PSOSZ:基于Internet群组命令传输算法  

PSOSZ: Group Command Transmission Algorithm Based on Internet

作　　者：章刚[1] 陈庆奎[1,2] 

机构地区：[1]上海理工大学管理学院,上海200093 [2]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《小型微型计算机系统》2015年第5期1021-1023,共3页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60970012)资助;高等学校博士学科点专项科研博导基金项目(20113120110008)资助;上海重点科技攻关项目(14511107902)资助;沪江基金研究基地专项(C14001)资助;上海市一流学科建设项目(XTKX2012)资助;上海市工程中心建设项目(GCZX14014)资助;上海重点科技攻关项目(09511501000;09220502800)资助

摘　　要：Internet尽力而为服务模式在支持群组命令传输过程中,容易产生资源竞争问题,从而导致部分命令传输丢失造成群组命令传输失败.提出基于Internet构建有效路径统计网络(Effective Path Statistics Network,EPSN),把基于Internet网络群组命令传输问题转换成基于EPSN网络群组多约束多目标优化问题(Group Multi-Constraints Multi-Objective Optimization Problem,GM CM OOP).提出基于独占区域粒子群优化算法(Particle Sw arm Optimization based on Sole Zone,PSOSZ).该算法,根据独占搜索空间划分思想搜索GMCMOOP问题的解.实验表明,在群组命令规模分别在最大量175和最小量75下,该模型在基于Internet环境部署的EPSN网络规模不断变化下,GCT成功率相对于经典路由算法DSA和Yen有较好性能,同时误差率相对基本粒子群有较好性能.It is a problem that the best-effort service model of Internet causes easily resources competition, which will generate GCT （ Group Command Transmission ） transmission failed giving rise to part of commands lost. To attack the problem, this paper firstly builds EPSN （Effective Path Statistics Network） based on Interact,and transforms the problem GCT based on Internet into the problem GMCMOOP （ Group Multi-Constraints Multi-Objective Optimization Problem ） based on the EPSN. Aiming at the problem GMCMOOP,this paper proposes PSOSZ （ Particle Swarm Optimization based on Sole Zone）. The algorithm makes use of PSO to searching GMCMOOP problem solutions according to sole search space tactic. The experiment shows that the algorithm, under the condition of scale of group commands between the maximum amount 175 and the minimum amount 75 ,the transmission success rate of GCT, relative to classic routing algorithms such as DSA and Yen, has better algorithm performance, according to scale of EPSN changing based on Internet. Meanwhile,the algorithm has better performance than Particle Swarm Optimization according to error rate.

关 键 词：群组命令传输 独占区域 粒子群优化算法 有效路径统计网络 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]