The Application of Dominant-Recessive Diploid Codes in MOGA

Abstract

When solving an optimal problem, different encoding method has an important effect on the performance of multi-objective genetic algorithm. This paper breaks the traditional binary coding ideas, introduces a new dominant-recessive diploid codes which applied in the MOGA. we analyses the impact on solution space by the binary multi-objective genetic algorithm and dominant-recessive diploid codes multi-objective genetic algorithm, which is operated by three basic operators of tournament selection, two-point crossing, and the basic bit mutation. Furthermore, by using the Numerical experiments of three Classic multi-objective optimization test functions, this algorithms and the efficient binary multi-objective algorithms named niched pare to genetic algorithms are compared. From the solution, we know that this paper’s algorithm is obviously superiors to the niched pare to genetic algorithm about the distribution, convergence of solution and the capability of anti-prematurity. Thus, it is interpreted that the algorithm is feasible from the aspects of theoretic analysis and numerical experiments, and the pare to solutions can be came to.