IB/AQA Mathematics: Numerical Methods Key Points | IB/AQA 数学：数值方法考点精讲

📚 IB/AQA Mathematics: Numerical Methods Key Points | IB/AQA 数学：数值方法考点精讲

Numerical methods are essential techniques for solving equations, integrating functions, and approximating solutions to differential equations when exact analytical methods are unavailable. In both IB Mathematics (Analysis & Approaches and Applications & Interpretation) and AQA A-level Mathematics, these methods appear repeatedly, not only in dedicated questions but also as tools in modelling and applied contexts. Mastering the core algorithms, error analysis, and conditions for convergence will significantly boost your exam performance. This revision guide condenses the most frequently tested numerical methods, explains their strengths and limitations, and provides practical tips for calculator use and exam technique.

数值方法是当精确解析法不可行时求解方程、积分函数以及逼近微分方程解的核心技术。在IB数学（分析与方法、应用与解释）和AQA A-level数学中，这些方法频繁出现，不仅以专门考题形式考查，还会作为建模和应用情境中的工具。掌握核心算法、误差分析以及收敛条件，将极大提升你的考试成绩。本复习指南浓缩了最常考的数值方法，解释其优缺点，并提供计算器使用和考试技巧的实用建议。

1. The Bisection Method | 二分法

The bisection method locates a root of a continuous function f(x) on an interval [a, b] where f(a) and f(b) have opposite signs. At each step, the interval is halved by evaluating the midpoint c = (a + b)/2. If f(c) = 0, c is the exact root; otherwise, the subinterval where the sign change occurs is retained. The method always converges provided f is continuous and the initial interval brackets a root, but convergence is linear and relatively slow.

二分法用于寻找连续函数f(x)在区间[a, b]上的根，且f(a)与f(b)异号。每一步通过计算中点c = (a + b)/2将区间二分；若f(c)=0，则c为精确根；否则保留发生符号改变的子区间。只要f连续且初始区间包含根，该方法必定收敛，但收敛速度为线性，相对缓慢。

After n iterations, the error bound is (b – a)/2ⁿ. This is often used to determine the number of steps needed to achieve a specified tolerance. In exam questions, you may be asked to perform two or three iterations by hand and to show the narrowing interval.

经过n次迭代后，误差界限为(b – a)/2ⁿ。这常被用来计算达到指定容差所需的步数。考试中可能要求手算进行两至三次迭代，并展示区间的缩小过程。

2. Newton-Raphson Method | 牛顿-拉弗森法

The Newton-Raphson method uses the tangent line at an approximation xₙ to find a better estimate xₙ₊₁ = xₙ – f(xₙ)/f'(xₙ). It converges quadratically once the estimate is sufficiently close to the root, making it much faster than bisection. However, it requires a differentiable function and a good initial guess; otherwise, it may diverge or oscillate, especially near stationary points or inflection points where f’ is small.

牛顿-拉弗森法利用近似点xₙ处的切线得到更好的估计值xₙ₊₁ = xₙ – f(xₙ)/f'(xₙ)。一旦估计值足够接近根，该方法为二次收敛，比二分法快得多。但要求函数可导且初始猜测良好；否则可能发散或震荡，尤其在f’很小的驻点或拐点附近。

On AQA exams, you might be given the iteration formula and asked to apply it starting from a given x₀, carrying out iterations until two successive values agree to a specified number of decimal places. You should also be able to derive the formula from a tangent equation or explain graphically how it works.

在AQA考试中，可能会给出迭代公式并要求从给定x₀开始计算，直到两个连续值在小数点后指定位数上一致。你还应能够从切线方程推导出公式，或用图形解释其原理。

3. The Secant Method | 割线法

The secant method approximates the derivative using a finite difference: xₙ₊₁ = xₙ – f(xₙ) × (xₙ – xₙ₋₁)/(f(xₙ) – f(xₙ₋₁)). It does not require an explicit derivative, which is useful when f’ is difficult to compute. Convergence is superlinear (order approximately 1.618) but still requires two initial guesses that are close enough to the root.

割线法用有限差分近似导数：xₙ₊₁ = xₙ – f(xₙ) × (xₙ – xₙ₋₁)/(f(xₙ) – f(xₙ₋₁))。它不需要显式的导数，在f’难以计算时非常实用。收敛为超线性（约1.618阶），但仍需要两个足够接近根的初始猜测。

In IB, you may encounter secant method as an alternative to Newton-Raphson in internal assessments or modelling tasks. Be aware that the method can be less stable if the two initial points are not well chosen.

在IB中，内部评估或建模任务里你可能遇到割线法作为牛顿-拉弗森法的替代。注意若两个初始点选择不当，该方法稳定性会降低。

4. Fixed-Point Iteration | 不动点迭代

Fixed-point iteration rearranges the equation f(x) = 0 into the form x = g(x). Starting from x₀, subsequent approximations are generated by xₙ₊₁ = g(xₙ). The method converges if |g'(x)| < 1 in the neighbourhood of the root; if |g'(root)| > 1, the iteration diverges. This is a key concept tested both theoretically and graphically with cobweb or staircase diagrams.

不动点迭代将方程f(x) = 0改写为x = g(x)。从x₀开始，由xₙ₊₁ = g(xₙ)生成后续近似值。若在根附近|g'(x)| < 1，迭代收敛；若|g'(根)| > 1，则发散。这是重要概念，常通过蛛网图或阶梯图从理论和图形两方面考查。

Examiners like questions that ask: “Show that the equation can be rearranged into …” and then require you to perform iterations and comment on convergence. You should also know how to choose a suitable rearrangement to ensure convergence.

考官喜欢这样的问题：“证明该方程可改写为…”然后要求你进行迭代并评论收敛性。你还应知道如何选择合适的改写形式以确保收敛。

5. Error Analysis and Stopping Criteria | 误差分析与停止准则

Understanding error types is crucial. True error = exact – approximation; relative error = true error / exact value. Since the exact value is unknown, we often use approximate error between successive iterations: |xₙ₊₁ – xₙ|. A common stopping criterion is when this difference falls below a tolerance, such as 10⁻³. But this does not guarantee the true error is equally small—the method might have stalled before reaching the root.

理解误差类型至关重要。真实误差 = 精确值 – 近似值；相对误差 = 真实误差 / 精确值。由于精确值未知，常使用连续迭代间的近似误差：|xₙ₊₁ – xₙ|。通常停止准则是该差值低于容差，如10⁻³。但这并不能保证真实误差同样小——迭代可能在到达根之前就已停滞。

For the bisection method, the error bound formula gives a guaranteed maximum error, while for Newton-Raphson, error estimates are more complicated. In calculator work, using the “ANS” button to iterate until digits stabilize is a practical time-saver, but you must still record values to the required accuracy.

对于二分法，误差界公式给出保证的最大误差，而牛顿-拉弗森法的误差估计更复杂。在计算器操作中，使用”ANS”键迭代直到数字稳定是省时实用技巧，但你仍需按要求的精度记录数值。

6. Numerical Integration: Trapezium Rule | 数值积分：梯形法则

The trapezium rule approximates ∫ₐᵇ f(x) dx using a series of trapezia. With n strips of equal width h = (b – a)/n, the formula is:∫ₐᵇ f(x) dx ≈ h/2 [f(x₀) + 2f(x₁) + 2f(x₂) + … + 2f(xₙ₋₁) + f(xₙ)]. The more strips used, the better the approximation, provided the function is not highly oscillatory.

梯形法则用一系列梯形近似∫ₐᵇ f(x) dx。将区间等分为n个宽度为h = (b – a)/n的条形，公式为：∫ₐᵇ f(x) dx ≈ h/2 [f(x₀) + 2f(x₁) + 2f(x₂) + … + 2f(xₙ₋₁) + f(xₙ)]。使用的条纹越多，近似越精确，前提是函数没有剧烈震荡。

In AQA and IB exams, you will be given a table of values and asked to apply the trapezium rule with a specific number of strips. Remember to check whether the question asks for an over- or under-estimate when the graph is concave. Drawing a small sketch can help justify your answer.

在AQA和IB考试中，会给出一个函数值表格，要求用特定数量的条纹应用梯形法则。当图形是凹的时，要记得检查题目是否要求判断高估还是低估。画一小幅草图有助于论证你的答案。

7. Simpson’s Rule | 辛普森法则

Simpson’s rule provides a more accurate integration estimate by fitting quadratic arcs through sets of three consecutive points. It requires an even number of strips (n even). The formula is:∫ₐᵇ f(x) dx ≈ h/3 [y₀ + 4(y₁ + y₃ + …) + 2(y₂ + y₄ + …) + yₙ], where yₓ = f(xₓ). The alternating coefficients 4 and 2 are easy to mix up, so double-check them.

辛普森法则通过在三相邻点之间拟合二次弧线得到更精确的积分估计。它要求条纹数为偶数（n为偶数）。公式为：∫ₐᵇ f(x) dx ≈ h/3 [y₀ + 4(y₁ + y₃ + …) + 2(y₂ + y₄ + …) + yₙ]，其中yₓ = f(xₓ)。交替的系数4和2容易混淆，务必仔细核对。

Simpson’s rule is exact for polynomials up to degree 3. Exam tasks often ask you to compare the trapezium rule and Simpson’s rule results, explaining why one is better. You may also need to state the number of strips or ordinates for a given Simpson’s rule application.

辛普森法则对三次及以下多项式是精确的。考题常常要求比较梯形法则和辛普森法则的结果，并解释为何一种更好。你还可能需要说明给定辛普森法则应用中的条纹数或纵坐标数。

8. Euler’s Method for Differential Equations | 微分方程欧拉方法

Euler’s method approximates the solution of a first-order differential equation dy/dx = f(x, y) with given initial condition y(x₀) = y₀. Using step size h, successive points are computed by:yₙ₊₁ = yₙ + h f(xₙ, yₙ) and xₙ₊₁ = xₙ + h. This tangent-line stepping is a building block for more sophisticated methods.

欧拉方法近似求解一阶微分方程dy/dx = f(x, y)及初始条件y(x₀) = y₀。使用步长h，依次计算：yₙ₊₁ = yₙ + h f(xₙ, yₙ)且xₙ₊₁ = xₙ + h。这种切线步进法是更复杂方法的基础。

In IB Applications & Interpretation and AQA, Euler’s method appears in both pure and applied contexts. You will typically fill in a table of x, y, and f(x,y). Pay attention to the required accuracy and whether the step size must be reduced to improve approximation. Diagrams showing the polygon path versus the true curve may also be assessed.

在IB应用与解释和AQA中，欧拉方法在纯数学和应用情境中都会出现。通常你需要填写x、y和f(x,y)的表格。注意要求的精度，以及是否需要减小步长以改进近似。试卷也可能考查显示折线路径与真实曲线的图形。

9. Convergence Conditions and Pitfalls | 收敛条件与易错点

Each numerical method has its own requirements. Bisection needs a sign change and continuity. Newton-Raphson needs a non-zero derivative near the root and a starting value sufficiently close. Fixed-point iteration requires |g'(x)| < 1 near the fixed point. When these conditions are violated, the method may fail to converge, oscillate, or converge to an unintended root.

每种数值方法都有其要求。二分法需要符号改变和连续性。牛顿-拉弗森法需要根附近导数非零且初值足够接近。不动点迭代需要在不动点附近|g'(x)| < 1。当这些条件被破坏时，方法可能不收敛、振荡或收敛到不想要的根。

Common exam traps include misapplying Simpson’s rule with an odd number of strips, forgetting to switch to radians when differentiating trigonometric functions for Newton-Raphson, and using inaccurate values from previous steps. Always show a check of the sign change when applying bisection, and test the derivative condition for Newton-Raphson where possible.

常见考试陷阱包括：在条纹数为奇数时错误应用辛普森法则、在对三角函数求导用于牛顿-拉弗森法时忘记使用弧度设置、以及使用前一步骤不准确的数值。应用二分法时务必展示符号改变的验证，并在可能时检验牛顿-拉弗森法的导数条件。

10. Calculator Techniques and Exam Strategy | 计算器技巧与考试策略

Modern calculators with iteration and table functions can accelerate numerical methods dramatically. For fixed-point iteration, enter the rearranged expression using ‘Ans’ and press ‘=’ repeatedly. For Euler’s method, you can set up a spreadsheet-like table, or use a programmable sequence. However, you must still record the values as required—showing intermediate results to the specified decimal places is essential for method marks.

现代计算器具有迭代和表格功能，可以大幅提高数值方法的速度。对于不动点迭代，使用’Ans’输入改写后的表达式并反复按’=’。对于欧拉方法，可以设置类似电子表格的表格，或使用可编程序列。但你必须按要求记录数值——在小数点后指定位数展示中间结果对获得方法分至关重要。

Time management is key. If asked to iterate until convergence, do three to four iterations manually, then switch to calculator for speed, but write down the values. In longer problems, note the required accuracy—often 4 or 5 decimal places. A sketch graph or sign analysis can help justify whether an estimate is an overestimate or underestimate, which frequently gains an extra mark.

时间管理是关键。若要求迭代至收敛，先手动进行三到四次迭代，然后利用计算器提速，但要写下数值。在较长的题目中，注意所需精度——通常是4到5位小数。草图或符号分析有助于说明估值是高估还是低估，这常常能赢得额外分数。

11. Summary Comparison of Methods | 方法对比总结

The following table summarises the key numerical methods you need to know:

下表总结了你需要掌握的关键数值方法：

Method	Convergence	Pros	Cons
Bisection	Linear, guaranteed	Always converges if sign change present	Slow; needs two starting points with opposite signs
Newton-Raphson	Quadratic (fast)	Very fast near root; needs only one start	May diverge; requires derivative; sensitive to initial guess
Secant	Superlinear	No derivative needed; faster than bisection	Needs two good initial points; less stable
Fixed-Point	Linear if |g’|<1	Simple to implement; easy rearrangement	Convergence depends on g'(x); may fail completely

Use this table to choose the appropriate method for a given scenario and to justify your choice in explanation questions.

利用此表在给定情境中选择合适的方法，并在解释题中为你的选择提供理由。

12. Final Exam Tips | 考前最后一击

When tackling numerical methods questions, always state the formula before substituting values. Check calculator modes (radians!), and if a question asks for a specific number of decimal places or significant figures, carry extra precision during intermediate steps to avoid rounding errors. Draw diagrams for cobweb or staircase patterns in fixed-point iteration to support your written explanation. Finally, practise writing concise but precise justifications for convergence or divergence—examiners look for words like ‘gradient’, ‘tangent’, ‘sign change’, and ‘interval’.

解答数值方法题目时，务必先写出公式再代入数值。检查计算器模式（弧度！），若题目要求特定的小数位数或有效数字，在中间步骤保留更高精度以避免舍入误差。对不动点迭代中的蛛网图或阶梯图进行绘制，以支持书面解释。最后，练习写出简洁而精确的收敛或发散理由——考官看重’斜率’、’切线’、’符号变化’、’区间’等关键词。

Consistent practice with past papers under timed conditions will build confidence. Numerical methods are highly scoring if you are methodical—write clear steps, label your iterations, and always box your final answer.

在限时条件下持续练习历年真题将建立信心。只要你条理清晰——写出清晰步骤、标注迭代过程、并框出最终答案——数值方法是极易得分的内容。

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）