Self-adaptation of control parameters is realized in classical evolution strategies (ES) using the appropriate mutation operators controlled by strategy parameters (i.e. mutation strengths) that are embedded into representation of individuals. The mutation strengths determine the direction and the magnitude of the changes on the basis of the new position of the individuals in the search space is determined. This chapter analyzes the characteristics of classical mutation operators, like uncorrelatedmutationwithonestepsizeanduncorrelatedmutationwithnstep sizes. In line with this, the uncorrelated mutation with n 4-dimensional vectors is proposed that beside the mutation strengths utilizes two additional strategy parameters embedded in the 4-dimensional structure used for definition of the change, i.e., shifting the location of normal distribution for a small shift angle and reversing the sign of the change. The promising results conducted on the suite of ten benchmark functions taken from the publications shown that the ES despite their maturity serve as an interesting area of future research.
CITATION STYLE
Fister, I., & Fister, I. (2015). On the mutation operators in evolution strategies. Adaptation, Learning, and Optimization, 18, 69–89. https://doi.org/10.1007/978-3-319-14400-9_3
Mendeley helps you to discover research relevant for your work.