Using predators and preys in evolution strategies

This poster presents an evolution strategy for single- and multi-objective optimization. The model uses the predatorprey approach from ecology to scale between both cases. Furthermore the main issue of adaptation working for singleand multi-objective problem-instances equally is discussed. Particular, the well proved self-adaptation mechanism for the mutation strengths in the single-objective case is adopted for the multi-objective one. This self-adaptation process is supported by a new strategy of competition between predators and preys. Six test functions are used to demonstrate the practicability of the model. Categories and Subject Descriptors I.6 [Simulation and Modeling]: Model Development General Terms Algorithms Keywords Predator-Prey, multi criterion, Evolution strategy

• Type: Conference
• Year: 2005
• Where: GECCO
• Authors: Karlheinz Schmitt, Jörn Mehnen, Thomas Michelitsch
• Keywords: Evolution Strategy, GECCO 2005, Self-adaptation Mechanism, Singleand Multi-objective Problem-instances