一种模拟退火伪并行遗传算法  

Pseudo Parallel Genetic Algorithm Based on Simulated Annealing

作　　者：郭栋[1] 时银水[1] 朱岩[1] 

机构地区：[1]防空兵指挥学院,河南郑州450052

出　　处：《指挥控制与仿真》2007年第3期37-40,共4页Command Control & Simulation

摘　　要：借鉴模拟退火算法的局部搜索能力,结合并行计算的思想设计了一种采用模拟退火机制的实数编码自适应交叉、全概率变异伪并行遗传算法,最后用这种方法对典型的多峰值函数求极值,并和基本遗传算法进行比较,结果表明:该算法具有较强的全局搜索能力和局部搜索能力,能够更有效地克服早熟收敛问题。In view of the local search ability of simulated annealing algorithm, a new style of optimal algorithm which combined simulated annealing algorithm with pseudo parallel genetic algorithms and adopts real coding, self-adaptive crossover and overall mutation is designed, then, compared with the standard genetic algorithms, the experimental results of typical multimodal function show that this algorithm can overcome premature convergence and find global optima more efficiently.

关 键 词：模拟退火 自适应交叉 全概率变异 伪并行遗传算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]