摘要
提出一种新的基于模糊控制策略的交叉和变异概率自适应调节算法.该算法以相邻两代群体之间平均适应度函数和标准差的差值作为输入,以交叉和变异概率的变化量作为输出.并提出了与输入相对应的自适应归一化算子以及新的基于启发式知识的模糊规则,用于交叉和变异概率的调节.对3种不同测试函数的数值仿真研究表明,与其他2种自适应模糊控制算法相比,该调节算法可使遗传算法具有更快的搜索速度和更高的搜索质量.
A new adaptive algorithm for regulating the probabilities of crossover and mutation based on fuzzy logic is proposed. The changes of average fitness value and standard deviation between two continuous generations are selected as input, while the changes of crossover probability and mutation probability as output. Two adaptive scaling factors are introduced for normalizing the input and new fuzzy rules based on domain heuristic knowledge are investigated for adjusting the probabilities of crossover and mutation. Numerical simulation studies of three different test functions are carried out, and the simulation results show that the genetic algorithm with the proposed algorithm exhibits improved search speed and quality compared with two other adaptive fuzzy control algorithms.
出处
《控制与决策》
EI
CSCD
北大核心
2008年第1期79-83,共5页
Control and Decision
基金
国家自然科学基金项目(60374032)
关键词
遗传算法
交叉概率
变异概率
模糊控制
Genetic algorithm Crossover probability
Mutation probability
Fuzzy control
