基于柯西烟花算法的大规模服务组合优化

Large-scale service portfolio optimization based on Cauchy fireworks algorithm

随着Web服务的日益丰富,如何动态地从大量候选服务集中选择出整体性能高的组合服务已成为服务组合优化领域亟待解决的问题。为稳定、高效地解决大数据集下的服务组合问题,提出一种柯西烟花算法。该算法在搜索过程中引入柯西变异算子增强了算法的全局搜索能力;采用精英候选策略有效降低了算法的时间开销。实验结果验证了该算法在处理大规模服务组合优化问题时的可行性和稳定性。

With the increasingly rich Web service, how to dynamically select the composite services with high overall performance has become a service portfolio optimization urgent problem to be solved from a large number of candidate services. In order to stabilize and efficiently solve the problem of service portfolio in large data sets, a Cauchy fireworks algorithm is proposed. In the search process, the algorithm enhances the global search capability of the algorithm though leading to Cauchy variation operator. Using elite selection strategy, the time cost of the algorithm is effectively reduced. The experimental results verify the feasibility and stability of this algorithm in the large-scale service combination optimization.