I have managed to solve this integral by using Taylor series expansion to approximate the $e^x=1+x$. However, I am not successful due to the integral is not converged.
Could you please give me a hint to simplify this one or express it in other function or terms. \begin{align} I= \int_0^{\infty } \frac{\text{exp}[-a x]}{(1+b x)^{p+2}(1+c x)^q} \, dx \end{align} where $a>0$, $b>0$, $c>0$ $p\ge 0$, $q \ge 0$.
Thank you very much!
You can't approximate $e^t\approx1+t$, since this is only valid for $t\approx0$. But, in this case, $t\in(0,\infty)$, since the upper integration limit is $\infty$. If the interval of integration would've been $(u,v)\subset(-1,1)$ instead of $(0,\infty)$, then indeed the approximation could have proven useful, but, as it stands, it is useless.
Now as to the actual integral: Begin first by evaluating $I(B,C)=\displaystyle\int_0^\infty\frac{e^{-ax}}{(B+bx)(C+cx)}~dx$,
which can be done by decomposing the integrand into partial fractions, and appealing to the definitions of certain special functions, such as the incomplete $\Gamma$ function, or the exponential integral. Then notice that for integer values of p and q, the integral can be expressed in terms of $I^{(p+1,~q-1)}(1,1)$. Alternately, split the interval of integration into $(0,1)$ and $(1,\infty)$, let $t=1/x$ on the latter, and expand the integrand into its binomial series, then reverse the order of summation and integration, and rewrite the new expression on terms of hypergeometric functions.