Doubt about an improper multiple integral

109 Views Asked by At

Study the convergence of the following integral $$\iint_A\frac{dxdy}{x+y}$$

Where $A=[0,1]^2$. If we consider $C=\{(x,y)\in \mathbb R : x^2+y^2\le1,\ x>0, y>0 \}$, we can say that $$\iint_A\frac{dxdy}{x+y}=\iint_{A-C}\frac{dxdy}{x+y}+\iint_C\frac{dxdy}{x+y}$$

Now the integral on $A-C$ is a proper integral, while the integral on $C$ is improper in $(x,y)=(0,0)$. Applying polar coordinates: $$\iint_C \frac{dxdy}{x+y}=\lim_{\varepsilon \to 0^+}\int_\varepsilon^1d\rho\int_0^{\frac \pi2}\frac{d\theta}{\cos \theta + \sin \theta}=\lim_{\varepsilon\to0^+}(1-\varepsilon)\int_0^{\frac \pi2}\frac{d\theta}{\cos \theta + \sin \theta}$$

Now we can say that $\cos \theta + \sin \theta$ is a continuous function and $C$ is compact, so for the Weierstrass theorem $\exists m>0 : \cos \theta + \sin \theta \ge m, \ \forall \theta \in[0,\frac \pi2]$, and he concludes that this last integral is convergent. I've doubts about:

1) Why do we need to say all the stuff about $m$ in the end? It's because we're doing a limit in two dimensions so we have to prove that it is uniform in $\theta$ so it exists? Or is for another reason?

2) After this affirmation about $m$ he just applies squeeze theorem like this $$\lim_{\varepsilon\to0^+}(1-\varepsilon)\int_0^{\frac \pi2}\frac{d\theta}{\cos \theta + \sin \theta}\le\lim_{\varepsilon\to0^+}(1-\varepsilon)\int_0^{\frac \pi2}\frac{d\theta}{m}$$

And since the last integral is convergent he concludes that the other one is convergent too, or is he using something else?

Thanks for your time.

1

There are 1 best solutions below

1
On

(2) is the reason for (1).

We need to affirm that (1) holds so that we can say (2) does.


$$\exists m{>}0\ ~\forall \theta\in[0;\pi/2] : (m\leq \cos \theta+\sin\theta)\tag 1$$

Which is true since the trig. sum is strictly positive over that interval.


$$0~\leq ~\lim_{\epsilon\to 0^+}(1-\epsilon)\int_0^{\pi/2}\dfrac{\mathrm d\theta }{\cos\theta+\sin\theta}~\leq~ \lim_{\epsilon\to 0^+}(1-\epsilon)\int_0^{\pi/2}\dfrac{\mathrm d\theta }{m}\tag 2$$

Which means the integral we seek must be less than a finite value, whatever it is.