Taylor Series | 泰勒级数

📚 Taylor Series | 泰勒级数

Taylor series is a powerful tool that expresses a function as an infinite sum of terms calculated from the values of its derivatives at a single point. Although it is not part of the standard GCSE AQA Mathematics syllabus, understanding the idea of polynomial approximation—especially linear approximation—can deepen your insight into gradients, rates of change, and estimation, all of which feature in GCSE. This article breaks down Taylor series from the ground up, connecting key concepts to familiar GCSE topics like straight-line graphs and percentage error.

泰勒级数是一种强大的工具，它通过函数在某一点的各阶导数值，将函数表达为无穷多项之和。尽管它不在 GCSE AQA 数学核心大纲中，但理解多项式近似（尤其是线性近似）的思想，可以帮助你更深入地理解梯度、变化率和估值，而这些正是 GCSE 会涉及的内容。本文从零开始拆解泰勒级数，并将关键概念与你熟悉的直线图和百分误差等 GCSE 主题联系起来。

1. What is a Taylor Series? | 什么是泰勒级数？

A Taylor series takes a smooth function and approximates it near a chosen point using a polynomial. The more terms you include, the closer the polynomial matches the function. The idea is that if you know everything about the function at one point—its value, slope, curvature, and higher-order rates of change—you can rebuild the function locally.

泰勒级数选取一个光滑函数，在选定点附近用一个多项式来逼近它。包含的项数越多，多项式就越贴近原函数。其核心思想是：如果你在某一点知道了函数的所有信息——函数值、斜率、弯曲程度以及更高阶的变化率——你就能局部地重建这个函数。

The simplest Taylor approximation is a constant: just the function value at the point. Adding a linear term gives the tangent line, which already captures the instantaneous rate of change. Quadratic, cubic, and higher terms then refine the fit by accounting for curvature and beyond.

最简单的泰勒近似就是一个常数：即该点处的函数值。加上一次项就得到了切线，它已经抓住了函数在该点的瞬时变化率。二次项、三次项以及更高次项则通过考虑曲率等因素进一步优化拟合效果。

2. Constant Approximation: The Zero-Order Term | 常数近似：零阶项

The crudest approximation of f(x) near x = a is simply f(a). This is called the zero-order Taylor polynomial. It is useful only when x is extremely close to a, and even then it ignores any change in the function. For example, if you approximate f(x) = x² near x = 2 by f(2) = 4, the estimate for 2.1² is 4, but the real value is 4.41—a large error.

在 x = a 附近对 f(x) 最粗糙的近似就是直接取 f(a)。这被称为零阶泰勒多项式。只有当 x 极其接近 a 时它才勉强有用，即便如此它也完全忽略了函数的变化。例如，如果你在 x = 2 附近用 f(2) = 4 来近似 f(x) = x²，那么对 2.1² 的估值就是 4，而真实值是 4.41，误差很大。

In GCSE terms, this is like estimating a value from a table but assuming no change—clearly not good enough when a graph shows a clear trend. A constant approximation gives no information about the direction or steepness of the graph.

用 GCSE 的思维来看，这就像从表格中估算数值却假设完全没有变化——当图表显示出明显趋势时，这样做显然是不够的。常数近似无法提供关于图像方向或陡峭程度的任何信息。

3. Linear Approximation and Tangents | 线性近似与切线

To improve the approximation, we include a linear term that captures the slope of the function at x = a. If the derivative of f at a is denoted f'(a) (or simply the gradient m), the first-order Taylor polynomial is:

f(x) ≈ f(a) + f'(a)(x − a)

This is exactly the equation of the tangent line to the curve at x = a. You have already met tangent lines in GCSE when discussing instantaneous gradient on distance-time or velocity-time graphs, or when finding the gradient of a curve by drawing a tangent.

为了提高近似精度，我们引入一次项来捕捉函数在 x = a 处的斜率。若用 f'(a)（或记作梯度 m）表示 f 在 a 点的导数，则一阶泰勒多项式为：f(x) ≈ f(a) + f'(a)(x − a)。这正是曲线在 x = a 处切线的方程。你在 GCSE 中讨论距离-时间图或速度-时间图的瞬时梯度、或通过画切线求曲线斜率时，其实已经接触过切线了。

The expression f(a) + f'(a)(x − a) is simply the point-slope form of a straight line. In GCSE language, you know a line’s value at a and its gradient m, so you can predict values nearby. The linear approximation does precisely this for any differentiable function.

表达式 f(a) + f'(a)(x − a) 正是直线的点斜式。用 GCSE 的语言来说，你知道直线在 a 点的值及其梯度 m，于是便可以预测附近的数值。线性近似就是为任何可导函数做了同样的事情。

4. Using Linear Approximation to Estimate Values | 使用线性近似估算数值

Linear approximation is extremely practical. Suppose you want to estimate √4.1 without a calculator. Let f(x) = √x, and choose a = 4, where you know f(4) = 2. The derivative f'(x) = 1/(2√x), so f'(4) = 1/(2×2) = 0.25. The linear approximation gives:

√4.1 ≈ 2 + 0.25 × (4.1 − 4) = 2 + 0.25 × 0.1 = 2.025

The true value is about 2.0248, so the error is only 0.0002. This technique works because the square root curve is relatively flat near x = 4.

线性近似非常实用。假设你想不用计算器就估算出 √4.1。设 f(x) = √x，选取 a = 4，已知 f(4) = 2。导数为 f'(x) = 1/(2√x)，故 f'(4) = 1/(2×2) = 0.25。线性近似给出：√4.1 ≈ 2 + 0.25 × (4.1 − 4) = 2.025。真实值约为 2.0248，误差仅 0.0002。这个技巧之所以奏效，是因为平方根曲线在 x = 4 附近较为平缓。

Another GCSE-friendly example is estimating (1.01)³ using f(x) = x³ at a = 1. f'(x) = 3x², so f'(1) = 3. Then (1.01)³ ≈ 1 + 3 × 0.01 = 1.03, while the exact value is 1.030301—close enough for many real-life situations.

另一个 GCSE 友好的例子是用 f(x) = x³ 在 a = 1 处估算 (1.01)³。f'(x) = 3x²，故 f'(1) = 3。于是 (1.01)³ ≈ 1 + 3 × 0.01 = 1.03，而精确值为 1.030301，对许多实际情况而言已足够接近。

5. Quadratic Approximation: Getting More Accurate | 二次近似：更加精确

A linear approximation only uses the first derivative. To capture the curvature of the graph, we add a quadratic term involving the second derivative f”(a). The second-order Taylor polynomial is:

f(x) ≈ f(a) + f'(a)(x − a) + (1/2) f”(a)(x − a)²

This fits a parabola that matches the function’s value, slope, and concavity at x = a. For √x at a = 4, f”(x) = −1/(4x^(3/2)), so f”(4) = −1/(4×8) = −1/32. The quadratic approximation for √4.1 becomes:

线性近似只用到了一阶导数。为了捕捉图像的弯曲程度，我们加入一个包含二阶导数 f”(a) 的二次项。二阶泰勒多项式为：f(x) ≈ f(a) + f'(a)(x − a) + (1/2) f”(a)(x − a)²。它拟合出的抛物线在 x = a 处与函数的值、斜率和凹凸性都保持一致。对于 f(x) = √x 在 a = 4 处的近似：f”(x) = −1/(4x^(3/2))，故 f”(4) = −1/32。于是 √4.1 的二次近似等于：

2 + 0.25 × 0.1 + (1/2) × (−1/32) × (0.1)² = 2.025 − 0.00015625 = 2.02484375

This is almost identical to the true value 2.02484567… The quadratic term corrected for the slight overestimate from the linear model.

这个结果几乎与真实值 2.02484567… 完全相同。二次项修正了线性模型略为高估的部分。

In practice, adding more terms reduces the error dramatically, but requires computing higher derivatives. GCSE students won’t be expected to compute these derivatives, but recognising that curved graphs can be approximated locally by polynomials is a valuable insight.

在实践中，增加更多项能显著减小误差，但这需要计算更高阶导数。GCSE 学生不需要实际计算这些导数，但认识到曲线图可以在局部用多项式来近似，是一个非常宝贵的洞见。

6. The General Form of a Taylor Polynomial | 泰勒多项式的一般形式

In general, the Taylor series of a function f about x = a is an infinite sum:

f(x) = f(a) + f'(a)(x − a) + (f”(a)/2!)(x − a)² + (f”'(a)/3!)(x − a)³ + …

Here n! (n factorial) is the product of all positive integers up to n. For example, 3! = 3 × 2 × 1 = 6. The coefficients involve derivatives evaluated at a, and each term contains a power of (x − a). This expansion continues indefinitely if the function is infinitely differentiable.

一般而言，函数 f 在 x = a 处的泰勒级数是一个无穷级数：f(x) = f(a) + f'(a)(x − a) + (f”(a)/2!)(x − a)² + (f”'(a)/3!)(x − a)³ + …。其中 n! 是 n 的阶乘，即前 n 个正整数的乘积，例如 3! = 6。系数包含在 a 点的各阶导数值，每一项则含有 (x − a) 的幂。如果函数无限次可导，这个展开式将无限进行下去。

The finite sum up to the n-th derivative is called the Taylor polynomial of degree n. Stopping after the linear term gives the tangent line; stopping after the quadratic term gives a parabolic fit; and so on.

直到第 n 阶导数的有限和被称为 n 次泰勒多项式。保留到一次项得到切线；保留到二次项得到抛物线拟合；以此类推。

7. Maclaurin Series: Expansion at Zero | 麦克劳林级数：在零点的展开

When the expansion point is chosen as a = 0, the Taylor series simplifies to a Maclaurin series:

f(x) = f(0) + f'(0)x + (f”(0)/2!)x² + (f”'(0)/3!)x³ + …

This is particularly neat because all (x − a) terms become just powers of x. Maclaurin series are used for many common functions like sine, cosine, and the exponential function.

当展开点取 a = 0 时，泰勒级数简化为麦克劳林级数：f(x) = f(0) + f'(0)x + (f”(0)/2!)x² + (f”'(0)/3!)x³ + …。这种形式特别简洁，因为所有 (x − a) 项都变成了 x 的幂。正弦、余弦、指数函数等许多常用函数的展开都采用麦克劳林级数。

For example, the exponential function eˣ has every derivative equal to eˣ, and e⁰ = 1. Thus its Maclaurin series is 1 + x + x²/2! + x³/3! + … Eventually, this sum converges to eˣ for any real x. At GCSE you may not have met e, but the idea is the same: we build a complicated function from simple building blocks.

例如，指数函数 eˣ 的各阶导数均等于 eˣ，且 e⁰ = 1。因此它的麦克劳林级数为 1 + x + x²/2! + x³/3! + …。最终，对于任意实数 x，该级数都会收敛到 eˣ。你在 GCSE 阶段可能没接触过 e，但原理是一样的：我们用简单的构件搭建出复杂的函数。

8. Taylor Series for Common Functions | 常见函数的泰勒级数

Here are some standard expansions with their intervals of convergence. These are derived from repeatedly differentiating and evaluating at one point.

Function f(x)	Maclaurin series (around 0)
sin x	x − x³/3! + x⁵/5! − x⁷/7! + …
cos x	1 − x²/2! + x⁴/4! − x⁶/6! + …
eˣ	1 + x + x²/2! + x³/3! + …
ln(1 + x)	x − x²/2 + x³/3 − x⁴/4 + …

These series show how functions that seem very different can all be represented by polynomials of infinite degree. In fields like physics and engineering, engineers often keep just the first few terms to get quick, decent approximations.

这些级数展示了看起来截然不同的函数如何都能用无穷次多项式来表示。在物理和工程等领域，人们常常只取前几项来快速得到不错的近似。

Even without understanding the derivations, you can appreciate that these patterns exist. For instance, substituting a small angle in radians into the first one or two terms of sin x gives a value remarkably close to the true sine, which is why small-angle approximations work.

即便不理解推导过程，你也能体会到这些规律的存在。例如，将一个以弧度为单位的小角度代入 sin x 的前一两项，就能得到一个与真实正弦值非常接近的数值，这就是小角度近似能奏效的原因。

9. Real-World Applications | 实际应用

Taylor series are not just theoretical; they are embedded in the technology you use every day. Your calculator approximates functions like sin, cos, and logarithms by evaluating the first several terms of their Taylor or Maclaurin series. The number of terms used depends on the required precision.

泰勒级数并非纯理论，它渗透在你日常使用的技术之中。计算器就是通过计算正弦、余弦、对数等函数的泰勒/麦克劳林级数前若干项来得出结果的。所用项数取决于所需精度。

In physics, the simple pendulum period T = 2π√(L/g) assumes a small swing angle. For larger angles, the exact formula involves an integral that can be approximated using a Taylor expansion to correct the small-angle formula. Similarly, satellite trajectories, electrical circuit analysis, and even economic models use Taylor series to linearise complex systems around an equilibrium point.

在物理学中，单摆周期公式 T = 2π√(L/g) 假设了摆角很小。对于较大摆角，精确公式包含一个积分，可利用泰勒展开对小角度公式进行修正。类似地，卫星轨道、电路分析甚至经济模型，都会借助泰勒级数在平衡点附近将复杂系统线性化。

Understanding that a curve behaves almost like its tangent line near a point is a recurring theme in science and engineering. It is the foundation of differential calculus and optimisation methods.

理解一条曲线在一点附近几乎与其切线行为一致，是科学和工程中反复出现的主题。它是微分学和最优化方法的基础。

10. Connection to GCSE Maths and Exam Tips | 与GCSE数学的联系及考试技巧

While you won’t be asked to compute a Taylor series in your GCSE AQA Mathematics exam, the underlying skills are highly relevant. Every time you draw a tangent to a curve to find an instantaneous gradient, you are effectively using a first-order Taylor linearisation. When you estimate a value from a graph by reading off a straight line segment, you are applying a similar piecewise-linear approximation.

虽然你在 GCSE AQA 数学考试中不会遇到需要计算泰勒级数的题目，但背后的技能密切相关。每次你画曲线的切线来求瞬时梯度，其实就是在用一种一阶泰勒线性化。当你通过读取直线段来估算图形上的数值时，你是在应用类似的分段线性近似。

Here are some tips linked to these ideas:

When estimating the area under a curve or a gradient at a point, use small intervals; this mirrors the idea that Taylor approximations improve as x gets closer to a.
Check the reasonableness of linear estimates by considering whether the curve is concave up or down (i.e., the sign of the second derivative). This helps you decide if your tangent line gives an overestimate or underestimate.
In problems about iterative methods or sequences, you are often stepping along a tangent line—the Newton-Raphson method, for instance, is built directly on first-order Taylor expansion.

以下是一些与之相关的技巧：

在估算曲线下方面积或某点的梯度时，尽量使用小区间；这呼应了 x 越接近 a、泰勒近似就越精确的原理。
通过考虑曲线是上凸还是下凸（即二阶导数的符号）来检验线性估计的合理性。这能帮你判断切线给出的估值是偏大还是偏小。
在迭代法或数列问题中，你往往正是沿着切线步进——例如牛顿-拉夫森方法就是直接建立在一阶泰勒展开之上。

Recognising that a straight line can momentarily represent a curve is a deep understanding that goes beyond formula memorisation. It connects graphical interpretation, ratio, and rate of change—all core GCSE competencies.

认识到一条直线可以在瞬间代表一条曲线，这是一种超越公式记忆的深层理解。它把图形解读、比值和变化率联系在一起——而这些都是 GCSE 的核心能力。

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

Taylor Series | 泰勒级数