一种基于杂草优化的全局Web服务选择算法

摘要

随着网络中的Web服务数量的暴增,基于QoS的服务选择成为保证组合服务质量和可靠性的关键环节.针对当前服务选择算法全局优化能力弱的问题,将服务选择建模为带约束的非线性最优化问题,并提出一种基于杂草优化的服务选择算法.首先随机产生一组服务选择可行解并将其编码为杂草个体,进而根据个体的QoS效用值确定个体的繁殖数,最后以高斯分布方式指导种群的扩散完成对解空间的快速搜索.理论分析与实验结果表明,该算法有效性和鲁棒性强,可获得相较于已有文献更优的全局解.%With the increasing number of Web services on network, QoS based service selection becomes a key factor to ensure the quality and reliability of service-oriented system.Existing service selection algorithm fail to achieve global results.To attack this problem, this paper modeled the service selection problem as a constrained non-linear optimization problem.Then it proposd a weed optimization based service selection algorithm.Firstly, it randomly generated and coded a group of feasible solutions to weed individual.Then, it calculated the reproduction numbers of weed individuals based on their QoS utility value.Finally, the weed population would spread in manner of Gaussian distribution to search the solution space.The theoretical analysis and experimental results demonstrate the efficiency and robustness of the proposed algorithm.This approach can achieve better global results than existing state-of-the-art approaches.