Optimization Without Derivatives: Prima Fortran Version and Inclusion in SciPy

Re:Optimization Without Derivatives: Prima Fortran Version a... by RockyMcNuts on Tuesday May 16, 2023 @09:20AM (#35961510)

I think there are a few derivative-free global optimizers, curious what sets this one apart? see e.g. https://www.microprediction.com/blog/optimize . If the current scipy-native one is unmaintained and this one is superior it certainly makes sense to update.

[ Parent | on HN ]

Re:Optimization Without Derivatives: Prima Fortran Version a... by icegreentea2 on Tuesday May 16, 2023 @09:30AM (#35961620)

The proposal isn't a "new" algorithm, rather a rewrite (plus bug fixes) of the existing scipy implementation of "COBYLA" (which I don't think it tested in your linked comparison).

[ Parent | on HN ]

Re:Optimization Without Derivatives: Prima Fortran Version a... by ckmiller on Tuesday May 16, 2023 @11:24AM (#35963110)

These algorithms work best on smooth functions with mid-to-low 100's of variables. They are very sample efficient, often obtaining superlinear convergence using just a linear number of evaluations of the objective function.

[ Parent | on HN ]