Menu

#3037 Integrate should not be null

None
closed
nobody
integrate (144)
5
2022-12-23
2015-10-20
Samuel
No
integrate(x*sin(x)*exp(-1/2x^2),x,minf,inf);

is zero whereas by ipp we get

integrate(cos(x)*exp(-1/2x^2),x,minf,inf);

which is
[\frac{\sqrt{2}\,\sqrt{\pi }}{\sqrt{e}}]

wxMaxima version: 13.4.0
Maxima version: 5.32.1
Maxima build date: 2014-01-10 01:52:55
Host type: x86_64-unknown-linux-gnu
Lisp implementation type: GNU Common Lisp (GCL)
Lisp implementation version: GCL 2.6.10

Discussion

  • Samuel

    Samuel - 2015-10-20

    Sorry : ipp stands for integration by parts.

     
  • Raymond Toy

    Raymond Toy - 2015-10-20

    Update your version of maxima. Current git maxima says

    (%i1) integrate(x*sin(x)*exp(-1/2*x^2),x,minf,inf);
    (%o1) (sqrt(2)*sqrt(%pi))/sqrt(%e)
    
     
  • sslavi

    sslavi - 2015-10-21

    But the current SBCL Maxima says "zero" again...

     
  • Raymond Toy

    Raymond Toy - 2015-10-21

    Don't know what's up with that. Perhaps an issue with sbcl on windows?

    I just checked out maxima 5.37.2 and the test on my linux box with cmucl, gcl, and sbcl. They all return the correct answer.

     
  • sslavi

    sslavi - 2015-10-22

    Very possibly so.
    Is it possible for you guys to release the GCL version of Maxima 5.37.2 for Windows too?
    SBCL seem to be related to a number of bug reports and CLISP is way too slow for complex probles.

     
  • Andrej Vodopivec

    Do you have any packages loaded. I can't reproduce this on Windows:

    (%i1) wxbuild_info()$
    wxMaxima version: 15.8.2
    Maxima version: 5.37.2
    Maxima build date: 2015-09-18 21:03:17
    Host type: i686-pc-mingw32
    System type: Win32 6.2.9200 X86
    Lisp implementation type: SBCL
    Lisp implementation version: 1.2.7
    (%i2) integrate(x*sin(x)*exp(-1/2*x^2),x,minf,inf);
    (%o2) (sqrt(2)*sqrt(%pi))/sqrt(%e)
    
     
    • sslavi

      sslavi - 2015-10-23

      Very interesting!!
      Yes, I have my personalized .mac file loaded by default, and I've found the culprit: the radexpand flag!

      If the radexpand is set to "true", the correct solution (sqrt(2)sqrt(%pi))/sqrt(%e) is returned.

      If the radexpand is set to "all", the wrong solution 0 is returned.

      Samuel, it would be interesting to see what is the default value of the radexpand flag in your Maxima session.

       

      Last edit: sslavi 2015-10-23
      • Samuel

        Samuel - 2015-10-23

        My default value of radexpand flag is true. But with radexpand is set to "false", I get the correct answer!
        Thanks a lot.

         
  • Samuel

    Samuel - 2015-10-22

    Where can I find an updated version of maxima for a 64 bit architecture?

     
  • Samuel

    Samuel - 2015-10-23

    Thank you, but my question was unprecise. I work on Linux and all the versions in the Maxima-Linux folder are labeled i386 so basically 32 bit. Even my current version 5.32.1 !!!
    But I get my answer for the main question and I know now that it is not a question of updating.

     
  • sslavi

    sslavi - 2015-10-24

    One question still remains.
    Why do I get the correct result with "radexpand" set to "true", and you don't?

     
  • Samuel

    Samuel - 2015-10-24

    Maybe there are some flags involved but I am not familiar with it. At least, do you get the correct answer when "radexpand" set to "false"?

     
  • sslavi

    sslavi - 2015-10-24

    Yes I do.

     
  • David Scherfgen

    David Scherfgen - 2022-12-23
    • status: open --> closed
     
  • David Scherfgen

    David Scherfgen - 2022-12-23

    This has been fixed by now, unclear in which commit.

     

Log in to post a comment.