Adaptive Genetic Algorithm Based on Population Diversity

摘要

Population diversity is the precondition of population evolution. Premature convergence is a problem in genetic algorithm. It is seriously influenced by the distribution properties of initial population. So an adaptive genetic algorithm based on diversity is proposed in this paper. It uses the max-min distance means and makes various individuals maintain some certain Hamming distance to produce good population distribution. Simultaneously the genetic operators are adaptively determined according to the population diversity and individual fitness. The diversity of population can be effectively maintained and a global optimal solution can be quickly obtained using the proposed method. Finally, four representative test functions are chosen to test the improved adaptive genetic algorithm's capability. The simulation and comparison results show the validity of this algorithm.