📚 The Chain Rule for Multivariable Composite Functions | 多元复合函数链式法则
In single-variable calculus, the chain rule tells us how to differentiate a composite function. When we move to multivariable calculus, the chain rule expands into a richer, more powerful tool that connects partial derivatives along different paths. This topic is a core requirement in IB Mathematics Analysis and Approaches HL, and it frequently appears in both Paper 2 and Paper 3. Mastering the multivariable chain rule not only helps you solve differentiation problems efficiently but also builds the foundation for understanding gradients, directional derivatives, and optimization in higher dimensions.
在单变量微积分中，链式法则告诉我们如何对复合函数求导。当我们进入多元微积分领域时，链式法则扩展为一个更丰富、更强大的工具，它连接了不同路径上的偏导数。这一主题是 IB 数学分析与方法 HL 课程的核心要求，经常出现在 Paper 2 和 Paper 3 中。掌握多元链式法则不仅能帮助你高效解决求导问题，还能为理解梯度、方向导数和多元优化奠定坚实基础。
1. Review of the Single-Variable Chain Rule | 一元链式法则回顾
ative of y with respect to x is given by the product of the derivative of f with respect to u and the derivative of g with respect to x:
在深入探讨多元推广之前，有必要回顾一元链式法则。如果 y = f(u) 且 u = g(x)，那么 y 关于 x 的导数等于 f 关于 u 的导数与 g 关于 x 的导数之积：
dy/dx = (dy/du) × (du/dx)
This formula expresses a simple idea: the rate of change of y with respect to x can be decomposed into the rate of change of y with respect to an intermediate variable u, times the rate of change of u with respect to x. The chain rule can be thought of as "cancelling" the du terms, although this is only a mnemonic device rather than a rigorous proof.
这个公式表达了一个简单的思想：y 关于 x 的变化率可以分解为 y 关于中间变量 u 的变化率乘以 u 关于 x 的变化率。链式法则可以被理解为"约去" du 项，但这仅仅是一种记忆方法，而非严格的证明。
2. Partial Derivatives and Total Differentials | 偏导数与全微分
e rate of change of f when only x changes, holding y fixed. The total differential of z combines both partial derivatives:
在多元微积分中，函数通常依赖于多个变量。对于函数 z = f(x, y)，偏导数 ∂f/∂x 衡量的是在保持 y 不变的条件下，仅 x 变化时 f 的变化率。z 的全微分结合了两个偏导数：
dz = (∂f/∂x) dx + (∂f/∂y) dy
The total differential provides a linear approximation to the change in z when both x and y change by small amounts. It plays a central role in understanding the multivariable chain rule, because when x and y are themselves functions of another variable t, the total differential allows us to compute dz/dt by dividing both sides of this expression by dt.
全微分提供了当 x 和 y 都发生微小变化时 z 的变化的线性近似。它在理解多元链式法则中起着核心作用，因为当 x 和 y 本身是另一个变量 t 的函数时，全微分允许我们通过将等式两边同时除以 dt 来计算 dz/dt。
3. The Simplest Multivariable Case: One Independent Variable | 最简单的多元情形：一个自变量
(t) are themselves functions of a single parameter t. In this case, the chain rule states:
最基本的多元链式法则情境涉及函数 z = f(x, y)，其中 x = g(t) 和 y = h(t) 本身是单一参数 t 的函数。在这种情况下，链式法则表述为：
dz/dt = (∂f/∂x)(dx/dt) + (∂f/∂y)(dy/dt)
Notice how each partial derivative of f is multiplied by the ordinary derivative of the corresponding independent variable with respect to t, and the two products are added together. This makes intuitive sense: a change in t affects z through both the x-channel and the y-channel, and the total effect is the sum of these separate contributions.
注意，f 的每个偏导数都乘以相应自变量关于 t 的普通导数，然后将两个乘积相加。这在直觉上很合理：t 的变化通过 x 通道和 y 通道两条路径影响 z，总效果是这两条独立贡献之和。
4. The General Multivariable Chain Rule | 多元链式法则的一般形式
Now consider a more general situation. Suppose z = f(x₁, x₂, … , xₙ), where each xᵢ is a differentiable function of m variables t₁, t₂, … , tₘ. Then for each variable tⱼ, the chain rule gives:
现在考虑更一般的情形。假设 z = f(x₁, x₂, … , xₙ)，其中每个 xᵢ 都是 m 个变量 t₁, t₂, … , tₘ 的可微函数。那么对于每个变量 tⱼ，链式法则给出：
∂z/∂tⱼ = (∂z/∂x₁)(∂x₁/∂tⱼ) + (∂z/∂x₂)(∂x₂/∂tⱼ) + … + (∂z/∂xₙ)(∂xₙ/∂tⱼ)
In compact summation notation, this is written as:
用紧凑的求和符号表示，即为：
∂z/∂tⱼ = Σᵢ₌₁ⁿ (∂z/∂xᵢ)(∂xᵢ/∂tⱼ)
This general formula covers all chain rule cases. The key idea is that the partial derivative of z with respect to any one of the intermediate variables tⱼ requires summing over all paths that connect z to tⱼ through the xᵢ variables.
这一通用公式涵盖了所有链式法则的情形。关键思想是：z 关于任一中间变量 tⱼ 的偏导数需要对所有通过 xᵢ 变量连接 z 到 tⱼ 的路径求和。
5. Vector and Matrix Interpretation | 向量与矩阵解释
ng vectors and matrices. Define the gradient of z as the row vector of partial derivatives ∇z = (∂z/∂x₁, ∂z/∂x₂, … , ∂z/∂xₙ). Then the chain rule for all m variables simultaneously can be written as:
多元链式法则可以用向量和矩阵优雅地表达。定义 z 的梯度为偏导数行向量 ∇z = (∂z/∂x₁, ∂z/∂x₂, … , ∂z/∂xₙ)。那么对所有 m 个变量的链式法则可以同时写成：
∇t z = ∇x z · J
where ∇t z is the vector (∂z/∂t₁, ∂z/∂t₂, … , ∂z/∂tₘ), ∇ the gradient with respect to x, and J is the Jacobian matrix whose (i, j)-entry is ∂xᵢ/∂tⱼ matrix perspective is particularly powerful in higher dimensions and in theoretical work.
其中 ∇t z 是向量 (∂z/∂t₁, ∂z/∂t₂, … , ∂z/∂tₘ)，∇x z 是关于 x 的梯度，J 是雅可比矩阵，其第 (i, j) 元素为 ∂xᵢ/∂tⱼ。这种矩阵视角在高维空间和理论研究中尤为强大。
6. A Practical Example with Two Paths | 双路径的实例演练
), where x = t² and y = eᵗ. We want to find dz/dt. First, compute the partial derivatives:
让我们将链式法则应用到一个具体问题上。考虑 z = x²y + sin(xy)，其中 x = t²，y = eᵗ。我们希望求 dz/dt。首先计算偏导数：
∂z/∂x = 2xy + y cos(xy), ∂z/∂y = x² + x cos(xy)
Next, compute dx/dt = 2t and dy/dt = eᵗ. Applying the chain rule:
接下来，计算 dx/dt = 2t 和 dy/dt = eᵗ。应用链式法则：
dz/dt = (2xy + y cos(xy))(2t) + (x² + x cos(xy))(eᵗ)
express dz/dt entirely in terms of t. Performing the substitution yields a complete expression that can be simplified further if needed. This step-by-step approach is exactly what examiners expect to see.
最后，代入 x = t² 和 y = eᵗ，将 dz/dt 完全表示为 t 的函数。完成代入后得到完整的表达式，如需化简可进一步整理。这种逐步求解的方式正是考官期望看到的。
7. Explicit Substitution vs. Chain Rule | 直接代入与链式法则的比较
t. Direct substitution gives z = cos²t + sin²t = 1, so dz/dt = 0. Applying the chain rule:
求 dz/dt 的另一种方法是将 x(t) 和 y(t) 直接代入 z = f(x, y)，得到 z 仅为 t 的函数，然后使用普通单变量技巧求导。考虑 z = x² + y²，其中 x = cos t，y = sin t。直接代入得 z = cos²t + sin²t = 1，因此 dz/dt = 0。应用链式法则：
dz/dt = 2x(−sin t) + 2y(cos t) = 2 cos t(−sin t) + 2 sin t(cos t) = 0
Both methods agree. Direct substitution is often simpler when the function has a simple algebraic form, but the chain rule is indispensable when substitution yields expressions too complicated to simplify, or when only the partial derivatives are known without explicit formulas.
两种方法结果一致。当函数形式简单时，直接代入往往更简便，但当代入后表达式过于复杂难以化简，或者只知道偏导数而没有显式公式时，链式法则就不可或缺。
8. Implicit Differentiation via the Chain Rule | 借助链式法则进行隐函数求导
ines y implicitly as a function of x. Differentiating both sides with respect to x, and using the chain rule for the y-dependence, we obtain:
多元链式法则为隐函数求导提供了简洁的推导。假设 F(x, y) = 0 隐式地将 y 定义为 x 的函数。等式两边关于 x 求导，对关于 y 的部分使用链式法则，得到：
∂F/∂x + (∂F/∂y)(dy/dx) = 0, hence dy/dx = −(∂F/∂x) / (∂F/∂y)
This famous formula is a direct consequence of the chain rule. It is particularly useful when F is too complicated to solve explicitly for y. Many IB exam questions on implicit differentiation are essentially testing whether you can apply this multivariable chain rule correctly.
这个著名公式是链式法则的直接推论。当 F 过于复杂而无法显式解出 y 时，它尤其有用。许多 IB 考试中关于隐函数求导的问题，本质上就是在测试你是否能正确运用多元链式法则。
9. Higher-Order Partial Derivatives and the Chain Rule | 高阶偏导数与链式法则
ond derivative d²z/dt² requires differentiating the first derivative expression with respect to t:
在计算复合函数的高阶导数时，链式法则必须重复应用，并且需要格外小心，因为一阶偏导数 ∂f/∂x 和 ∂f/∂y 本身也是 x 和 y 的函数。例如，如果 z = f(x, y)，其中 x = g(t)，y = h(t)，那么二阶导数 d²z/dt² 需要对一阶导数表达式关于 t 再次求导：
d²z/dt² = d/dt[(∂f/∂x)(dx/dt) + (∂f/∂y)(dy/dt)]
∂y, and ∂²f/∂y². This process is algebraically intensive but follows a systematic pattern.
每个偏导数 ∂f/∂x 和 ∂f/∂y 本身都必须使用链式法则求导，这会引入二阶偏导数 ∂²f/∂x²、∂²f/∂x∂y 和 ∂²f/∂y²。这一过程代数运算量大，但遵循系统化的模式。
10. Dependency Diagrams: A Visual Method | 依赖关系图：可视化方法
A dependency diagram (or tree diagram) is a powerful visual tool for organising the terms in the multivariable chain rule. To construct one, write the dependent variable z at the top, draw branches to each intermediate variable x, y, … , and from each intermediate variable draw branches to the final independent variables t₁, t₂, … .
依赖关系图（或称树形图）是一种强大的可视化工具，用于整理多元链式法则中的各项。构建方法是：将因变量 z 写在最上方，绘制分支连接到每个中间变量 x, y, … ，再从每个中间变量绘制分支连接到最终自变量 t₁, t₂, … 。
Each path from z to a given independent variable tⱼ contributes one term to ∂z/∂tⱼ and z → y → t. The chain rule sum simply adds the products of derivatives along each path.
从 z 到给定自变量 tⱼ 的每条路径贡献 ∂z/∂tⱼ 中的一项。例如，如果 z 依赖于 x 和 y，且 x 和 y 都依赖于 t，那么从 z 到 t 有两条路径：z → x → t 和 z → y → t。链式法则的和式就是将每条路径上的导数乘积相加。
11. Common Mistakes and Exam Strategies | 常见错误与应试策略
derivatives: ∂x/∂t is not the same as dx/dt when x depends on more than one variable. A third mistake is evaluating partial derivatives at the original variables rather than at the substituted expressions.
学生在应用多元链式法则时，常犯几种可预见的错误。一种常见错误是遗漏某条路径，尤其是当存在三个或更多中间变量时。另一个常见错误是混淆偏导数与普通导数：当 x 依赖多个变量时，∂x/∂t 不同于 dx/dt。第三种错误是在原变量处而非代入后的表达式处计算偏导数。
To avoid these pitfalls, always start by drawing a dependency diagram, then systematically write down each path term, and finally substitute all expressions before simplifying. In IB exams, showing the intermediate chain rule expression without simplifying often earns full marks, so do not sacrifice accuracy for speed.
为避免这些陷阱，一开始就绘制依赖关系图，然后系统地写出每条路径的项，最后在化简前代入所有表达式。在 IB 考试中，展示中间的链式法则表达式而不化简往往能获得满分，因此不要为了速度而牺牲准确性。
12. Summary and Key Takeaways | 总结与核心要点
with partial derivatives and summing over all dependency paths. The general formula ∂z/∂tⱼ = Σᵢ (∂z/∂xᵢ)(∂xᵢ/∂tⱼ) is universal and can be applied to any composite function scenario.
多元链式法则是一元版本的自然推广，用偏导数替代普通导数，并对所有依赖路径求和。通用公式 ∂z/∂tⱼ = Σᵢ (∂z/∂xᵢ)(∂xᵢ/∂tⱼ) 具有普适性，适用于任何复合函数情境。
Key takeaways for the IB exam are: identify the dependency structure clearly, use the chain rule (not direct substitution) when the expression is complex, remember that implicit differentiation is a special case of the chain rule, and always draw a dependency diagram for multi-path problems.
IB 考试的核心要点是：清晰识别依赖结构，当表达式复杂时使用链式法则（而非直接代入），记住隐函数求导是链式法则的特例，并且对多路径问题始终绘制依赖关系图。
Familiarity with the vector-matrix formulation provides a deeper understanding and prepares you for university-level mathematics. With consistent practice on past-paper problems, the multivariable chain rule will become one of the most reliable tools in your differentiation arsenal.
熟悉向量-矩阵表述能加深理解，并为大学阶段的学习做好准备。通过持续练习历年真题，多元链式法则将成为你求导工具箱中最可靠的工具之一。
Published by TutorHao | IB 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