How to solve this definite integral?

Click For Summary

Discussion Overview

The discussion revolves around solving a complex definite integral involving the incomplete gamma function and exponential terms. Participants explore various integration techniques and transformations, discussing both the structure of the integral and the methods for evaluating it. The conversation includes technical reasoning and mathematical manipulation, focusing on the integral's properties and potential solutions.

Discussion Character

  • Technical explanation
  • Mathematical reasoning
  • Debate/contested

Main Points Raised

  • One participant presents the integral to be solved and mentions previous attempts using integration by substitution without success.
  • Another participant identifies the integrand as a combination of exponential and incomplete gamma functions, questioning the nature of the two-variable function involved.
  • A participant provides a transformation of the integral, expressing it in terms of a new variable and discussing the integration of the incomplete gamma function.
  • Further elaboration on the integral's structure is provided, including a breakdown of the terms and the introduction of limits for the integration process.
  • One participant expresses difficulty in accessing the thread due to layout issues, indicating a potential barrier to engagement in the discussion.
  • Another participant revisits a previous suggestion and acknowledges that the original poster has already utilized that approach.
  • A participant claims to have reached a final solution for the definite integral, presenting a complex expression that incorporates various mathematical functions and constants.

Areas of Agreement / Disagreement

The discussion features multiple competing views and approaches to solving the integral, with no consensus on a single method or solution. Participants express differing perspectives on the integral's complexity and the effectiveness of various techniques.

Contextual Notes

Participants reference the incomplete gamma function and exponential decay, but the discussion does not resolve the assumptions or limitations inherent in the integral's formulation. There are also unresolved mathematical steps and transformations discussed throughout the thread.

Ad VanderVen
Messages
169
Reaction score
13
TL;DR
How to solve the integral ##\int_{0}^{x}\!-{\frac {\lambda\,{{\rm e}^{-\lambda\,t}}{\beta}^{\alpha} \left( -\lambda+\beta \right) ^{-\alpha} \left( -\Gamma \left( \alpha \right) +\Gamma \left( \alpha, \left( -\lambda+\beta \right) t \right) \right)}{\Gamma \left( \alpha \right) }}\,{\rm d}t##
I would like to solve the integral underneath:

$$\displaystyle \int_{0}^{x}\!-{\frac {\lambda\,{{\rm e}^{-\lambda\,t}}{\beta}^{\alpha} \left( -\lambda+\beta \right) ^{-\alpha} \left( -\Gamma \left( \alpha \right) +\Gamma \left( \alpha, \left( -\lambda+\beta \right) t \right) \right)}{\Gamma \left( \alpha \right) }}\,{\rm d}t$$

I tried it before using Integration by Substitution, but that didn't help either.
 
Last edited:
Physics news on Phys.org
The integrand is written as
Ae^{-\lambda t}+Be^{-\lambda t}\Gamma (\alpha ,Ct)
The first term is easy to integrate. What is two variable function of ##\Gamma(\alpha, Ct)## you wrote ?

[EDIT] It is an incomplete gamma function.
 
Last edited:
The only non-trivial part is <br /> \int_0^x e^{-\lambda t} \Gamma(\alpha, kt)\,dt<br /> = \frac 1k \int_0^{kx} e^{-\mu s}\Gamma(\alpha, s)\,ds<br /> where k = \beta - \lambda and \mu = \lambda/k. Now <br /> \Gamma(\alpha, s) = \int_s^\infty t^{\alpha - 1} e^{-t}\,dt = \Gamma(\alpha) - \int_0^s t^{\alpha - 1} e^{-t}\,dt so <br /> \int_0^{X} e^{-\mu s} \Gamma(\alpha, s)\,ds = \int_0^X e^{-\mu s} \Gamma(\alpha)\,ds -<br /> \int_0^X \int_0^s e^{-\mu s - t} t^{\alpha - 1}\,dt\,ds. The first integral is trivial; in the second we can set (u,v) = (\mu s + t, t) and we find that <br /> \int_0^X \int_0^s e^{-\mu s - t} t^{\alpha - 1}\,dt\,ds = \frac{1}{\mu}\int_0^{X}<br /> \int_{u/(1+ \mu)}^u e^{-u} v^{\alpha - 1}\,dv\,du <br /> + \frac{1}{\mu} \int_X^{(1 + \mu)X} \int_{u/(1 + \mu)}^X e^{-u} v^{\alpha - 1}\,dv\,du and after doing the inner integral over v you will be left with integrals of e^{-u} (trivial) and e^{-u} u^\alpha (expressible in terms of incomplete gamma functions).

EDIT: The limits should be 0 \leq u \leq (1 + \mu)X and \max\{0, u - \mu X\} \leq v \leq u/(1 + \mu). That introduces a term of the form \int_{\mu X}^{(1 + \mu)X} e^{-u} (u - \mu X)^\alpha \,du = e^{-\mu X} \int_0^{X} e^{-z}z^\alpha\,dz.
 
Last edited:
  • Like
Likes   Reactions: anuttarasammyak, ergospherical, fresh_42 and 1 other person
Hi folks, I'm trying to respond to your answers, but I can't read them because of the right column, which prevents me from seeing the whole page. I don't know how to click away that column.
 
pasmith said:
The only non-trivial part is <br /> \int_0^x e^{-\lambda t} \Gamma(\alpha, kt)\,dt<br /> = \frac 1k \int_0^{kx} e^{-\mu s}\Gamma(\alpha, s)\,ds<br /> where k = \beta - \lambda and \mu = \lambda/k.

This can be tidied up considerably: <br /> \begin{align*}<br /> \frac{1}{k}\int_0^{kx} e^{-\mu t} \Gamma(\alpha, t)\,dt &amp;= \frac 1k \int_0^{kx} \int_t^\infty e^{-\mu t-s}s^{\alpha - 1}\,ds\,dt \\<br /> &amp;= \frac 1{\mu k} \int_0^\infty v^{\alpha - 1} \int_v^{U(v)} e^{-u}\,du\,dv \\<br /> &amp;= \frac 1{\mu k} \int_0^\infty v^{\alpha - 1} (e^{-v} - e^{-U(v)})\,dv<br /> \end{align*} using (u,v) = (\mu t+s,s) where U(v) = \min\{(1 + \mu)v, v + \mu kx\}. Hence <br /> \begin{align*}<br /> \frac{1}{k} \int_0^{kx} e^{-\mu t} \Gamma(\alpha, t)\,dt &amp;= \frac{\Gamma(\alpha)}{\mu k}<br /> - \frac{1}{\mu k}\int_0^{kx} v^{\alpha - 1}e^{-(1 + \mu)v}\,dv <br /> - \frac{e^{-\mu kx}}{\mu k} \int_{kx}^\infty v^{\alpha-1} e^{-v}\,dv \\<br /> &amp;= \frac{\Gamma(\alpha)}{\mu k} - \frac{\gamma(\alpha, (1+\mu)kx)}{\mu k(1 + \mu)^\alpha} <br /> - \frac{e^{-\mu kx} \Gamma(\alpha,kx)}{\mu k}.<br /> \end{align*}
 
pasmith

For me the non-trivial part is:

$$\displaystyle \int_{0}^{x}\!{\frac {\Gamma \left( \alpha,\beta\,t-\lambda\,t \right) }{{{\rm e}^{\lambda\,t}}}}\,{\rm d}t$$

Please, would you be so kind to start from this equation.
 
pasmith

Sorry for my earlier reply. Now that I've looked again, I've seen that you've actually started from the equation I've suggested.
 
pasmith

With your help I got the final solution for the definite integral:

$$\displaystyle {\frac {{\beta}^{\alpha} \left( {{\rm e}^{\lambda\,x}}-1 \right) {{\rm e}^{-\lambda\,x}}}{ \left( -\lambda+\beta \right) ^{\alpha}}}-{\frac {{\beta}^{\alpha}}{\alpha\, \left( 1+\alpha \right) \left( -\lambda+\beta \right) ^{\alpha}\Gamma \left( \alpha \right) } \left( -{{\rm e}^{-x\beta}} \left( {\frac {\beta}{-\lambda+\beta}} \right) ^{-\alpha} \left( x\beta \right) ^{\alpha}\alpha- \left( x\beta \right) ^{\alpha/2}{{\rm e}^{-1/2\,x\beta}}{{WhittakerM}_{\alpha/2,\,\alpha/2+1/2}\left(x\beta\right)} \left( {\frac {\beta}{-\lambda+\beta}} \right) ^{-\alpha}-{{\rm e}^{-x\beta}} \left( {\frac {\beta}{-\lambda+\beta}} \right) ^{-\alpha} \left( x\beta \right) ^{\alpha}-{{\rm e}^{-\lambda\,x}}\Gamma \left( \alpha, \left( -\lambda+\beta \right) x \right) {\alpha}^{2}+\Gamma \left( \alpha \right) {\alpha}^{2}-{{\rm e}^{-\lambda\,x}}\Gamma \left( \alpha, \left( -\lambda+\beta \right) x \right) \alpha+\Gamma \left( \alpha \right) \alpha \right) }$$
 

Similar threads

  • · Replies 14 ·
Replies
14
Views
3K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 16 ·
Replies
16
Views
2K
  • · Replies 12 ·
Replies
12
Views
3K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 2 ·
Replies
2
Views
1K
  • · Replies 1 ·
Replies
1
Views
3K
Replies
46
Views
5K
  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 3 ·
Replies
3
Views
3K