Remark
- In Question 3, we use simply compare values to determine the maximum or minimum on the boundary. While we acknowledge that this method is not entirely rigorous, we have not covered a more rigorous method in this course.
- It is recommended that you review the double angle formula before the midterm.
- We did not explicitly introduce the second method used in Question 6. We suggest that you review the method.
Let $f(x,y)=\sqrt{1+x+2y}$.
- Find Gradient of $f$, and explain what the meaning of gradient.
- Find vectors $\vec{v}$ such that the directional derivative of $f$ along the vector $\vec{v}$ at $(2,3)$ is $0$.
- Find an equation of the tangent plain to the graph $z=f(x,y)$ when $x=2$ and $y=3$.
- Use the linearization of $f$ at $(2,3)$ to estimate the value of $f(2.01,2.99)$.
Answer
- $f_x = \frac{1}{2\sqrt{1+x+2y}}$, $f_y = \frac{1}{\sqrt{1+x+2y}}$. The gradient of a function $f$ at a point $(a,b)$ is a vector pointing to the direction where has largest directional derivative at the point, i.e. \(\frac{\partial f}{\partial\nabla f} = \max_\vec{v}\left\\{\frac{\partial f}{\partial \vec{v}}\right\\}.\)
- We notice that $\vec{v}$ being perpendicular to $\nabla f$ follows from the following equation \(\frac{\partial f}{\partial\vec{v}} = \vec{v}\cdot\nabla f =0.\) Therefore, $\vec{v}= \left(\frac{\partial f}{\partial y}(2,3),-\frac{\partial f}{\partial x}(2,3)\right) = \left(-\frac{1}{3},\frac{1}{6}\right)$
- $\frac{1}{6}(x-2)+\frac{1}{3}(y-3)=z-3$
- Let $L(x,y)=3+\frac{1}{6}(x-2)+\frac{1}{3}(y-3)$. We have $f(2.01,2.99)\approx 3-\frac{0.01}{3}$.
Let $f(x,y)=x^2y-xy^2$. Suppose both $x$ and $y$ are functions of $t$. Say $x(t)=\ln(t)$ and $y(t)=\tan(t)$. Find $\frac{\partial F}{\partial t}$.
Answer
$\frac{\partial f}{\partial t} = \frac{\partial f}{\partial x}\frac{\partial x}{\partial t} + \frac{\partial f}{\partial y}\frac{\partial y}{\partial t} = \frac{2\ln(t)-\tan^2(t)}{t} + (2\ln(t)\tan(t)-\tan^2(t))\sec^2(t)$.
Suppose the function $f(x,y)=ax^2+bxy+cy^2+3x$ has a saddle point when $x = -\frac{6}{5}$ and $y = \frac{3}{5}$, and suppose $f(0,1)=1$.
- Find $a$, $b$, and $c$.
- Find the local maximal and local minimal on the region $x^2+y^2\leq 9$.
Answer
- Since $\left(-\frac{6}{5},\frac{3}{5}\right)$ is a critical point of $f$, so we may set the following system of equations \(\nabla f\left(-\frac{6}{5},\frac{3}{5}\right) = (0,0)\Rightarrow \begin{cases} 4a-b=3\\\\ b-c=0 \end{cases}.\) Moreover, $c=1$ follows $f(0,1)=1$. Hence, we have $a=1$ and $b=1$.
- We already know $\left(-\frac{6}{5},\frac{3}{5}\right)$ is a critical point in the region $x^2+y^2\leq 9$. It remains to identify critical points on the boundary. By Lagrange method, we have to solve, for $g(x,y)=x^2+y^2$,
\(\begin{cases}
\nabla f =\lambda\nabla g\\\\
g = 9
\end{cases}\Rightarrow
\begin{cases}
2x + y + 3 = \lambda 2x\\\\
x + 2y = \lambda 2y\\\\
x^2 + y^2 = 9
\end{cases}.\)
We devide the first equation by the second, and derive
\(\begin{cases}
\frac{2x+y+3}{x+2y}=\frac{x}{y}\\\\
x^2 + y^2 = 9
\end{cases}\Rightarrow
\begin{cases}
x^2 - y^2 = 3y\\\\
x^2 + y^2 = 9
\end{cases}\Rightarrow
2y^2 + 3y - 9 = 0.\)
Therefore, we have $y=\frac{3}{2}$ or $-3$, and $x=\pm\frac{3\sqrt{3}}{2}$ or $0$ respectively.
For $(x,y) = (\frac{3\sqrt{3}}{2},\frac{3}{2})$, $f(x,y) = \frac{27(1+\sqrt{3})}{4}$ (max).
For $(x,y) = (0,3)$, $f(x,y) = 9$ (min).
Evaluate $\int_0^4\int_{\frac{\sqrt{y}}{2}}^1\sqrt{x^3+3}dxdy$.
Answer
To interchange the order of integration, we have to rewrite the limits of $\Omega: 0<y<4,\ \frac{\sqrt{y}}{2}<x<1$. It is \(\Omega: 0<x<1,\ 0<y<4x^2.\) Thus, the integral is \(\int_0^1\int_0^4x^2 \sqrt{x^3+3}dydx = \int_0^1 4x^2\sqrt{x^3+3}dx = \left.\frac{8}{9}(x^3+3)^{3/2}\right|^1_0.\)
Calculate the area of the region inside a half cardioid \(r=1-\sin(\theta)\)
Answer
The area is the following double integral: \(\begin{aligned} \int_{-\pi/2}^{\pi/2}\int_0^{1-\sin(\theta)} r drd\theta &= \frac{1}{2}\int_{-\pi/2}^{\pi/2} (1-\sin\theta)^2 d\theta\\\\ &= -\int_{-\pi/2}^{\pi/2} \sin\theta d\theta + \frac{1}{2}\int_{-\pi/2}^{\pi/2} 1 + \sin^2\theta d\theta\\\\ &= \left. \cos\theta\right\vert_{-\pi/2}^{\pi/2} + \frac{1}{2}\int_{-\pi/2}^{\pi/2}\frac{3}{2} - \frac{1}{2} + \sin^2\theta d\theta\\\\ &= 0 + \left. \frac{3}{4}\theta \right\vert_{-\pi/2}^{\pi/2} - \frac{1}{2}\int_{-\pi/2}^{\pi/2} \cos2\theta d\theta\\\\ &= \frac{3}{4}\pi + \left. \frac{1}{4}\sin 2\theta \right\vert_{-\pi/2}^{\pi/2} = \frac{3}{4}\pi. \end{aligned}\)
Evaluate $\int\int_\Omega (x^2-y^2)dxdy$ where $\Omega$ is enclosed by $y-x=1$, $y-x=0$, $xy=2$ and $xy=1$.
Answer
Consider a change of variable $(u,\ v) = \Phi(x,y) = (y-x,\ xy)$. Then, we have \(\Omega' = \Phi(\Omega): 0\leq u\leq 1,\ 1\leq v\leq 2.\)
- Method 1: If we desire to change variables from $x$ and $y$ to $u$ and $v$, we have to find “inverse” function of $\Phi$. We notice that \(y+x = \sqrt{u^2+4v}.\) Thus, we can consider $\Psi(u,v) = \left(\frac{\sqrt{u^2+4v}-u}{2},\frac{u+\sqrt{u^2+4v}}{2}\right)$, and note that $\Phi\circ\Psi = Id(u,v)$, the identity map on $uv$-coordinate. We compute Jacobian for the change of variable \(J_{\Psi}(u,v) = \frac{1}{\sqrt{u^2+4v}}.\) Then, the double integral is \(\int_1^2\int_0^1 u(\sqrt{u^2+4v})\frac{1}{\sqrt{u^2+4v}}dudv = \int_1^2\int_0^1 u dudv.\) It follows that the value of the double integral is $1$.
- Method 2: Observing $dxdy = J_{\Psi}(u,v)dudv$ and $dudv = J_{\Phi}dxdy$, we notice that $J_{\Psi}(u,v) = J_{\Phi}(x,y)^{-1}$. Therefore, we have \(\begin{aligned} \int\int_\Omega (x^2-y^2)dxdy &= \int\int_{\Omega'}(x^2-y^2)J_{\Psi}(u,v)dudv\\\\ &= \int\int_{\Omega'}(x^2-y^2)J_{\Phi}(x,y)^{-1}dudv. \end{aligned}\) where \(J_{\Phi}(x,y) = -x-y.\) Thus, it follows that \(\int\int_{\Omega'}(x^2-y^2)J_{\Phi}(x,y)^{-1}dudv = \int\int_{\Omega'}-(x-y)dudv = \int_1^2\int_0^1 ududv = 1.\)