摘要
建立了集群负载均衡问题的数学模型,并提出改进多态蚁群算法来对其进行求解的策略.首先,算法中侦察蚁以每个处理节点为中心,作局部侦察并设置侦察信息素;其次,搜索蚁利用侦察蚁提供的辅助信息做全局搜索,通过多态蚂蚁间的协作,能更快地搜索到问题的优化解.最后,通过一个试验与最小加权连接算法,传统多态蚁群算法进行了对比.结果表明,对于负载均衡问题,改进多态蚁群算法比前述算法在算法稳定性,负载的均衡能力,计算速度方面更具有优势.
A novel mathematical model has been developed to address the complicated issue of the Clus- ter-based Load balancing. Therefore, an improved and polymorphic ant colony algorithm（IPACA） has been brought forward to solve the problem of Load balancing. First, spy ants fulfill the reconnaissance to the local route which is beside every processing nodes and set reconnoitering pheromones on the pro- cessing node. Then, search ants search the feasible path by the auxiliary information from spy ants. The cooperating among polymorphic ants can significantly improve the speed to find the optimum solution. Finally, a case study is presented to compare IPACA with Weighted minimum connection algorithm and polymorphic ant colony algorithm. The test results show that the proposed algorithm is of more advan- tage than fore mentioned algorithms in computational results stability, the ability to load balance and computational speed for Load balancing.
出处
《四川大学学报（自然科学版）》
CAS
CSCD
北大核心
2009年第5期1311-1315,共5页
Journal of Sichuan University(Natural Science Edition)
基金
科技部科技型中小企业创新基金(06C26225101730)
四川省科技公关项目(05GG021-003-2)
关键词
LINUX集群
负载均衡
改进多态蚁群算法
软件测试平台
数学模型
Linux clusters, Load balancing, an improved and polymorphic ant colony algorithm, software testing platform, mathematical model
