摘要
针对满足用户多QoS约束条件的Web服务组合优化问题,提出一种基于图的三阶段Web服务组合方法Sky-MCSP-R。该方法从候选服务空间中筛选出Skyline服务,直接在优质候选服务的基础上构造基于图的Web服务组合模型,减少模型的结点规模。利用引入了过约束机制的MCSP-K算法进行服务组合,弱化约束强度使MCSP-K算法产生尽可能多的可行解。运用Relax算法求得最优解。实验结果表明,该方法在保持较高优化率的基础上提高了组合效率,减少了无解现象。
Aiming at the optimization problem of Web service composition which satisfies user's multiple QoS constraints, a Web service composition method in three phases based on graph named Sky-MCSP-R is proposed. First, Skyline services are selected from the candidate service space, thus the model of Web serviee composition based on graph is conducted directly on these high quality candidate services, the scale of nodes of the model is reduced. Then, MCSP-K algorithm using the over-constraint mecha nism is employed to compose services, the constraint intensity is weakened to make the algorithm MCSP-K produce as many so lutions as possible. At last the method uses the algorithm Relax to get the optimal solution. Experiments show that the method improves the composition efficiency on the basis of keeping a higher optimization rate and reduces the no solution phenomenon.
出处
《计算机工程与设计》
CSCD
北大核心
2014年第1期148-152,共5页
Computer Engineering and Design
基金
国家自然科学基金项目(61202097)