Here is my question:

Starting from the relation $$\int_{0}^{+\infty}t^{a-1}e^{-nt}\,dt=n^{-a}\Gamma(a)\qquad a>0$$ and differentiating $m-$times under the integral sign we can get to

$$\int_{0}^{+\infty}t^{a-1}\ln^mte^{-nt}\,dt=n^{-a}\sum_{t=0}^{m}(-1)^{m-r}\binom{m}{r}\Gamma^{(r)}(a)\ln^{m-r}n\quad a>0,\;m=0,1,2,\ldots\quad (1)$$ where $\Gamma^{(r)}(a)$ means the $r-$th derivative of $\Gamma$ function at $a$.

Is there a way to reach to similar relation for $\mathbb N\not\ni m>0$ ? (with the $\ln$ in absolute value in this case) The relation doesn't have to be an "exact" result (I haven't find it in any table of Laplace transforms) but I would expect an asymptotic estimate for $n\to+\infty$ or sth could be found.

First thought was to write the above sum in $(1)$ as \begin{align}n^{-a}\sum_{r=0}^{m}(-1)^r\binom{m}{r}\Gamma^{(m-r)}(a)\ln^rn&=n^{-a}\sum_{r=0}^{m}\binom{-m+r-1}{r}\Gamma^{(m-r)}(a)\ln^rn\notag \\&=n^{-a}\sum_{r=0}^{m}\binom{-r-1}{m-r}\Gamma^{(r)}(a)\ln^{m-r}n\notag \\&=n^{-a}\sum_{r=0}^{+\infty}\binom{-r-1}{m-r}\Gamma^{(r)}(a)\ln^{m-r}n\notag \\&=n^{-a}\sum_{r=0}^{+\infty}\frac{\Gamma(-r)}{\Gamma(-m)\Gamma(m-r+1)}\Gamma^{(r)}(a)\ln^{m-r}n\notag\end{align}

and try to find an argument to pass from the validity of the equality on the positive integers to positive reals but I didn't succeed.

Second thought was to start doing repeated integration by parts on LHS of $(1)$. This way I managed to get the first summand but things get really messy and I cannot proceed to get the derivatives of $\Gamma$. Any thoughts?

Thank you very much in advance,



Your Answer

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

Browse other questions tagged or ask your own question.