QoS动态变化的服务选择算法

Service selection algorithm with QoS dynamic changing

由于现代战争的快节奏和异常激烈,在面向服务的军事综合电子信息系统中候选服务的服务质量往往随时间快速变化,有时还有服务的加入和退出,现有组合服务选择方法很难应对这种场景.提出了一种基于危险理论的动态约束多目标免疫克隆算法(DCMOICADT)用于QoS动态变化的服务选择.首先将基于QoS的军事信息服务选择问题建模为带QoS约束的动态多目标组合优化问题,接着采用基于危险理论的动态约束多目标免疫克隆算法同时优化多个目标函数,最终产生一组满足约束条件的Pareto最优解服务组合集.对比实验结果表明,DCMOICADT设计了环境感知因式用于描述QoS动态变化,使用Pareto-占优集和有益不可行解协同的免疫进化方案,能根据当前环境的变化快速且自适应地调整各免疫操作,所得最优解集具有较好的多样性和较强的逼近性,能有效解决QoS动态变化的军事信息服务选择问题.

In the service-oriented C4ISR system, current service selection approaches can＇t deal with thesituation in which the QoS of candidate services are dynamically changing in the fast and intensely modernwar, even there are services arisen or disappeared. A dynamic constrained multi-objective optimizationimmune clone algorithm based on danger theory （DCMOICADT） is proposed for above situation. Thealgorithm firstly transforms the problem of dynamic Web Service selection with QoS global optimal intoa dynamic multi-objective services composition optimization with QoS constraints. The DCMOICADT isutilized to produce a set of optimal Pareto services composition with constraint principle by means of op-timizing various objective functions simultaneously. Experimental results indicate that the DCMOICADTdesigns the environment apperceive equation to describe the QoS dynamic change, and the Pareto dominateset and the helpfully infeasible solutions are corporately immune evolving. The algorithm has adaptivelyshifted the immune operators according to the change of environment and has better performance in di-versity, strong imminence and large distribution which solves the QoS dynamic changing military serviceselection problem feasibility and efficiency.