Last modified: 15 December 2025

Using SciPy and optimagic optimisers [New]


[New] New in CIAO 4.18 is direct support for a number of optimisers from scipy.optimize or optimagic.

The in-built Sherpa optimisers are:

sherpa> list_methods()
['gridsearch', 'levmar', 'moncar', 'neldermead', 'simplex']

If you have SciPy installed then this will expand to:

['gridsearch',
 'levmar',
 'moncar',
 'neldermead',
 'scipy_basinhopping',
 'scipy_differentialevolution',
 'scipy_direct',
 'scipy_dualannealing',
 'scipy_minimize',
 'scipy_shgo',
 'simplex']

and if you have optimagic installed then it will grow to (optimagic installs tend to include SciPy so the SciPy ones will also be included):

['gridsearch',
 'levmar',
 'moncar',
 'neldermead',
 'optimagic',
 'scipy_basinhopping',
 'scipy_differentialevolution',
 'scipy_direct',
 'scipy_dualannealing',
 'scipy_minimize',
 'scipy_shgo',
 'simplex']

Unlike SciPy, the optimagic optimizers use the same name ("optimagic") and options are used to choose the particular optimiser.