高等学校计算数学学报
高等學校計算數學學報
고등학교계산수학학보
NUMERICAL MATHEMATICS A JOURNAL OF CHINESE UNIVERSITIES
2009年
3期
221-231
,共11页
hybrid genetic algorithm%integer programming%Markov chain
In this paper a hybrid genetic algorithm for integer programming,which integrates both a genetic algorithm and a simulated annealing algorithm by a new method to avoid large drop of the algorithm's efficiency is proposed. A new "nature select" rule is added into the algorithm to ensure that it converges to optimal solution with probability 1. And also individuals more than one are generated by a pair of parents to improve individuals' diversity and algorithm's local search ability. The algorithm's convergence is proved using the theory of Markov chain,and effectiveness is examined by several numerical experiments.
