new methode to initialize nm_simplex

Brian Gough bjg@network-theory.co.uk
Fri Mar 17 16:13:00 GMT 2006


Ivo Alxneit-Kamber writes:
 > i fear you missunderstood (maybe my comments were not really sufficient)
 > or i do not really get your point here.
 > 

Just that you don't need to modify the API if your algorithm only
needs an initial scalar step instead of a vector -- use the magnitude
of the vector (and ignore the direction if necessary, or put it to
some use).  That fits the existing API.

-- 
Brian Gough



More information about the Gsl-discuss mailing list