期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2006
卷号：6
期号：2B
页码：239-239~245
出版社：International Journal of Computer Science and Network Security
摘要：It is quite necessary that optimizing complicated systems using multi-population coevolutionary genetic algorithms implemented on networks. But there is no effective way to allocate computational resources to improve the performance of algorithms. A decision-making model for resource allocation in multi-population coevolutionary genetic algorithms implemented on networks is established, considering the case that all the populations coevolve synchronously. In this model, the optimized objective is to minimize the maximal computational time for average one-step iteration of a computational node, and the constraint is that each evolutionary population is allocated to one computational node at most. A method for solving the presented model based on genetic algorithms is also given. The validity of the model is shown through comparative examples. The results obtained in this paper provide a feasible mode for resource allocation in multi-population coevolutionary genetic algorithms implemented on networks.
关键词：genetic algorithms, coevolution, multi-population, implemented on networks, resource allocation.
