摘要
利用分布式约束满足的方法求解分布式配置问题时,在过约束和欠约束条件下都不能得到令人满意的结果.文中将分布式配置问题抽象为分布式组合最优化问题,把遗传退火算法扩展到分布式计算环境以求解分布式配置问题.以SOAP为基础搭建实验平台,在各种约束情况下,文中算法都给出了令人满意的实验结果.可见分布式遗传退火算法可以求解各种约束条件下的分布式配置问题.
DCSP(distributed constraint satisfaction problem) algorithms can not give satisfactory result while solving distributed configuration in over-constrained or under-constrained situation. Then distributed configuration was regarded as distributed combinatorial optimization problem. The genetic annealing algorithms were expanded to distributed genetic annealing algorithms(DGAA) to solve distributed configuration problem. A test software platform was programmed based on SOAP(simple object access protocol). The computing results were satisfactory in any constraint situation. As a result, DGAA can solve distributed configuration problem in any constraint situation.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2005年第6期1335-1340,共6页
Journal of Computer-Aided Design & Computer Graphics
基金
国家自然科学基金(60273056)
关键词
分布式配置
遗传退火算法
分布式组合最优化
SOAP
distributed configuration
genetic annealing algorithms
distributed combinatorial optimization
SOAP