From: SourceForge.net <no...@so...> - 2009-03-15 21:06:49
|
Bugs item #2687962, was opened at 2009-03-15 22:06 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=104933&aid=2687962&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: Open Resolution: None Priority: 5 Private: No Submitted By: Dieter Kaiser (crategus) Assigned to: Nobody/Anonymous (nobody) Summary: hgfred([-3/2,1],[-1/2],-t) division by zero Initial Comment: Working on the Laplace transfom for the Exponential Integral E, I get a problem with half integral values for the order of the Exponential Integral E. This is one of the hypergeometric functions needed by the algorithm: %i17) hgfred([-3/2,1],[-1/2],-t); Division by 0 -- an error. To debug this try debugmode(true); It works for positive half integral values and a=-1/2 but not for a=-3/2, a=-5/2, a=-7/2, ... This would be a solution for a=-3/2,b=1,c=-1/2 (see wolfram.function.com): hgfred([-3/2,1],[-1/2],t) = 1+3*t-3*t^(3/2)*atan(sqrt(t)) Dieter Kaiser ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=104933&aid=2687962&group_id=4933 |
From: SourceForge.net <no...@so...> - 2009-03-20 03:41:09
|
Bugs item #2687962, was opened at 2009-03-15 17:06 Message generated for change (Comment added) made by rtoy You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=104933&aid=2687962&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 Private: No Submitted By: Dieter Kaiser (crategus) Assigned to: Nobody/Anonymous (nobody) Summary: hgfred([-3/2,1],[-1/2],-t) division by zero Initial Comment: Working on the Laplace transfom for the Exponential Integral E, I get a problem with half integral values for the order of the Exponential Integral E. This is one of the hypergeometric functions needed by the algorithm: %i17) hgfred([-3/2,1],[-1/2],-t); Division by 0 -- an error. To debug this try debugmode(true); It works for positive half integral values and a=-1/2 but not for a=-3/2, a=-5/2, a=-7/2, ... This would be a solution for a=-3/2,b=1,c=-1/2 (see wolfram.function.com): hgfred([-3/2,1],[-1/2],t) = 1+3*t-3*t^(3/2)*atan(sqrt(t)) Dieter Kaiser ---------------------------------------------------------------------- >Comment By: Raymond Toy (rtoy) Date: 2009-03-19 23:41 Message: To compute the function Maxima was transforming F(-1/2,1;1/2;z) to F(-1/2,1;-1/2;z) and applying A&S 15.2.5. But a = c which causes a division by zero in that formula. Change implementation to handle the case where we want to adjust both a and c by the same amount. Use A&S 15.2.7 and 15.2.8 to do that. This fixes the problem. Fixed in hyp.lisp, rev 1.102 And for the record the correct answer has atanh, not atan. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=104933&aid=2687962&group_id=4933 |