Andreas Ostermeier

TL;DR: This paper puts forward two useful methods for self-adaptation of the mutation distribution - the concepts of derandomization and cumulation and reveals local and global search properties of the evolution strategy with and without covariance matrix adaptation.

TL;DR: A new formulation for coordinate system independent adaptation of arbitrary normal mutation distributions with zero mean enables the evolution strategy to adapt the correct scaling of a given problem and also ensures invariance with respect to any rotation of the fitness function (or the coordinate system).

TL;DR: A new adaptation scheme for adapting arbitrary normal mutation distributions in evolution strategies is introduced which can adapt correct scaling and correlations between object parameters and reliably adapts mutation distributions corresponding to hyperellipsoids with high axis ratio.

TL;DR: The performance of Evolution Strategies depends on a suitable choice of internal strategy control parameters and the number of necessary step-sizes equals the dimension of the problem.

TL;DR: This paper presents a derandomized scheme of mutative step-size control that facilitates a reliable self-adaptation of individual step-sizes and indicates that the adaptation by this concept declines due to an interaction of the random elements involved.