3
$\begingroup$

I'm beginning my way through Coddington's Intro to ODE's and I'm a little thrown off in the preliminary section in a proof regarding complexed valued functions. ( I should note that I've taken a course in ODES, but my background in Complex anything-besides-the-basics is subpar. )

Particularly the book shows that:

$$ \left| \int_a^b f(x) \, dx \right| \leq \int_a^b \left| f(x) \right| \, dx $$

The proof is pretty short, so I guess I'll just map it out until the me-thrown-off point.

First, let

$$ F \, = \, \int_a^b f(x) \, dx \quad $$

and

$$ u = \cos(\theta) + i\sin(\theta). $$

Then, let $ F \, = \, \left| F \,\right| u $ where $ F \neq 0 $.

Since $u\overline{u} = 1$,

$$ \left| F \right| = \,\overline{u} F \, = \;\overline{u} \int_a^b f(x) \, dx = \;...$$

and the step that loses me:

$$ ...\; = \, Re\left[ \; \overline{u} \int_a^b f(x) \, dx \; \right] \, = \; ... $$

Maybe my unfamiliarity with complex-valued functions is making me miss something obvious, but I'm stumped. For instance, I've tried expanding $ \overline{u} \int_a^b f(x) \, dx $ to:

$$ ( \cos(\theta) - i \sin(\theta) ) \int_a^b f(x) \, dx $$ $$ = \; \cos(\theta) \int_a^b f(x) \, dx - i sin(\theta) \int_a^b f(x) \, dx $$ $$ = \; \cos(\theta) \left( \int_a^b \left ( Re \, f \, \right) (x) \, dx + i \int_a^b \left ( Im \, f \, \right) (x) \, dx \right)- i sin(\theta) \left( \int_a^b \left ( Re \, f \, \right) (x) \, dx + i \int_a^b \left ( Im \, f \, \right) (x) \, dx \right) $$ $$ = \; \cos(\theta) \int_a^b \left ( Re \, f \, \right) (x) \, dx + i \cos(\theta) \int_a^b \left ( Im \, f \, \right) (x) \, dx - i sin(\theta) \int_a^b \left ( Re \, f \, \right) (x) \, dx + sin(\theta) \int_a^b \left ( Im \, f \, \right) (x) \, dx $$ $$ = \; \overline{u} \int_a^b \left( Re \, f \right) (x) \, dx + \left( i \cos(\theta) + \sin(\theta)\right) \int_a^b \left ( Im \, f \, \right) (x) \, dx $$

I felt as if I was on the right track but I hit a wall at this point, and it began to feel like I was convoluting something simple. Anyway the proof finishes with:

$$ ... \; = \; \int_a^b Re\left[ \overline{u} f(x) \right] \, dx \; \leq \int_a^b \left| f(x) \right| \, dx $$

Any help is appreciated :)

$\endgroup$
  • $\begingroup$ Maybe this is not really answering your question (So I put it in comments, but if you let $b$ be a function of $x$, and differentiate both sides, then you get: $$\operatorname{sgn}\int_a^b f(x) \mbox{d}x\mbox{ }f(x)\leq|f(x)|$$, which is obvious $\endgroup$ – Abhimanyu Pallavi Sudhir Jun 9 '13 at 3:07
  • $\begingroup$ Related: math.stackexchange.com/questions/2090022 $\endgroup$ – Watson Nov 27 '18 at 14:55
4
$\begingroup$

By definition, the quantity $|F|$ is real. Since $$|F| = \overline{u}\int_a^bf(x)\,dx,$$ it follows that $$Re\left[\overline{u}\int_a^b f(x)\,dx\right] = Re\,|F| = |F| = \overline{u}\int_a^b f(x)\,dx.$$

$\endgroup$
  • $\begingroup$ So $ Re[ ... ] $ is shown to reiterate that is it real, so that the next step logically follows. Hmm I think I got it. Thanks! $\endgroup$ – Chester May 31 '12 at 0:40
1
$\begingroup$

The real part is inserted not for emphasis, but so you can use the bound $\Re (z) \leq |z|$, vis-à-vis:

$$|\int_a^bf(x)\,dx| = \overline{u}\int_a^bf(x)\,dx = \int_a^b \overline{u} f(x)\,dx = \Re (\int_a^b \overline{u} f(x)\,dx)$$

(Sorry, I don't know how to continue the chain properly...)

$$\Re (\int_a^b \overline{u} f(x)\,dx) = \int_a^b \Re (\overline{u} f(x))\,dx \leq \int_a^b |\overline{u} f(x)|\,dx = \int_a^b |f(x)|\,dx.$$

$\endgroup$
  • $\begingroup$ Ah ok. Now it makes perfect sense. The world is momentarily right again. Thank you. I guess I got hung up on trying to explicity show that $$ \overline{u} \int_a^b f(x) \, dx $$ only contains real parts when it wasn't really necessary since the derivation starts with $$ \left| \int_a^b f(x) \, dx \right| $$ D'oh! $\endgroup$ – Chester May 31 '12 at 11:00
  • $\begingroup$ Happens to all of us. $\endgroup$ – copper.hat May 31 '12 at 15:17

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.