PDA

View Full Version : Integrate problem with undefined constant


sukharef
Jul29-11, 11:06 AM
Hello!

I'm trying to calculate this integral:

Integrate [Exp[-I a *y] Exp[I*Sin[y]] Exp[-y*Sign[y]], {y, -Infinity, Infinity}] , where "a" - constant that i'll define later.

So the result of the calculation acсording to Mathematica is the same expression : Exp[-I a *y] Exp[I*Sin[y]] Exp[-y*Sign[y]].

What can i do to calculate my integral without defining my constant "a" and using NIntagrate?

Thank you in advance.

Simon_Tyler
Jul29-11, 07:35 PM
I doubt that Mma returns the integrand - it probably returns an unevaluated integral, no matter what choice you make for the constant a. The integral probably does not have a nice closed form in terms of any common or existing special functions.

Mma doesn't even evaluate the "simpler" integral Exp[I*Sin[y] - y^2].

If you just need to use this result just define the numerical integral (with memoization)

int[a_?NumericQ] := int[a] = NIntegrate[Exp[-I a*y] Exp[I*Sin[y]] Exp[-y*Sign[y]], {y, -Infinity, Infinity}]//Chop

It's a little slow... but you can get a reasonable plot of the integral

Plot[int[x], {x, -10, 10}, PlotPoints -> 15, MaxRecursion -> 3, PlotRange -> All]
http://www.physicsforums.com/attachment.php?attachmentid=37617&stc=1&d=1311986079

sukharef
Jul30-11, 01:32 AM
Thank you so much for answer!

Simon_Tyler
Jul31-11, 12:33 AM
Not a problem. I hope it was useful.
If you do find a closed form to that integral - don't forget to post it here!