A Novel Approach to Reduce the Effect of Initial Population on Simple Genetic Algorithm

摘要

In this paper, a novel technique in reducing the sensitivity of initial population on Simple Genetic Algorithm (SGA) is reported. Instead of solely relying on mutation to introduce new schema, a small group of migrant chromosomes is injected at each generation to the mating pool. This mechanism provides more chances for correct schema to be included with less probability of destroying existing ones. Experimental results demonstrate that the new approach adopted in SGA can attain satisfactory performance with complete ill-formed initial population.