Abstract:
Genetic Algorithms (GA) is a method that imitates evolution of life. It is a searching and optimization algorithm. In the past thirty years, genetic algorithms is succeed in solving some complex global optimization problem. In this paper, an improved GA is given. It adopts a new hybrid operator, and a hybrid coding method. At the same time, we discuss a kind of random dynamic programming problems and construct hybrid genetic algorithms for these problems. On the basis of analyzing and simulating examples of manufacturing decision problem, it shows that the algorithm is efficient to solve these dynamic programming problems. The algorithm is more robust and more convenient for user, it not only avoids plunging into local optimum solutions but also solves the dynamic programming problem quickly without any other mathematics knowledges.