[TRNSYS-users] negative sign in minimum result of cost function trnopt

Jeff Thornton thornton at tess-inc.com
Sun Nov 11 12:28:03 PST 2018



I'm not sure we understand your question. 2e-14 and -2e-14 differ by
only 0.0000000000004 if I counted my zeroes correctly. And an error with
a magnitude of e-14 is not a great choice for an error function. That
falls well within precision of the variables... 


Jeff Thornton
President - TESS LLC
3 N. Pinckney Street, Suite 202, Madison WI USA 53703
Office: (608) 274-2577 Fax: (608) 278-1475
E-Mail: thornton at tess-inc.com

On 11/10/2018 4:33 pm, Michel Tamarzians via TRNSYS-users wrote: 

> Dear all, 
> When doing an optimization one of the minimum results is positive 2.1E-14 but another minimum result is negative 2.1E-14. Can someone please elaborate why the results changes to negative while the minimum found is already much smaller than zero..?? 
> Thank you :) 
> Kind regards, 
> Michel 
> _______________________________________________
> TRNSYS-users mailing list
> TRNSYS-users at lists.onebuilding.org
> http://lists.onebuilding.org/listinfo.cgi/trnsys-users-onebuilding.org [1]

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.onebuilding.org/pipermail/trnsys-users-onebuilding.org/attachments/20181111/52ede43d/attachment.html>

More information about the TRNSYS-users mailing list