摘要
为了确保在服务组合中获得Pareto最优解集,把服务组合建模为多个服务质量属性同时优化的多目标优化问题,提出了一种依据服务质量属性类型的通用预处理方法,采用多个信息素表和单个启发式信息表的多目标蚁群算法,蚂蚁随机选择一种信息素表建构可行解,每个蚁群周期完成后所有信息素都会蒸发,但每个优化函数只有一个最优解获得信息素增加,经过多过蚁群周期后即可解获得最优解集。实验结果表明,该方法可为Web服务组合提供一种很好的优化方案,具有很高的准确率。
To obtain the Pareto optimal solution set of Web services composition, Web services composition is modeled to multi- obiective optimization of multiple quality of service. A generic pretreated method based on service quality attributes type is proposed. Several pheromone tables and single heuristic information table are adopted in this multi-objective ant colony algorithm. The ant selected a pheromone table randomly and constructed a feasible solution. All pheromone evaporate partly after each ant colony cycle, but only the optimal solution's members obtains pheromone increase for each optimization function. The optimal solution set could obtain after a lot of ant colony cycle. Simulation shows that the method provide a good optimized scheme for web services composition with high accuracy.
出处
《计算机工程与设计》
CSCD
北大核心
2012年第3期885-889,共5页
Computer Engineering and Design
基金
贵州省科学技术基金项目(QKHJZi[2008]2124
QKHJZi[2010]2102)
关键词
PARETO最优解集
多目标优化
服务质量
WEB服务组合
蚁群优化
Pareto optimal solution
multi-objective optimization quality of service web services composition ant colony opti-mization