提出了一种基于改进模拟退火算法的Qo S动态服务组合方法.根据用户功能需求进行服务筛选,构造服务组合所需的候选服务集,再对候选服务进行分类产生同类服务集合.根据改进的模拟退火算法从候选服务集中选取满足用户Qo S需求的组合服务.当组合服务或构件服务接近Qo S临界值时,综合使用局部贪心算法和改进模拟退火算法进行服务重组.案例分析表明,该方法在动态服务组合方面是可行而有效的.
This paper proposes a QoS dynamic services composition based on an improved simulated annealing algorithm. First, classification services builds a set of candidate ser- vices from the service repository according to user's functional requirements. Optimal composite services are computed using an improved simulated annealing (ISA) algorithm, and then recommended to the user. When the quality of composite service is close to a critical value of QoS, a local greedy algorithm and global ISA algorithm are used to re-implement service composition. Feasibility and effectiveness of the proposed method is shown by a case study.