摘要
给出了压缩遗传算法的模式定理以及收敛性和运算参数的分析,并提出一种快速压缩遗传算法(fcGA).该算法用压缩遗传算法(cGA)运行少量代数得到的概率值及其运行代数组成一个观测样本,借助于统计学中的最小二乘法估算几万代以后的概率值,组成新的概率矩阵并根据该矩阵产生新的个体,用这些新的个体更新概率矩阵.旅行商问题(TSP)的仿真证明,该算法是一种十分高效的遗传算法.
The schema theorem of the cGA (compact genetic algorithm) and the analysis of cGA's convergence and the parameters are given. A kind of fast compact genetic algorithm (fcGA) is proposed. With the probability values that the cGA gets in the beginning generations, the probability values in thousands of generations are estimated by the least square approach. The new probability matrix is composed, from which the new offspring is generated. Based on their fitness the probability matrix is updated. The simulations on the traveling salesman problem show that this algorithm is of high efficiency.
出处
《控制与决策》
EI
CSCD
北大核心
2004年第6期683-686,共4页
Control and Decision
基金
国家自然科学基金资助项目(59889505
70071017).
关键词
压缩遗传算法
最小二乘法
旅行商问题
Computer simulation
Least squares approximations
Markov processes
Probability distributions
Traveling salesman problem
