The article presents an investigation of heuristic behaviour of search algorithms applied to numerical problems. The aim is to compare the abilities of Particle Swarm Optimisation, Differential Evolution and Free Search to adapt to variety of search spaces without the need for constant re-tuning of algorithms parameters. The article focuses on several advanced characteristics of Free Search and attempts to clarify specifics of its behaviour. The achieved experimental results are presented and discussed.
|Title of host publication||PROCEEDINGS of the 20th International Conference SAER-2006|
|Number of pages||6|
|Publication status||Published - 1 Sept 2006|