[SciPy-dev] Brent's Principal Axis Algorithm

[Sturla Molden]

Benny Malengier skrev:
> To return to the original question, yes, it would be great if this
> could be included in scipy optimize.
> If it needs to be rewritten, if somebody makes a clean description of
> how to do it by looking at the code, I'm sure some people would be
> willing to reimplement it in C for scipy.
>   
I took a quick look at the code. It seems to optimize with Brent's 
method along one dimension, jump randomly to a point in the vicinity, 
optimize along conjugate dimensions, etc. I should not be too hard to 
reimplement with a version of brent for one dimension and a prng.

S.M.