[Scipy-tickets] [SciPy] #1588: scipy.optimize.cobyla not consistant in Windows

SciPy Trac scipy-tickets@scipy....
Tue Jan 24 15:56:16 CST 2012


#1588: scipy.optimize.cobyla not consistant in Windows
----------------------------+-----------------------------------------------
 Reporter:  casperskovby    |       Owner:  somebody    
     Type:  defect          |      Status:  needs_review
 Priority:  normal          |   Milestone:  Unscheduled 
Component:  scipy.optimize  |     Version:  0.9.0       
 Keywords:                  |  
----------------------------+-----------------------------------------------
Changes (by casperskovby):

  * status:  needs_info => needs_review
  * version:  0.10.0 => 0.9.0


Comment:

 Maybe you are right. The example above produces two different solutions:

 [   10.03757281    19.00251316    28.01307958    37.00515082    46.006885
     55.00541918    64.00338297    73.00529579    82.00094877
 91.00473487
    100.79408054   361.10189117   784.72829641  1369.38714815
 2116.63289219
   3025.59712551  4096.43042584  5329.77681657  6724.15961497
 8281.86631326]
 and
 [   10.04021941    19.00317346    28.01289247    37.00549988
 46.00679365
     55.00545058    64.00356037    73.00517599    82.00126994
 91.00486307
    100.84905325   361.1232957    784.72018973  1369.40159306
 2116.63416335
   3025.59715952  4096.46017614  5329.75391513  6724.20850515
 8281.88890401]

 If this is small differences I do not know but it is of course not
 satisfactory. Do you still think that rounding errors and alignment of
 data in memory are causing the different solutions? If so how can you
 avoid this?

 I am using Enthought 7.1 with Scipy 0.9.0.

-- 
Ticket URL: <http://projects.scipy.org/scipy/ticket/1588#comment:4>
SciPy <http://www.scipy.org>
SciPy is open-source software for mathematics, science, and engineering.


More information about the Scipy-tickets mailing list