Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

show that the sequence-indexed with $a_n$ , $${1\over{1+t^2}} - {e^{-ta_n}\over{(1+t^2)}}(\cos a_n + t\sin a_n)$$ is bounded from above by an integrable function for a sufficiently large $a_n$ ($\lim_{n\to\infty}a_n= + \infty$)

i tried this :

$(\cos a_n+t\sin a_n)\ge(-1-t)$

$1-e^{-ta_n}(\cos a_n+t\sin a_n)\le1+ {(1+t)e^{-ta_n}}$

${1\over{1+t^2}}(1-e^{-ta_n}(\cos a_n+t\sin a_n))\le{1\over{1+t^2}}(1+ {(1+t)e^{-ta_n}})$

and $g_n(t)={1\over{1+t^2}}(1+ {(1+t)e^{-ta_n}})$ is integrable ?

share|improve this question
    
That's true. But is it terribly clear why $g_n(t)$ is integrable? (BTW, there's a missing minus sign in the final line) –  user108903 Jan 19 '13 at 0:49
    
for me , i didnt understand the concept of "an integrable function for a sufficiently large $a_n$" , i just wanted to prove that a bounded integrable function exist , even if my solution is correct it is not complete ! the teacher solution goes to calculate the limit of $g_n$ then replacing $te^{-ta_n}$ with an $\varepsilon \, \forall n \ge N$ , i didn't understand why $te^{u}$ form and not $e^{u}$ and why he did all of that ! –  Lofaif Jan 19 '13 at 1:10
    
It sounds like the $\epsilon$ business is directed towards an application of the dominated convergence theorem in order to show that the limit of the integrals is zero. –  user108903 Jan 19 '13 at 1:20
    
yeah correct ! , he worked in his solution with $f_n(t)={1\over{1+t^2}}e^{-ta_n}(cosa_n+tsina_n)$ and he said : lets prove that : $\int_0^\infty f_n(t) dt=0$ .. and he proved that but our sequence is different , i didnt understand why he cut the sequence into two part , and i just tried to prove it as above ! –  Lofaif Jan 19 '13 at 1:34
add comment

1 Answer

up vote 0 down vote accepted

Consider $f_n(t)=\dfrac1{1+t^2} -\dfrac {\mathrm e^{-ta_n}}{1+t^2}(\cos a_n + t\sin a_n)$. The goal is not to find some integrable functions $(g_n)$ such that $f_n\leqslant g_n$, presumably on $(0,+\infty)$, for every $n$ large enough, but to find an integrable function $g$ such that $f_n\leqslant g$ on $(0,+\infty)$ for every $n$ large enough.

To do so, let $N$ such that $a_n\geqslant1$ for every $n\geqslant N$ and note that $\cos a_n + t\sin a_n\geqslant-(1+t)\geqslant-2(1+t^2)$ and that $\mathrm e^{-a_n t}\leqslant\mathrm e^{-t}$ for every $n\geqslant N$, hence $$ f_n(t)\leqslant\frac1{1+t^2}+2\mathrm e^{-t}=g(t), $$ and $g$ is obviously integrable on $(0,+\infty)$.

If, as I suspect, the goal of the exercise is actually to find an integrable function $g$ such that $|f_n|\leqslant g$ on $(0,+\infty)$ for every $n$ large enough, use $|\cos a_n + t\sin a_n|\leqslant1+t\leqslant2(1+t^2)$ $$ |f_n(t)|\leqslant\frac1{1+t^2}+2\mathrm e^{-t}=g(t). $$ Finally, note that somewhat more refined arguments allow to show that, irrespectively of the value of $a_n$ and for every $n$, for every $t\geqslant0$, $$ |f_n(t)|\leqslant\frac3{1+t^2}. $$

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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