📚 The Amendment Process in Numerical Iterations | 数值迭代中的修正过程
In A-Level Mathematics, solving equations that cannot be rearranged algebraically often calls for numerical methods. At the heart of these methods lies a systematic way to improve an initial guess: each step “amends” the current estimate, pushing it closer to the true root. This iterative amendment process is not only a central topic in Edexcel’s Pure Mathematics but also a powerful illustration of how mathematicians refine approximate solutions through repetition and correction.
在A-Level数学中，许多方程无法通过代数方法直接求解，数值方法便成为关键工具。这些方法的核心在于一种系统化地改进初始猜测的方式：每一步都在“修正”当前估计值，使其逐渐逼近真实根。这一迭代修正过程不仅是爱德思纯数学中的重要专题，也生动展示了数学家如何通过重复与校正不断逼近真解。
1. Why Iterative Amendment Matters | 为何迭代修正如此重要
Exact solutions to equations like cos x = x or eˣ e expressed in closed form. Instead, we construct a sequence of approximations x₀, x₁, x₂, … uces the error systematically, turning a rough starting value into a highly accurate root.
诸如 cos x = x 或 eˣ = 3x 这类方程无法给出封闭形式的精确解。我们转而构造一个近似序列 x₀, x₁, x₂, … ，其中每一项都会修正前一项。修正机制旨在系统性地减小误差，从而将一个粗糙的初始值转化为高精度的根。
2. The General Form of an Iterative Amendment | 迭代修正的一般形式
: it must be designed so that the sequence converges to the desired root α where α = g(α).
典型的迭代格式写作 xₙ₊₁ = g(xₙ)。将前一个值 xₙ 代入修正函数 g，输出 xₙ₊₁ 即为修正后的估计值。g 的选取至关重要：它必须使序列收敛到所需的根 α，且满足 α = g(α)。
3. Simple Iteration as a Basic Amendment Strategy | 简单迭代作为基本修正策略
If we wish to solve f(x) = 0, we can rearrange it into x = g(x). For instance, from x² – ach step adjusts the current value using the information encoded in the original equation.
若要解 f(x) = 0，可将其改写为 x = g(x) 的形式。例如，由 x² – 5x + 3 = 0 可得到 x = (x² + 3)/5。从某个 x₀ 出发，反复应用修正 xₙ₊₁ = (xₙ² + 3)/5。每一步都利用原方程所蕴含的信息对当前值进行调整。
4. The Newton–Raphson Amendment Process | 牛顿–拉夫森修正过程
The Newton–Raphson formula is arguably the most celebrated amendment process in A-Level maths. Given an estimate xₙ, the next value is obtained by subtracting a correction term:
xₙ₊₁ = xₙ – f(xₙ) / f′(xₙ)
The correction –f(xₙ)/f′(xₙ) uses the slope of the function to amend the guess. When the method works, the number of correct decimal places roughly doubles at each amendment, making it extremely efficient.
牛顿–拉夫森公式堪称A-Level数学中最著名的修正过程。给定估计值 xₙ，下一个值通过减去一个修正项得到：
xₙ₊₁ = xₙ – f(xₙ) / f′(xₙ)
修正项 –f(xₙ)/f′(xₙ) 利用函数的斜率来修正猜测值。当该方法有效时，每次修正大致都能使正确的小数位数翻倍，效率极高。
5. Geometric Interpretation of the Amendment | 修正过程的几何解释
y = f(x), the point (xₙ, f(xₙ)) lies on the curve. The tangent at this point crosses the x‑ on shifts the guess horizontally from a curve-tangent intersection toward the actual root.
在 y = f(x) 的图像上，点 (xₙ, f(xₙ)) 位于曲线上。该点处的切线与 x 轴的交点即为 xₙ₊₁。因此，修正过程在局部将曲线替换为其线性近似。从视觉上看，每次迭代都将猜测值从切线与 x 轴的交点水平移向真正的根。
6. Conditions for a Reliable Amendment | 可靠修正的条件
Not every amendment process converges. For simple iteration x = g(x), the magnitude of |g′(x)| near the root must be less than 1. If |g′(α)| ≥ 1, the amendments may diverge – hes of y = g(x) and y = x with staircase or cobweb diagrams help visualise this condition.
并非所有修正过程都能收敛。对于简单迭代 x = g(x)，在根附近的 |g′(x)| 必须小于 1。若 |g′(α)| ≥ 1，修正可能发散——每一次校正都会将估计值推得更远。利用阶梯图或蛛网图绘制 y = g(x) 与 y = x，可以直观展示这一条件。
7. Error Analysis: How Much Does Each Amendment Improve? | 误差分析：每次修正改善了多大程度？
Let eₙ = xₙ – α be the error after the n‑th step. For Newton–Raphson, the relationship eₙ₊₁ ∝ eₙ² indicates quadratic convergence. For a simple iteration with 0 < |g′(α)| uickly the error shrinks: the smaller its absolute value, the more powerful the amendment.
设 eₙ = xₙ – α 为第 n 步后的误差。对牛顿–拉夫森法，关系式 eₙ₊₁ ∝ eₙ² 表明其具有二次收敛性。对满足 0 < |g′(α)| < 1 的简单迭代，则有 eₙ₊₁ ≈ g′(α) eₙ。修正因子 g′(α) 决定了误差缩小的速率：其绝对值越小，修正的效力越强。
8. Stopping Criteria: When to Cease Amending | 终止准则：何时停止修正
An amendment process must have a sensible stopping rule. Common criteria include |xₙ₊₁ – xₙ| < tolerance (e.g. 0.0001) or |f(xₙ₊₁)| < tolerance. This avoids endless cycle and signals that further amendments will not meaningfully change the answer within the required accuracy.
修正过程必须有一个合理的终止规则。常用的准则是 |xₙ₊₁ – xₙ| < 容差（例如 0.0001）或 |f(xₙ₊₁)| < 容差。这可以避免无限循环，并提示在该精度要求下，继续修正已无实质意义。
9. Choosing the Starting Value for the Amendment Loop | 为修正循环选择初值
A poor initial guess can cause divergence even for Newton– is then amended by the iterative scheme. Sometimes plotting the two functions, e.g. y = eˣ and y = 3x, gives a visual hint for the neighbourhood of the root.
即便对于牛顿–拉夫森法，糟糕的初值也可能导致发散。爱德思考题通常要求通过符号变化（f(a)f(b) < 0）确定初值，随后由迭代格式进行修正。有时绘制两个函数，如 y = eˣ 和 y = 3x，可以直观地提示根所在的邻域。
10. Amendment in Context: Comparing Methods | 修正方法之比较
| Method | Correction type | Convergence speed | Risk |
| Simple iteration | xₙ₊₁ = g(xₙ) | Linear | Divergence if |g'(α)| ≥ 1 |
| Newton–Raphson | xₙ₊₁ = xₙ – f(xₙ)/f'(xₙ) | Quadratic | Fails when f'(xₙ) ≈ 0 |
| Iteration with rearrangement | x = (x + kf(x)) etc. | Depends on k | Choosing suitable k |
Each method offers a different balance between simplicity and amendment power, and Edexcel exams expect candidates to understand these trade-offs.
每种方法都在简便性与修正能力之间达成了不同的平衡，爱德思考查要求考生理解这些权衡。
11. Real-World Example: Amending the Estimate for √2 | 实例：修正 √2 的估计值
To approximate √2, solve f(x) = x² – 2 = 0 with Newton–Raphson. Starting with x₀ = 1.5:
x₁ = 1.5 – (1.5² – 2) / (2×1.5) = 1.416666…
57. In just two amendments the estimate moves from 1.5 to 1.4142, already correct to four decimal places.
为求 √2，解 f(x) = x² – 2 = 0，使用牛顿–拉夫森法。从 x₀ = 1.5 出发：
x₁ = 1.5 – (1.5² – 2) / (2×1.5) = 1.416666…
此次修正减去大约 0.0833。下一次迭代得到 x₂ ≈ 1.4142157。仅两次修正，估计值就从 1.5 变为 1.4142，已精确到小数点后四位。
12. Pitfalls in the Amendment Process | 修正过程中的陷阱
The derivative f′(xₙ) may be zero or very small, causing an enormous correction that overshoots the root.
Oscillation can occur when the amendment overshoots and then bounces back, never settling.
Forgetting to use radian mode on a calculator when iterating trigonometric functions leads to completely wrong amendments.
导数 f′(xₙ) 可能为零或非常小，导致修正幅度极大，越过了根。
当修正越过根后反弹，可能产生振荡，永远无法稳定下来。
在迭代三角函数时忘记将计算器设为弧度模式，会导致完全错误的修正结果。
13. Linking Amendment to the Edexcel A-Level Exam | 修正与爱德思A-Level考试的关联
Edexcel papers frequently present an incomplete iteration formula and ask students to carry out several amendments, deducing the root to a specified accuracy. Questions also test the ability to derive a Newton–Raphson formula from a given f(x) and to identify whether a chosen rearrangement will produce a convergent amendment process.
爱德思试卷常给出一个不完整的迭代公式，要求学生执行数次修正，并推断出达到指定精度的根。考题也会测试从给定的 f(x) 推导牛顿–拉夫森公式的能力，以及判断所选变形是否能产生收敛的修正过程。
14. Summary: The Power of Systematic Correction | 总结：系统化校正的力量
The amendment process in numerical methods transforms a rough guess into a precise solution through repeated, structured corrections. Whether using a simple rearrangement or the sophisticated Newton–Raphson step, the core idea remains the same: measure the current error and adjust by a calculated amount. Mastering this iterative mindset not only secures marks on A-Level papers but also builds intuition for algorithms used across science and engineering.
数值方法中的修正过程通过有结构的反复校正，将粗糙猜测转化为精确解。无论是使用简单的变形还是复杂的牛顿–拉夫森步，核心思想始终如一：衡量当前误差，并按照计算出的量进行调整。掌握这种迭代思维不仅能在A-Level考试中稳拿分数，还能为理解科学与工程中广泛应用的算法建立直观认识。
Published by TutorHao | Mathematics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图