Double Integral and PDE

Calculus Level 4

f ( x , y ) f(x,y) satisfies the partial differential equation 2 f x 2 + 2 f y 2 = 1. \frac {\partial^2f}{\partial x^2}+\frac {\partial^2f}{\partial y^2}=1. Evaluate D ( x f x + y f y ) d x d y , \iint_D\left(x\frac {\partial f}{\partial x}+y\frac {\partial f}{\partial y}\right)dxdy, where D = { ( x , y ) R 2 x 2 + y 2 1 } . D=\{(x,y) \in \mathbb{R}^2 \mid x^2+y^2 \leq 1\}.


The answer is 0.785398.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

2 solutions

Mark Hennings
Jun 11, 2018

Stokes' Theorem in the plane tells us that, for smooth functions A , B A,B D ( x A + y B ) d x d y = C ( A d y B d x ) \iint_D (\partial_xA + \partial_yB)\,dx\,dy \; = \; \oint_C(A\,dy - B\,dx) where C C is the positively oriented unit circle. Thus D ( x ( r 2 x f ) + y ( r 2 y f ) ] d x d y = C r 2 ( x f d y y f d x ) = C ( x f d y y f d x ) = D ( x ( x f ) + y ( y f ) ) d x d y = D 2 f d x d y = D d x d y \begin{aligned} \iint_D \Big(\partial_x(r^2\partial_xf) + \partial_y(r^2\partial_yf)\Big]\,dx\,dy & = \; \oint_Cr^2 (\partial_xf\,dy - \partial_yf\,dx) \\ & = \; \oint_C (\partial_xf\,dy - \partial_yf\,dx) \\ & = \; \iint_D \Big(\partial_x(\partial_xf) + \partial_y(\partial_yf)\Big)\,dx\,dy \; = \; \iint_D \nabla^2f\,dx\,dy \; = \; \iint_D \,dx\,dy \end{aligned} so that D ( r 2 2 f + 2 ( x x f + y y f ) ) d x d y = D d x d y \iint_D\Big(r^2\nabla^2f + 2\big(x\partial_xf + y\partial_yf\big)\Big)\,dx\,dy \; =\; \iint_D\,dx\,dy and hence D ( x x f + y y f ) d x d y = 1 2 D ( 1 r 2 ) d x d y = π 0 1 ( 1 r 2 ) r d r = 1 4 π \iint_D \big(x\partial_xf + y\partial_yf\big)\,dx\,dy \; = \; \tfrac12\iint_D(1-r^2)\,dx\,dy \; = \; \pi\int_0^1 (1-r^2)r\,dr \; = \; \boxed{\tfrac14\pi}

Steven Chase
Jun 7, 2018

Note: My assumed form for the solution is somewhat lacking in generality, as @Mark Hennings has pointed out in his solution

Taking double derivatives and ending up with constants means that we're talking about quadratics. The general form should be:

f = A x 2 + B x + C y 2 + D y + E 2 A + 2 C = 1 f = A x^2 + B x + C y^2 + D y + E \\ 2 A + 2 C = 1

Evaluating the integrand:

x f x + y f y = x ( 2 A x + B ) + y ( 2 C y + D ) = 2 A x 2 + 2 C y 2 + B x + D y = 2 A r 2 c o s 2 θ + 2 C r 2 s i n 2 θ + B r c o s θ + D r s i n θ x \, \frac{\partial{f}}{\partial{x}} + y \, \frac{\partial{f}}{\partial{y}} = x (2 A x + B) + y (2 C y + D) \\ = 2 A x^2 + 2 C y^2 + B x + D y \\ = 2 A \, r^2 \, cos^2 \theta + 2 C \, r^2 \, sin^2 \theta + B r \, cos \theta + D r \, sin \theta

Integrating sine or cosine over the unit disk results in zero overall, so the lower order terms drop out, and we're effectively left with:

x f x + y f y = 2 A r 2 c o s 2 θ + 2 C r 2 s i n 2 θ x \, \frac{\partial{f}}{\partial{x}} + y \, \frac{\partial{f}}{\partial{y}} = 2 A \, r^2 \, cos^2 \theta + 2 C \, r^2 \, sin^2 \theta

The integral becomes:

0 2 π 0 1 ( 2 A r 2 c o s 2 θ + 2 C r 2 s i n 2 θ ) r d r d θ = 0 2 π 0 1 ( 2 A r 3 c o s 2 θ + 2 C r 3 s i n 2 θ ) d r d θ \int_0^{2 \pi} \int_0^1 (2 A \, r^2 \, cos^2 \theta + 2 C \, r^2 \, sin^2 \theta ) r \, d r \, d \theta \\ = \int_0^{2 \pi} \int_0^1 (2 A \, r^3 \, cos^2 \theta + 2 C \, r^3 \, sin^2 \theta ) d r \, d \theta

Substituting in for C in terms of A yields:

0 2 π 0 1 ( 2 A r 3 c o s 2 θ + ( 1 2 A ) r 3 s i n 2 θ ) d r d θ \int_0^{2 \pi} \int_0^1 (2 A \, r^3 \, cos^2 \theta + (1 - 2 A) \, r^3 \, sin^2 \theta ) d r \, d \theta

The integrals of sine squared and cosine squared over the unit disk are both equal, so these components now cancel, and we're left with:

0 2 π 0 1 r 3 s i n 2 θ d r d θ = π 4 \int_0^{2 \pi} \int_0^1 r^3 \, sin^2 \theta \, d r \, d \theta = \frac{\pi}{4}

It is not true that we are forced to consider only quadratics. We are looking at 1 2 x 2 \tfrac12x^2 plus a general solution of Laplace's equation (that is nonsingular at the origin). Thus we are looking at a combination (possibly infinite) of the functions r k cos k θ , r k sin k θ k N r^k \cos k\theta \;,\; r^k\sin k\theta \hspace{2cm} k \in \mathbb{N} together with the constant function 1 1 . You are restricting attention to k = 1 , 2 k=1,2 only! You have to use Stokes' Theorem to prove this result completely. See my proof.

Mark Hennings - 2 years, 12 months ago

Log in to reply

Interesting. Yes, ensuring sufficient generality in the form of the equation can be tricky.

Steven Chase - 2 years, 12 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...