Differentiating an Integral with Respect to a Parameter | 对参数求导的积分

📚 Differentiating an Integral with Respect to a Parameter | 对参数求导的积分

In calculus, we often meet integrals that contain a parameter in addition to the variable of integration. Differentiating such an integral with respect to that parameter is a powerful technique, known as differentiation under the integral sign, or the Leibniz integral rule. This method appears frequently in IB higher-level analysis, integration theory, and applied mathematics, allowing us to evaluate otherwise intractable definite integrals and to solve differential equations with integral terms.

在微积分中，我们经常会遇到除积分变量外还含有一个参数的积分。对这种积分关于参数求导是一种强大的技巧，称为积分号下求导法或莱布尼茨积分法则。这个方法经常出现在IB高水平分析、积分理论和应用数学中，使我们能计算原本难以处理的有定积分，并求解带积分项的微分方程。

1. Understanding Parameterized Integrals | 理解含参积分

A parameterized integral has the form I(α) = ∫_a^b f(x,α) dx, where x is the dummy variable and α is a parameter on which the integrand may depend. The limits a and b can be constants or functions of α. The goal is to find dI/dα without evaluating I(α) directly, or to use differentiation to evaluate the original integral.

含参积分的常见形式为 I(α) = ∫_a^b f(x,α) dx，其中 x 是哑变量，α 是被积函数可能依赖的参数。上下限 a 和 b 可以是常数，也可以是 α 的函数。我们的目标是不直接计算 I(α) 而求出 dI/dα，或者利用求导来计算原积分。

In IB analysis, you might see problems like "Find d/dα ∫₀¹ sin(αx) dx". Rather than integrating first, differentiating under the integral sign gives a shortcut.

在IB分析中，你可能会遇到“求 d/dα ∫₀¹ sin(αx) dx”这类问题。与其先积分，不如在积分号下直接求导来得简便。

---

2. The Leibniz Rule Statement | 莱布尼茨法则的表述

limits and the integrand depend on a parameter t, the Leibniz integral rule states: d/dt ∫_a(t)^b(t) f(x,t) dx = f(b(t),t)·b'(t) − f(a(t),t)·a'(t) + ∫_a(t)^b(t) (∂f/∂t)(x,t) dx.

在上下限与被积函数均依赖参数 t 的情况下，莱布尼茨积分法则为：d/dt ∫_a(t)^b(t) f(x,t) dx = f(b(t),t)·b'(t) − f(a(t),t)·a'(t) + ∫_a(t)^b(t) (∂f/∂t)(x,t) dx。

lf. If limits are constant, the first two terms vanish, leaving ∂f/∂t inside the integral.

等式右边的三个部分分别对应：积分上限的变化、积分下限的变化以及被积函数本身的变化。如果积分限为常数，前两项消失，只剩下积分内的 ∂f/∂t。

---

3. Conditions for Differentiating Under the Integral Sign | 积分号下求导的条件

(t), b(t) are differentiable. Then the rule holds. Deeper theorems involve dominated convergence for improper integrals.

从数学上讲，只有在一定的正则条件下才能进行交换。在IB阶段，我们假设 f(x,t) 及其偏导数 ∂f/∂t 在有界矩形 [a,b] × [c,d] 上连续，且 a(t), b(t) 可导，那么上述法则成立。对反常积分，则需要控制收敛定理等更深的知识。

When these conditions fail, naive differentiation can lead to incorrect results. You should always verify continuity of the integrand in the region of interest before applying the rule.

当这些条件不成立时，贸然求导可能导致错误。在应用法则前，你应当验证被积函数在相关区域上的连续性。

---

4. Proof Outline | 证明大纲

proof uses the definition of derivative as a limit of a difference quotient. Let F(t) = ∫_a(t)^b(t) f(x,t) dx. Write the difference F(t+h) − F(t) by splitting the integration intervals and using the mean value theorem. After careful limit management, you obtain the three terms. Rigorous justification requires uniform continuity and the ability to pass the limit inside the integral.

证明利用导数作为差商极限的定义。令 F(t) = ∫_a(t)^b(t) f(x,t) dx。将差 F(t+h) − F(t) 通过拆分积分区间并利用中值定理写出。仔细取极限后，便得到三个部分。严格的证明需要一致连续性以及极限与积分交换的合理性。

is enough: [F(t+h)−F(t)]/h = ∫ [f(x,t+h)−f(x,t)]/h dx → ∫ ∂f/∂t dx by uniform convergence.

在IB课堂上，对常限且 ∂f/∂t 连续的情形，给出如下概述： [F(t+h)−F(t)]/h = ∫ [f(x,t+h)−f(x,t)]/h dx → ∫ ∂f/∂t dx （由一致收敛保证）。

---

5. Special Case: Constant Limits | 特例：定限积分

The most frequently used form occurs when a and b are constants (independent of t): d/dt ∫_a^b f(x,t) dx = ∫_a^b ∂f/∂t dx. This simply says we can swap differentiation and integration.

最常用的形式是当 a 和 b 为常数（与 t 无关）时：d/dt ∫_a^b f(x,t) dx = ∫_a^b ∂f/∂t dx。这表明我们可以交换求导与积分的次序。

For example, if I(α) = ∫₀¹ e^αx dx, then I'(α) = ∫₀¹ x e^αx dx, which can be integrated by parts or used to find I(α) directly.

例如，若 I(α) = ∫₀¹ e^αx dx，则 I'(α) = ∫₀¹ x e^αx dx，可以通过分部积分求出，或用来直接解出 I(α)。

---

6. Variable Limits of Integration | 变积分限

When limits are functions of the parameter, they contribute extra terms. These arise from the fundamental theorem of calculus: differentiate an integral with a variable upper limit gives the integrand evaluated there times the derivative of the limit.

当积分限是参数的函数时，它们会贡献额外的项。这来自微积分基本定理：对变上限积分求导，得到被积函数在上限处的值乘以上限的导数。

The full rule elegantly merges those contributions with the term from the integrand's dependence. For instance, d/dx ∫₀^x sin(xt) dt requires careful application because both the upper limit and integrand contain x.

完整法则优雅地将这些贡献与被积函数依赖项合并。例如，求 d/dx ∫₀^x sin(xt) dt 就需要仔细运用法则，因为上限和被积函数都含有 x。

---

广告 / Sponsored

7. Example 1: A Basic Parameter Inside the Integrand | 例1：被积函数内的参数

Consider I(α) = ∫₀¹ g the Leibniz rule. Since limits are constant, we differentiate under the integral sign: I'(α) = ∫₀¹ ∂/∂α sin(αx) dx = ∫₀¹ x cos(αx) dx.

考虑 I(α) = ∫₀¹ sin(αx) dx。用莱布尼茨法则计算 dI/dα。由于上下限为常数，可以在积分号下求导：I'(α) = ∫₀¹ ∂/∂α sin(αx) dx = ∫₀¹ x cos(αx) dx。

Now integrate by parts: let u = x, dv = cos(αx)dx, so du = dx, v = (1/α) sin(αx). Then ∫₀¹ x cos(αx) dx = [x·(1/α) sin(αx)]₀¹ - ∫₀¹ (1/α) sin(αx) dx = (1/α) sin α + (1/α²)(cos α − 1). Hence I'(α) = (α sin α + cos α − 1)/α².

现在进行分部积分：令 u = x, dv = cos(αx) dx，则 du = dx, v = (1/α) sin(αx)。那么 ∫₀¹ x cos(αx) dx = [x·(1/α) sin(αx)]₀¹ - ∫₀¹ (1/α) sin(αx) dx = (1/α) sin α + (1/α²)(cos α − 1)。因此 I'(α) = (α sin α + cos α − 1)/α²。

this to the direct method: I(α) = (1 − cos α)/α, differentiate to verify the same result.

可与直接法比较：I(α) = (1 − cos α)/α，求导可验证得到相同结果。

---

8. Example 2: Differentiation with Variable Upper Limit | 例2：变上限求导

Find F'(x) if F(x) = ∫₀^x² e^t² dt. Here a = 0 (constant), b(x) = x², and f(t,x) = e^t² does not depend on x explicitly. The Leibniz rule gives: F'(x) = e^(x²)²·2x − e^0²·0 + ∫₀^x² 0 dt = 2x e^x⁴.

求 F'(x)，其中 F(x) = ∫₀^x² e^t² dt。这里 a = 0 为常数，b(x) = x²，而被积函数 f(t,x) = e^t² 不显含 x。由莱布尼茨法则得：F'(x) = e^(x²)²·2x − e^0²·0 + ∫₀^x² 0 dt = 2x e^x⁴。

hes the fundamental theorem of calculus combined with the chain rule directly: if G(u) = ∫₀^u e^t² dt, then F(x) = G(x²) → F'(x) = G'(x²)·2x = e^x⁴·2x.

注意链式法则的效果：上限的导数作为乘子乘以被积函数在上限处的值。该结果与微积分基本定理结合链式法则得出的一致：若 G(u) = ∫₀^u e^t² dt，则 F(x) = G(x²) → F'(x) = G'(x²)·2x = e^x⁴·2x。

---

9. Example 3: Error Function and Gaussian Integral | 例3：误差函数与高斯积分

The error function is defined as erf(x) = (2/√π) ∫₀^x e^−t² entiating with respect to x using the Leibniz rule (or FTC) yields: d/dx erf(x) = (2/√π) e^−x². This makes the error function a powerful tool in probability and statistics.

误差函数定义为 erf(x) = (2/√π) ∫₀^x e^−t² dt。用莱布尼茨法则（或微积分基本定理）对 x 求导得：d/dx erf(x) = (2/√π) e^−x²。这使得误差函数成为概率与统计中的有力工具。

More interestingly, the technique can be used to evaluate ∫₀^∞ e^−x² dx by introducing a parameter. Let I(α) = ∫₀^∞ e^{−α x²} er integrals) yields a differential equation whose solution gives the classic result √π/2.

更有趣的是，该技巧可用来计算 ∫₀^∞ e^−x² dx。引入参数 α > 0，令 I(α) = ∫₀^∞ e^{−α x²} dx。在积分号下求导（并为反常积分交换次序提供理由）得到一个微分方程，求解后得到经典的 √π/2 结果。

---

10. Common Mistakes to Avoid | 需避免的常见错误

ot constant. Always apply the full Leibniz rule. Another is overlooking that the integrand' be differentiated as a partial derivative ∂f/∂t, treating the dummy variable as constant.

一个常见错误是在积分限不是常数时忘记对上下限求导。务必使用完整的莱布尼茨法则。另一个错误是忽略被积函数对参数的依赖要按偏导数 ∂f/∂t 求导，即将哑变量视为常数。

Students also sometimes attempt to differentiate under the integral sign without checking continuity conditions, leading to incorrect results for improper integrals with singularities. In IB exams, ensure the problem guarantees the necessary smoothness.

学生有时未检查连续性条件就在积分号下求导，这对于有奇点的反常积分可能导致错误结果。在IB考试中，要确认题目已隐含必要的平滑条件。

---

11. Exam-Style Practice Question | 考试风格练习题

Question: Let I(α) = ∫₀^π ln(1 + α cos x) dx for |α| < 1. Find I'(α) by differentiating under the integral sign. Then, by recognising a standard integral, determine I(α).

题目：设 I(α) = ∫₀^π ln(1 + α cos x) dx，其中 |α| < 1。通过积分号下求导求 I'(α)。然后，利用一个标准积分确定 I(α)。

Solution sketch: For fixed α with |α| integrand is continuous and its partial derivative w.r.t α is cos x / (1+α cos x). Thus I'(α) = ∫₀^π / (1+α cos x) dx. Rewrite the integrand as (1/α)[1 − 1/(1+α cos x)] and integrate to get I'(α) = (π/α)(1 − 1/√1−α²). Then integrate back with I(0)=0 to find I(α) = π ln[(1+√1−α²)/2].

解答简概：对于满足 |α|<1 的固定 α，被积函数连续，且对 α 的偏导数为 cos x / (1+α cos x)。因此 I'(α) = ∫₀^π cos x / (1+α cos x) dx。将被积函数改写为 (1/α)[1 − 1/(1+α cos x)] 并积分，得 I'(α) = (π/α)(1 − 1/√1−α²)。然后利用 I(0)=0 积分回代，得 I(α) = π ln[(1+√1−α²)/2]。

---

12. Summary and Key Takeaways | 总结与要点

The Leibniz integral rule is an essential tool for differentiating integrals with respect to a parameter. Always remember the three-part formula, check constant versus variable limits, and treat the integrand's derivative as a partial derivative. This technique not only simplifies differentiation but can unlock clever integration strategies.

莱布尼茨积分法则是关于对参数求导的基本工具。务必记住三部分公式，检查积分限是常数还是变限，并将被积函数的导数视为偏导数。这一技巧不仅简化求导，还能开启巧妙的积分策略。

Mastering this topic will strengthen your calculus foundations and prepare you for university-level analysis, differential equations, and mathematical physics. Practise with both simple and cleverly parameterised integrals to gain confidence.

掌握这一专题将巩固你的微积分基础，并为大学阶段的分析、微分方程和数学物理做好准备。通过练习简单和巧妙参数化的积分，不断增强信心。

Published by TutorHao | IB Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）