2
$\begingroup$

I was trying to evaluate

$$\int \sec x dx \text{ using Complex Analysis }$$

I thought of rewriting it as $$\int \dfrac{1}{\Re(e^{\iota x})}$$ However, I cannot think of how to proceed from here. Could somebody please assist me in solving this Integral in the way I have tried? I would indeed be most grateful for your help. Many thanks!

$$$$

Note: A similar method to evaluate $I=\int \cos x dx$:

$$\Re(e^{\iota x})=\cos x$$ $$\Longrightarrow \int \cos x dx = \Re \int e^{\iota x} dx$$ $$I=\Re(\dfrac{e^{\iota x}}{\iota}) = -\iota \Re(\cos x +\iota \sin x) = \sin x$$

$\endgroup$
  • $\begingroup$ I came across this method here $\endgroup$ – Ishan Jun 20 '15 at 8:33
  • $\begingroup$ Just to mention: I've never heard anybody refer to or intentionally write $i$ as $\iota$. If you want to do so, you can but I don't recommend it. $\endgroup$ – George V. Williams Jun 30 '15 at 16:12
  • $\begingroup$ I'm really sorry Sir. I had always thought that it was iota and not 'i'. I'll rectify it from now on. Thanks for telling me, Sir. $\endgroup$ – Ishan Jul 1 '15 at 9:10
1
$\begingroup$

A hint:

Subsitute the definition of $\cos$ to have:

$$ I = \int \frac{2}{e^{ix} + e^{-ix}} \, \mathrm{d}x = \frac{2}{i}\int \frac{i e^{ix}}{(e^{ix})^2 + 1} \, \mathrm{d}x = - 2 i \arctan e^{ix}$$

Then, you can use the definition of the inverse tangent when the argument is complex.

PS: Simplification should lead to a real result.

$\endgroup$
  • $\begingroup$ Thanks you Sir, but how would this use $\Im(e^{\iota x})$ or $\Re(e^{\iota x})$? $\endgroup$ – Ishan Jun 20 '15 at 20:18
  • $\begingroup$ Sir, could I ask for one last favour? Sir, could you please evaluate $$\int_0^{\pi/2} \dfrac{\sin^2(x)}{\sin (x) + \cos(x)} dx?$$ I keep getting my answer as $$\dfrac{\ln(\sqrt{2}+1)}{\sqrt{2}}$$However, my book lists a completely different answer. $\endgroup$ – Ishan Jun 20 '15 at 20:20
  • $\begingroup$ I think rewriting the integral as $\int \mathrm{d}x/\mathrm{Re}(e^{ix})$ may not lead you to a simple evaluation of the result. In general, it's not true that: $1/\mathrm{Re}(z) = \mathrm{Re}(1/z)$, for a complex number $z$ so you cannot use your approach anymore. $\endgroup$ – Dmoreno Jun 20 '15 at 20:27
  • 1
    $\begingroup$ Alright Sir. Thanks a lot for your help. Regarding the second problem I think I've understood where I've made a mistake although I still have to check. If I still get it wrong, I'll put itin a new post. $\endgroup$ – Ishan Jun 20 '15 at 20:32
  • 1
    $\begingroup$ Oh, sorry Sir, I forgot to click it. I just accepted it Sir. I'm really really sleepy and am now going to bed (it's past 2 at night here in India). Goodnight Sir.$$$$ PS. I just checked my answer again and it comes out to be correct. There's a misprint in my book due to which my answer had seemed to be wrong. $\endgroup$ – Ishan Jun 20 '15 at 20:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.