CMA-ES
From Wikipedia, the free encyclopedia
CMA-ES stands for Covariance Matrix Adaptation Evolution Strategy. The covariance matrix adaptation is a derandomized method to adapt the covariance matrix of the multivariate normal mutation distribution in the Evolution strategy. The covariance matrix describes the pairwise dependencies between the variables. The adaptation of the covariance matrix leads to learning a second order model of the underlying objective function similar to the approximation of the inverse Hessian matrix in the Quasi-Newton method in classical optimization.
[edit] Principle
The adaptation principle is based on the idea to increase the probability of a successful mutation step. The covariance matrix is changed such that the likelyhood of the successful step(s) of the last generation to appear again is increased.
