基于双种群协同进化的QoS全局最优Web服务选择算法

Dual population co-evolutionary Web services selection algorithms with QoS global optimization

针对服务质量(quality of service,QoS)全局最优Web服务选择问题,提出了一种双种群协同进化QoS全局最优Web服务选择算法。算法在多目标离散粒子群算法基础上设计一种双种群协同进化框架以同步进行非支配排序和精英粒子保留,并定义了一种新的离散粒子位置更新算子。同时为保证粒子的多样性和算法的全局收敛能力,算法采用基于距离的粒子多样性度量算子、基于适应值排序的粒子选择算法和基于轮盘赌的全局最优解选择策略。仿真实验结果表明该算法能同时优化多个目标,并得到一组满足约束的Pareto最优解,且具有较好的性能和鲁棒性,解集的质量和分布也优于非支配排序遗传(nondominated sorting genetic algorithm,NSGA)算法的改进算法NSGA-Ⅱ,能有效解决QoS全局最优的Web服务选择问题。

To solve the problem of Web services selection with quality of service （QoS） global optimization, a novel dual population co-evolutionary Web services selection algorithm is proposed. Inspired by the improved nondominated sorting genetic algorithm （NSGA-II） and based on multi-objective discrete particle swarm optimization algorithm, the proposed algorithm uses a dual population co-evolutionary framework to do non-dominated sorting and elitist maintaining synchronously, and defines a new discrete particle position update operator. Furthermore, a distance-based diversity measurement operator, an adaptive value aware particle selection operator and a roulette-based global optimal solution selection strategy are designed to ensure particles diversity and achieve better global convergence ability. The proposed algorithm can optimize multiple objectives simultaneously and finally obtain a set of constrained Pareto optimum composite service solutions. Theoretical analysis and experimental results indicate that the proposed algorithm not only owns satisfied performance and robustness, but also has better solution quality and distribution than NSGA-II, all of which indicate that the proposed algorithm is an efficient method applied to QoSaware Web service selection with global optimization.