Self-adaptation using multi-chromosomes

Adaptation of the parameters and operators in Evolutionary Algorithms is an important research area as it tunes the algorithm to the problem while solving the problem. Self-adaptation where we let the parameter values and operator probabilities evolve is important as here we do not have to design the feedback mechanism or rules to implement the adaption. In this paper we extend self-adaptation to non-numeric problems in Genetic Algorithms by using a multi-chromosome representation. We modify a genetic algorithm for a Cutting Stock Problem to self-adapt two strategy parameters; the results indicate that the approach works quite well.