Optimization via Parameter Mapping with Genetic Programming

Abstract. This paper describes a new approach for parameter optimization that uses a novel representation for the parameters to be optimized. By using genetic programming, the new method evolves functions that transform initial random values for the parameters into optimal ones. This new representation allows the incorporation of knowledge about the problem being solved to improve the search. Moreover, the new approach addresses the scalability problem by using a representation that, in principle, is independent of the size of the problem being addressed. Promising results are reported, comparing the new method with differential evolution and particle swarm optimization on a test suite of benchmark problems.

• Type: Conference
• Year: 2004
• Where: PPSN
• Authors: João Carlos Figueira Pujol, Riccardo Poli
• Keywords: Distributed And Parallel Computing, Initial Random Values, Method Evolves Functions, Parameters, PPSN 2004