Production Rules As Chromosomes of GA for Robotic Swarm Applications

Fairly complicated data structures, other than a string of binary digits, as chromosome representations of genetic algorithms (GAs) have been tried, and published, by various researchers. How to choose an appropriate representation for a GA problem is not only necessary, but also a critical step. Once the genome 1 is not a string of binary digits, the recombination operators have to be re-defined. However, the merits gained through the suitability of genome representation, like faster convergence, easier effect interpretation etc. outweigh the extra effort required. This paper presents results about an experiment on utilization of production rules as chromosomes of a GA problem on a common multiple robots cooperative task: environment exploration. It provides details of implementation and gives a narration about the new devised recombination operators like crossover and mutation.