Mutative σ-self-adaptation can beat cumulative step size adaptation when using weighted recombination
Hans-Georg Beyer, Alexander Melkozerov · 2008
This paper proposes the σ-self-adaptive weighted multirecombination evolution strategy (ES) and presents a performance analysis of this newly engineered ES. The steady state behavior of this strategy is investigated on the sphere model and a formula for the optimal choice of the learning parameter is derived allowing the ES to reach maximal performance. A comparison between weighted multirecombination ES with σ-self-adaptation (σSA) and with cumulative step size adaptation (CSA) shows that the σ-self-adaptive ES can exhibit the same performance and can even outperform its CSA counterpart for a range of learning parameters.