Multi-objectivization evolutionary algorithm based on evolvability in dynamic environment
Hao Chen · Journal of systems engineering · 2011
Evolvability can characterize a dynamic performance of an evolutionary system.Evolvability of individuals can directly influence the quality and efficiency of population evolutionary.Fitness improvement and genotypic change are used for measuring the evolvability of individuals.This paper proposes two multiobjectivization evolutionary algorithms by introducing the fitness improvement and by introducing genotypic change respectively.The performances of the new algorithms are verified in dynamic environments with different complexity.The results indicate that the new algorithms have the better performance of dynamic optimization.