基于模拟退火机制的精英协同进化算法

Elite Co-evolutionary Genetic Algorithm Based on Simulated Annealing Mechanism

为了获取更好的全局寻优性能,同时保持较快的收敛速度,文中结合精英策略、协同进化思想和模拟退火机制,提出了一种基于模拟退火机制的精英协同进化算法(SACEA)。算法维持三个种群:精英种群、普通种群和随机种群。精英个体组团,并和其他组员个体协作或对其引导来达到进化目的。SACEA算法在精英组团过程中引入随机种群以增加种群多样性,同时随机个体和精英个体的合作采用快速模拟退火机制来实现,使算法获得了更好的全局寻优性。通过对15组标准测试函数的仿真,并和已有的算法进行对比,很容易得出:SACEA算法具有更强的全局寻优能力,同时收敛速度也有所提高。

In order to get better global optimization performance and maintain the fast convergence speed,combined with the elite strategy and the concept of co-evolutionary and simulated annealing mechanism,put forward a new algorithm,that is the elite co-evolutionary ge-netic algorithm based on simulated annealing method （ SACEA） . The algorithm maintains three populations including elite population, common population and stochastic population. And then elite individuals form teams and exchange information with other team members with the cooperating operation or leading operation. SACEA introduces the stochastic population to evolution to improve diversity of pop-ulation,at the same time,the stochastic individual and the elite individual using fast simulated annealing method to realize the purpose of cooperation. Through all above,the algorithm gets the better global optimization performance. Simulation on 15 standard test simulation and compared with existing algorithms,it is clearly shown that SACEA has better ability of searching globally optimal solution and makes an improvement in convergence speed.