Encoding multiple solutions in a linear genetic programming chromosome

Abstract

Linear Genetic Programming (LGP) is a Genetic Programming variant that uses linear chromosomes for solution encoding. Each LGP chromosome is a sequence of C language instructions. Each instruction has a destination variable and several source variables. One of the variables is usually chosen to provide the output of the program. In this paper, we enrich the LGP technique by allowing it to encode multiple solutions for a problem in the same chromosome. Numerical experiments show that the proposed Multi-Solution LGP significantly outperforms the standard Single-Solution LGP on the considered test problems. © Springer-Verlag 2004.