/pyMACE

Python implementation of the MACE Bayesian optimization algorithm, with GPy used as the backend GP library

Primary LanguageC++

README

About

The origin MACE algorithm, re-implemented in python

  • MCMC hyperparameter integration
  • Matern52 kernel

Dependencies

  • GPy for the GP regression
  • toml for config file
  • Platypus for multi-objective optimization

TODO

  • Improve the interface
  • Combine MACE with other parallel strategy like kriging believer or local penalization