关键词: 双变异率, 海明距离, 局部小变异, 全局大变异

Abstract: An improved genetic algorithm based on dual mutation is proposed to overcome slow convergent speed，poor seeking optimization capabilities and easy to fall into a local optimum of the standard genetic algorithm.In the evolution process，a concept of generalized hamming distance is introduced.When the crossover operator which is controlled by the generalized hamming distance have individuals less than the population size，a small local mutation carries on original population，which can avoid inbreeding，expand the search space，increase the population diversity and effectively curb the premature convergence.Following，the overall big mutation ensures global convergence of the whole process.The simulation with the typical test functions indicates that this new genetic algorithm can significantly improve the quality of solutions and convergence speed.

Key words: dual mutation, hamming distance, small local mutation, overall big mutation