坐标整理好再代入公式。最终答案务必联系实际情境——不论是“精确至 3 位小数的根”还是“近似面积，由于……所以是高估值”。
By understanding the underlying geometry and the conditions for convergence, you can answer not just computational tasks but also conceptual questions about why a method might fail or how to improve accuracy. Numerical methods are a true integration of algebra, functions, and calculus — master them, and you add a versatile set of tools to your mathematical toolkit.
通过理解其背后的几何意义和收敛条件，你不仅能解答计算题，也能回答为什么某种方法可能失效或如何提高精度等概念性问题。数值方法是代数、函数与微积分的真正融合——掌握它们，你的数学工具箱将再添一套多用法宝。
Published by TutorHao | Numerical Methods 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