Group Abstract

Message Boards

WOLFRAM COMMUNITY

4.3K Views

3 Replies

4 Total Likes

View groups...

Follow this post

Share this post:

GROUPS:

Mathematics Calculus Wolfram Language

Problem in integrating Erfc function

Seb Fig

Posted 4 years ago

Hello folks, I'm having some troubles on the analytical integration over the variable z of this function using Mathematica: z^6 Erfc[b \[Rho] - d \[Rho] - a n \[Rho] - c z \[Rho] - a m z \[Rho]] With the following assumptions: Assumptions -> \[Rho] > 0 && a \[Element] Reals && b \[Element] Reals && c \[Element] Reals && d \[Element] Reals && m \[Element] Reals && n \[Element] Reals] Thus, I use the following command: Integrate[ z^6 Erfc[ b \[Rho] - d \[Rho] - a n \[Rho] - c z \[Rho] - a m z \[Rho]], z, Assumptions -> \[Rho] > 0 && a \[Element] Reals && b \[Element] Reals && c \[Element] Reals && d \[Element] Reals && m \[Element] Reals && n \[Element] Reals] I can't understand why Mathematica cannot integrate this function, while if I lower the degree of *z^6* the integration is performed without any problems. For instance, this function gets integrated fine and smoothly over z and using the same assumptions: z^5 Erfc[b \[Rho] - d \[Rho] - a n \[Rho] - c z \[Rho] - a m z \[Rho]] Thank you for the time spent on reading this. Looking forward to your help! S

Hello folks,

I'm having some troubles on the analytical integration over the variable z of this function using Mathematica:

     z^6 Erfc[b \[Rho] - d \[Rho] - a n \[Rho] - c z \[Rho] - a m z \[Rho]]

With the following assumptions:

Assumptions -> \[Rho] > 0 && a \[Element] Reals && b \[Element] Reals && c \[Element] Reals && d \[Element] Reals && m \[Element] Reals && n \[Element] Reals]

Thus, I use the following command:

Integrate[
 z^6 Erfc[
   b \[Rho] - d \[Rho] - a n \[Rho] - c z \[Rho] - 
    a m z \[Rho]], z, 
 Assumptions -> \[Rho] > 0 && a \[Element] Reals && 
   b \[Element] Reals && c \[Element] Reals && d \[Element] Reals && 
   m \[Element] Reals && n \[Element] Reals]

I can't understand why Mathematica cannot integrate this function, while if I lower the degree of z^6 the integration is performed without any problems. For instance, this function gets integrated fine and smoothly over z and using the same assumptions:

z^5 Erfc[b \[Rho] - d \[Rho] - a n \[Rho] - c z \[Rho] - a m z \[Rho]]

Thank you for the time spent on reading this. Looking forward to your help!

POSTED BY: Seb Fig

3 Replies

Sort By:

Mariusz Iwaniuk

Mariusz Iwaniuk, engineer, math hobbyist.

Posted 4 years ago

A simple workaround: Integrate[z^6 Erfc[A1 + A2 z + A3 z], z] /. A1 -> b ρ - d ρ - a n ρ /. A2 -> -c * ρ /. A3 -> -amρ // Simplify (1/56 ((1/((c + a m)^7 Sqrt[π] ρ^7)) 2 E^(-(-b + d + a n + c z + a m z)^2 ρ^2) (24 + 87 (-b + d + a n)^2 ρ^2 + 40 (-b + d + a n)^4 ρ^4 + 4 (-b + d + a n)^6 ρ^6 - 4 (c + a m)^5 (-b + d + a n) z^5 ρ^6 + 4 (c + a m)^6 z^6 ρ^6 + 4 (c + a m)^4 z^4 ρ^4 (3 + (-b + d + a n)^2 ρ^2) + 2 (c + a m)^3 (b - d - a n) z^3 ρ^4 (11 + 2 (-b + d + a n)^2 ρ^2) + 2 (c + a m)^2 z^2 ρ^2 (12 + 15 (-b + d + a n)^2 ρ^2 + 2 (-b + d + a n)^4 ρ^4) + (c + a m) (b - d - a n) z ρ^2 (57 + 36 (-b + d + a n)^2 ρ^2 + 4 (-b + d + a n)^4 ρ^4)) - ( 8 (-b + d + a n)^7 Erf[(b - d - a n - c z - a m z) ρ])/(c + a m)^7 - ( 105 (-b + d + a n) Erf[(b - d - a n - c z - a m z) ρ])/((c + a m)^7 ρ^6) - ( 210 (-b + d + a n)^3 Erf[(b - d - a n - c z - a m z) ρ])/((c + a m)^7 ρ^4) - ( 84 (-b + d + a n)^5 Erf[(b - d - a n - c z - a m z) ρ])/((c + a m)^7 ρ^2) + 8 z^7 Erfc[(b - d - a n - c z - a m z) ρ]) )

A simple workaround:

 Integrate[z^6 Erfc[A1 + A2 z + A3 z], z] /. A1 -> b ρ - d ρ - a n ρ /. A2 -> -c * ρ /. A3 -> -a*m*ρ // Simplify

 (*1/56 ((1/((c + a m)^7 Sqrt[π] ρ^7))
    2 E^(-(-b + d + a n + c z + a m z)^2 ρ^2) (24 + 
       87 (-b + d + a n)^2 ρ^2 + 40 (-b + d + a n)^4 ρ^4 + 
       4 (-b + d + a n)^6 ρ^6 - 
       4 (c + a m)^5 (-b + d + a n) z^5 ρ^6 + 
       4 (c + a m)^6 z^6 ρ^6 + 
       4 (c + a m)^4 z^4 ρ^4 (3 + (-b + d + a n)^2 ρ^2) + 
       2 (c + a m)^3 (b - d - a n) z^3 ρ^4 (11 + 
          2 (-b + d + a n)^2 ρ^2) + 
       2 (c + a m)^2 z^2 ρ^2 (12 + 15 (-b + d + a n)^2 ρ^2 + 
          2 (-b + d + a n)^4 ρ^4) + (c + a m) (b - d - 
          a n) z ρ^2 (57 + 36 (-b + d + a n)^2 ρ^2 + 
          4 (-b + d + a n)^4 ρ^4)) - (
    8 (-b + d + a n)^7 Erf[(b - d - a n - c z - a m z) ρ])/(c + 
      a m)^7 - (
    105 (-b + d + a n) Erf[(b - d - a n - c z - a m z) ρ])/((c + 
       a m)^7 ρ^6) - (
    210 (-b + d + a n)^3 Erf[(b - d - a n - c z - a m z) ρ])/((c +
        a m)^7 ρ^4) - (
    84 (-b + d + a n)^5 Erf[(b - d - a n - c z - a m z) ρ])/((c + 
       a m)^7 ρ^2) + 
    8 z^7 Erfc[(b - d - a n - c z - a m z) ρ])
 *)

POSTED BY: Mariusz Iwaniuk

W. Craig Carter

W. Craig Carter, MIT

Posted 4 years ago

The WL language may miss some integrals. It happens. For this one, you could try: sol = SolveValues[ b \[Rho] - d \[Rho] - a n \[Rho] - c z \[Rho] - a m z \[Rho] == \[Phi], z][[1]] cov = D[sol, \[Phi]] integral = cov (-(1/((c + a m) \[Rho]))) Integrate[(( b \[Rho] - d \[Rho] - a n \[Rho] - \[Phi])/((c + a m) \[Rho]))^6 Erf[\[Phi]], \[Phi]] gives a closed form. checking: dintegral = D[integral /. \[Phi] -> b \[Rho] - d \[Rho] - a n \[Rho] - c z \[Rho] - a m z \[Rho], z]; Simplify[dintegral] (z^6 Erf[(b - d - a n - c z - a m z) \[Rho]])

The WL language may miss some integrals. It happens. For this one, you could try:

sol = SolveValues[
   b \[Rho] - d \[Rho] - a n \[Rho] - c z \[Rho] - 
     a m z \[Rho] == \[Phi], z][[1]]

cov = D[sol, \[Phi]]

integral = cov (-(1/((c + a m) \[Rho]))) Integrate[((
     b \[Rho] - d \[Rho] - 
      a n \[Rho] - \[Phi])/((c + a m) \[Rho]))^6 Erf[\[Phi]], \[Phi]]

gives a closed form.

checking:

dintegral = 
  D[integral /. \[Phi] -> 
     b \[Rho] - d \[Rho] - a n \[Rho] - c z \[Rho] - a m z \[Rho], z];

Simplify[dintegral]  (*z^6 Erf[(b - d - a n - c z - a m z) \[Rho]]*)

POSTED BY: W. Craig Carter

Seb Fig

Posted 4 years ago

Thanks a lot for your reply! The integral gives indeed a closed form. I just miss the part where the variable "*cov*" seems being squared before the integral. Why is that? cov^2 Integrate[(( b \[Rho] - d \[Rho] - a n \[Rho] - \[Phi])/((c + a m) \[Rho]))^6 Erf[\[Phi]], \[Phi]] Since: cov=-(1/((c + a m) \[Rho])) Inspired by your answer, I would have set the integral this way: z=((b \[Rho] - d \[Rho] - a n \[Rho] - \[Phi])/((c + a m) \[Rho])) cov = dz/d\[Phi] = -(1/((c + a m) \[Rho])) Integrate[((b \[Rho] - d \[Rho] - a n \[Rho] - \[Phi])/((c + a m) \[Rho]))^6 Erf[\[Phi]] (-(1/((c + a m) \[Rho]))), \[Phi]] Am I right or am I missing something?

Thanks a lot for your reply! The integral gives indeed a closed form.

I just miss the part where the variable "cov" seems being squared before the integral. Why is that?

cov^2 Integrate[((
     b \[Rho] - d \[Rho] - 
      a n \[Rho] - \[Phi])/((c + a m) \[Rho]))^6 Erf[\[Phi]], \[Phi]]

Since:

cov=-(1/((c + a m) \[Rho]))

Inspired by your answer, I would have set the integral this way:

z=((b \[Rho] - d \[Rho] - a n \[Rho] - \[Phi])/((c + a m) \[Rho]))
cov = dz/d\[Phi] = -(1/((c + a m) \[Rho]))

Integrate[((b \[Rho] - d \[Rho] - a n \[Rho] - \[Phi])/((c + a m) \[Rho]))^6 Erf[\[Phi]] (-(1/((c + a m) \[Rho]))), \[Phi]]

Am I right or am I missing something?

POSTED BY: Seb Fig

Reply to this discussion

Reply Preview

Attachments

Remove Add a file to this post

Follow this discussion

or Discard

Feedback