Neural network optimization based on improved diploidic genetic algorithm

In this paper, a kind of improved method of diploid genetic algorithm (DGA) without considering the dominant and recessive of the allele is given directed at the disadvantages of DGA which are easy to fall into premature convergence and have low efficiency in late period local searching. Improved the genetic operation process by imitating the reproductive processes of diplont and adopting the process of gametes recombination and homologous chromosomes chiasma. United the advantages of genetic algorithm and neural network, a new neural network structure contacted with the diploid genetic algorithm closely is designed. This scheme combines the strong global search capability of genetic algorithm and self-learning ability of neural network. Then applied the method to the complex multi-peak function optimization. Simulation results show that the improved algorithm can keep the population diversity and repressed the premature convergence effectively. The neural network optimization based on diploid genetic algorithm increased the convergence speed and accuracy, and ensured the global optimal.