【摘要】 为了更有效地抑制文化遗传算法的早熟收敛现象和提高收敛速度,提出了一种多策略结合的文化遗传算法。该算法在信念空间,使用与文化算法不同的接受函数、影响函数和更新函数,在群体空间,针对种群采取多种群化,并采用自适应的交叉变异操作且多种群之间加入竞争机制的遗传算法,这样使得改进后的算法具有更强的全局寻优能力和局部寻优能力,有效避免陷入局部最优,抑制了早熟收敛,提高了收敛效率。用上述算法对几个典型函数进行优化,实验证明了多种群自适应的文化遗传算法的有效性和可行性,新的算法不易陷入早熟收敛,此外全局搜索能力和局部搜索能力得到有效平衡,收敛率高。更多还原

【Abstract】 A multi-population adaptive cultural genetic algorithm(IMFICGA) is proposed in this paper. Base on the improved cultural genetic algorithm(ICGA), the novel algorithm introduces the new acceptance function, influence function and update function in cultural algorithm and combines adaptive crossover and mutation operation in genetic algorithm. Besides, it uses the multi-population parallel algorithm and adopts competitive mechanism between populations. Thus IMFICGA has stronger ability of globe search and enables to avoid trapping into local optimum effectively, and suppresses the premature convergence. In the experiment, MFICGA was used to optimize series of classical functions. Simulation show that the algorithm is difficult to trap into premature convergence, in addition, achieves a good balance between ability of global search and local search and higher convergence rate. Thus IMFICGA is effective and feasible.更多还原