Sciweavers

CEC
2008
IEEE

Scalarization versus indicator-based selection in multi-objective CMA evolution strategies

14 years 5 months ago
Scalarization versus indicator-based selection in multi-objective CMA evolution strategies
Abstract—While scalarization approaches to multicriteria optimization become infeasible in the case of many objectives, for few objectives the benefits of populationbased methods compared to a set of independent singleobjective optimization trials on scalarized functions are not obvious. The multi-objective covariance matrix adaptation evolution strategy (MO-CMA-ES) is a powerful algorithm for real-valued multi-criteria optimization. This populationbased approach combines mutation and strategy adaptation from the elitist CMA-ES with multi-objective selection. We empirically compare the steady-state MO-CMA-ES with different scalarization algorithms, in which the elitist CMA-ES is used as single-objective optimizer. Although only bicriteria benchmark problems are considered, the MO-CMA-ES performs best in the overall comparison. However, if the scalarized problems have a structure that can easily be exploited by the CMA-ES and that is less apparent in the vector-valued fitness functi...
Thomas Voß, Nicola Beume, Günter Rudolp
Added 29 May 2010
Updated 29 May 2010
Type Conference
Year 2008
Where CEC
Authors Thomas Voß, Nicola Beume, Günter Rudolph, Christian Igel
Comments (0)