mystic: a simple model-independent inversion framework http://www.its.caltech.edu/~mmckerns/software.html
# Version 0.2a1: 05/22/10 # Highlights First alpha version for second minor release. Extending optimizers to parallel and distributed computing. Solvers: - Differential Evolution (x2) - Nelder-Mead Simplex - Powell's Directional Search Method - Pseudo-global Scattershot - Pseudo-global Batch Grid API: - solvers share a common interface - solvers can be called as a unique function or using API - solvers with built-in optimization control handlers - configurable solvers can be bound or unbound - configurable solvers have user-provided or random initial points - configurable termination conditions - configurable mutation strategies (for DE solver) - parallel mapping of optimization jobs - distributed parallel mapping of optimization jobs - parallel multiprocess mapping of optimization jobs - launching with job schedulers Tools: - configurable 2-variable monitors - wrap function with counter or bounds - cost-function generator - standard set of optimization test models - set of example scripts for test cases - math tools Documentation: - User's Guide with tutorials - online Reference Manual --- Mike McKerns California Institute of Technology http://www.its.caltech.edu/~mmckerns -- http://mail.python.org/mailman/listinfo/python-announce-list Support the Python Software Foundation: http://www.python.org/psf/donations/
