跳到主要内容

京都大学 情報学研究科 知能情報学専攻 2021年2月実施 基礎科目 F1-2

Author​

祭音Myyura (co-authored with GPT 5.6 SOL)

Description​

Q.1​

Given

z=2y−x4y2,y=cos⁡x,z=2y-x^4y^2,\qquad y=\cos x,

compute ∂z/∂x\partial z/\partial x and ∂z/∂y\partial z/\partial y. Using these results, determine dz/dxdz/dx.

Q.2​

Evaluate

∫−10(11−x+1−x) dx.\int_{-1}^{0}\left(\frac{1}{1-x}+\sqrt{1-x}\right)\,dx.

Q.3​

Find the maximum rate of change of

f(x,y)=x2+y3f(x,y)=\sqrt{x^2+y^3}

at (1,1)(1,1) and the direction in which this maximum rate occurs.

题目描述​

  1. 已知 z=2y−x4y2z=2y-x^4y^2、y=cos⁡xy=\cos x,求 ∂z/∂x\partial z/\partial x、∂z/∂y\partial z/\partial y,并据此求 dz/dxdz/dx。

  2. 计算

    ∫−10(11−x+1−x) dx.\int_{-1}^{0}\left(\frac{1}{1-x}+\sqrt{1-x}\right)\,dx.
  3. 求 f(x,y)=x2+y3f(x,y)=\sqrt{x^2+y^3} 在 (1,1)(1,1) 处的最大变化率及其方向。

Kai​

Q.1​

Treating xx and yy as independent variables,

∂z∂x=−4x3y2,∂z∂y=2−2x4y.\boxed{\frac{\partial z}{\partial x}=-4x^3y^2}, \qquad \boxed{\frac{\partial z}{\partial y}=2-2x^4y}.

Since dy/dx=−sin⁡xdy/dx=-\sin x, the chain rule gives

dzdx=∂z∂x+∂z∂ydydx=−4x3cos⁡2x−(2−2x4cos⁡x)sin⁡x.\begin{aligned} \frac{dz}{dx} &=\frac{\partial z}{\partial x} +\frac{\partial z}{\partial y}\frac{dy}{dx}\\ &=-4x^3\cos^2x-(2-2x^4\cos x)\sin x. \end{aligned}

Therefore,

dzdx=−4x3cos⁡2x−2sin⁡x+2x4sin⁡xcos⁡x.\boxed{\frac{dz}{dx} =-4x^3\cos^2x-2\sin x+2x^4\sin x\cos x}.

Q.2​

An antiderivative is

−log⁡(1−x)−23(1−x)3/2.-\log(1-x)-\frac23(1-x)^{3/2}.

Hence

∫−10(11−x+1−x) dx=log⁡2+23(23/2−1)=log⁡2+42−23.\begin{aligned} \int_{-1}^{0}\left(\frac{1}{1-x}+\sqrt{1-x}\right)\,dx &=\log2+\frac23\left(2^{3/2}-1\right)\\ &=\boxed{\log2+\frac{4\sqrt2-2}{3}}. \end{aligned}

Q.3​

The gradient is

∇f(x,y)=(xx2+y3,3y22x2+y3).\nabla f(x,y)= \left( \frac{x}{\sqrt{x^2+y^3}}, \frac{3y^2}{2\sqrt{x^2+y^3}} \right).

At (1,1)(1,1),

∇f(1,1)=(12,322).\nabla f(1,1)= \left(\frac1{\sqrt2},\frac3{2\sqrt2}\right).

The maximum directional derivative is the gradient norm:

∥∇f(1,1)∥=12+98=264.\boxed{\|\nabla f(1,1)\| =\sqrt{\frac12+\frac98} =\frac{\sqrt{26}}4}.

It occurs in the unit direction of the gradient,

u=∇f(1,1)∥∇f(1,1)∥=113(2,3).\boxed{\boldsymbol{u} =\frac{\nabla f(1,1)}{\|\nabla f(1,1)\|} =\frac1{\sqrt{13}}(2,3)}.