自适应调整选择压力的灾变元胞遗传算法

Self-adaptive Cellular Genetic Algorithms with Disaster Based on Selection Pressure

对于遗传算法而言,全局探索和局部寻优能力之间的平衡影响算法的性能,选择压力就代表着这个平衡。只有当全局探索和局部寻优之间的平衡达到最佳化才能够使算法又快又精确的寻求到全局最优解。随着算法运行,种群结构不断的变化,选择压力也在不断变化。分析研究了灾变元胞遗传算法的选择压力,根据种群多样性和种群收敛度,提出一种基于灾变参数调节选择压力的自适应元胞遗传算法。通过两个典型函数优化实验,表明选择压力自适应调节可提高算法性能,并得出这两个函数在寻优过程中的最佳选择压力变化规律,这为自适应算法设计提供了一种新的途径。

For the GA, the trade-off between exploration and exploitation may affect the performance of the algorithm. Only when the trade-off reaches optimization, the algorithm can search the global optimal solution quickly and accurately. As the algorithm runs generation by generation, the construction of population changes unceasing, and the selection pressure is changed too. A self-adaptive algorithm was put forward based on the population diversity and convergence degree to ensure the selection pressure, and adjust the selection pressure by changing disturbances parameter, and the change rule of the best selection pressure for some class problem can be found out. As the two typical functions optimize problem, the experiment indicates that the selection pressure self-adaptive can advance the performance of the algorithm.