From: SourceForge.net <noreply@so...>  20060328 04:57:19

Bugs item #1451351, was opened at 20060316 10:11 Message generated for change (Comment added) made by rtoy You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=104933&aid=1451351&group_id=4933 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: None Group: None >Status: Closed >Resolution: Fixed Priority: 5 Submitted By: Raymond Toy (rtoy) Assigned to: Nobody/Anonymous (nobody) Summary: integrate(x*exp((kk+1)*x)/(exp(x)+3),x,minf,inf) seems wrong Initial Comment: assume(kk<0,kk+1>0); integrate(x*exp((kk+1)*x)/(exp(x)+3),x,minf,inf); Answer kk is not an integer. If you exponentialize and factor the result of the integral you get (log(3)^2+%pi^2)/2*exp(log(3)*kk+%i*%pi*kk) Since 1 < kk < 0, exp(%i*%pi*kk) is complex. Hence, the integral of a realvalued function is complex. That's wrong. This integral uses rectzto%pi2 to evaluate it, which is a rectangle of height 2*%pi. The algorithm seems correct, but the result is clearly wrong.  >Comment By: Raymond Toy (rtoy) Date: 20060327 23:57 Message: Logged In: YES user_id=28849 This should be fixed in CVS. rectzto%pi2 no longer succeeds on this integral because the exponential part is not a rational function of exp(x). The code has been enhanced to convert this integral to another form using the substitution x=log(y), which results in an integral that maxima knows how to handle. Closing this bug.  You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=104933&aid=1451351&group_id=4933 