0
$\begingroup$
Integrate[(1 - (-Gamma[γ/ξ, -((p λ)/ξ)] + 
 Gamma[γ/ξ, (p (-1 + z) λ)/ξ])/(
Gamma[γ/ξ] - Gamma[γ/ξ, -((
   p λ)/ξ)])) z^(μ/ξ1) (1 - 
 z)^(η/ξ1 - γ/ξ - 1) Exp[λ*p*
z (-(1/ξ1) + 1/ξ)] , {z, 0, 1}, 
Assumptions -> ∞ > η > 0 && ∞ > γ > 
0 && ∞ > ξ > 0 && ∞ > λ > 
0 && ∞ > μ > 0 && ∞ > p > 
0 && ∞ > ξ1 > 0 && 1 >= z >= 0]

I am trying to integrate the above function, Mathematica tool is not any showing error. Is the above function is divergent? Could you please suggest.

$\endgroup$
5
  • $\begingroup$ Crossposted here. $\endgroup$ Commented May 25, 2020 at 19:22
  • $\begingroup$ Looking at a simpler integral: this will work provided $f,g \in\mathbb{Z^+}$ and they are small. $$e^z z^f (1-z)^g \left(1-\frac{\Gamma (a,b (z-1))-\Gamma (a,b)}{\Gamma (a)-\Gamma (a,b)}\right)$$ $\endgroup$
    – flinty
    Commented May 25, 2020 at 20:28
  • $\begingroup$ @flinty ,I tried to run below code ,not getting any results ,Could you suggest what is wrong ,Integrate[ E^z (1 - z)^f z^ g (1 - (-Gamma[a, b] + Gamma[a, b (-1 + z)])/( Gamma[a] - Gamma[a, b])), {z, 0, 1}, Assumptions -> f < 10 && g < 10] $\endgroup$
    – Pramod
    Commented May 26, 2020 at 5:05
  • $\begingroup$ This is not a convergence error because there is nothing to converge here---a symbolic integration is requested. Not every symbolic integral can be computed. $\endgroup$
    – yarchik
    Commented May 26, 2020 at 8:50
  • $\begingroup$ In my code above you need to replace $f,g$ with integer values. $\endgroup$
    – flinty
    Commented May 26, 2020 at 10:59

1 Answer 1

1
$\begingroup$

It is unlikely that the integral can be computed for any values of parameters. However, in one special nontrivial case of $\gamma=\xi$ the integrand takes a simpler form

f=(1-(-Gamma[γ/ξ,-((p λ)/ξ)]+Gamma[γ/ξ,(p (-1+z) λ)/ξ])/(Gamma[γ/ξ]-Gamma[γ/ξ,-((p λ)/ξ)])) z^(μ/ξ1) (1-z)^(η/ξ1-γ/ξ-1) Exp[λ*p*z (-(1/ξ1)+1/ξ)]/.γ->ξ//FullSimplify

$$\frac{(1-z)^{\frac{\eta }{\text{$\xi $1}}-2} z^{\mu /\text{$\xi $1}} \left(e^{\frac{\lambda p}{\xi }}-e^{\frac{\lambda p z}{\xi }}\right) e^{-\frac{\lambda p z}{\text{$\xi $1}}}}{e^{\frac{\lambda p}{\xi }}-1},$$

and the integral can be computed

Integrate[f,{z,0,1},Assumptions->∞>η>0&&∞>ξ>0&&∞>λ>0&&∞>μ>0&&∞>p>0&&∞>ξ1>0&&1>=z>=0]

$$\frac{\Gamma \left(\frac{\eta }{\text{$\xi $1}}-1\right) \Gamma \left(\frac{\mu +\text{$\xi $1}}{\text{$\xi $1}}\right) \left(e^{\frac{\lambda p}{\xi }} \, _1\tilde{F}_1\left(\frac{\mu +\text{$\xi $1}}{\text{$\xi $1}};\frac{\eta +\mu }{\text{$\xi $1}};-\frac{p \lambda }{\text{$\xi $1}}\right)-\, _1\tilde{F}_1\left(\frac{\mu +\text{$\xi $1}}{\text{$\xi $1}};\frac{\eta +\mu }{\text{$\xi $1}};p \lambda \left(\frac{1}{\xi }-\frac{1}{\text{$\xi $1}}\right)\right)\right)}{e^{\frac{\lambda p}{\xi }}-1}$$ for $\eta >\text{$\xi $1}$.

$\endgroup$
3
  • $\begingroup$ Thanks, the gamma and Xi have a different meaning in the model in which I am currently working. why it could the general integration does not exist, do you have any suggestions. $\endgroup$
    – Pramod
    Commented May 26, 2020 at 8:20
  • 1
    $\begingroup$ @Pramod_achar This is not a bug, this is not a convergence error that general integration does not exist. This is a very typical situation. Most of the integrals cannot be computed. P.S. I am wondering why you write gamma small and Xi capital, shouldn't it be "xi"? $\endgroup$
    – yarchik
    Commented May 26, 2020 at 8:48
  • 1
    $\begingroup$ Thanks, not noticed previously Xi, I wrote it while I need to write in Mathematica [Xi]. $\endgroup$
    – Pramod
    Commented May 26, 2020 at 10:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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