4. pyopus.optimizer — Optimization algorithms

Optimization algorithms and test function suites module

This module provides unconstrained and bound constrained optimization algorithms.

Nothing from the submodules of this module is imported into the main optimizer module. The optimizer module provides only the optimizerClass() function for on-demand loading of optimizer classes.

Optimization algorithms search for the argument which results in the lowest possible value of the cost function. The search can be constrained meaning that only certain values of the argument are allowed.

Iteration of the optimization algorithm is another name for the consecutive number of cost function evaluation.

pyopus.optimizer.optimizerClass(className)

Returns the class object of the optimizer named className. Raises an exception if the optimizer class object is not found.

This function provides on-demand loading of optimizer classes.

To create an optimizer object of the class HookeJeeves that minimizes function f and put it in opt use:

from pyopus.optimizer import optimizerClass
OptClass=optimizerClass('HookeJeeves')
opt=OptClass(f)

• 4.1. pyopus.optimizer.base — Base classes for optimization algorithms and plugins
• 4.2. pyopus.optimizer.cache — Caching of function and constraint values, and annotations
• 4.3. pyopus.optimizer.filter — Filter-based point acceptance
• 4.4. pyopus.optimizer.coordinate — Box constrained coordinate search optimizer
• 4.5. pyopus.optimizer.hj — Box constrained Hooke-Jeeves optimizer
• 4.6. pyopus.optimizer.nm — Unconstrained Melder-Mead simplex optimizer
• 4.7. pyopus.optimizer.grnm — Grid-restrained Nelder-Mead simplex optimizer
• 4.8. pyopus.optimizer.sanm — Unconstrained successive approximation Nelder-Mead simplex optimizer
• 4.9. pyopus.optimizer.boxcomplex — Box’s constrained simplex optimizer
• 4.10. pyopus.optimizer.de — Box constrained differential evolution optimizer
• 4.11. pyopus.optimizer.psade — Box constrained parallel SADE global optimizer
• 4.12. pyopus.optimizer.qpmads — Mesh adaptive direct search with quadratic programming