基于多Agent的制造网格资源调度新算法  

A New Algorithm Research into Scheduling for Manufacture Grid Resources Based on Multi-Agent

作　　者：曲巨宝[1,2] 刘胜[2] 李玲瑞[2] 

机构地区：[1]武夷学院电子工程系 [2]武夷学院数学与计算机系,福建武夷山354300

出　　处：《重庆师范大学学报（自然科学版）》2010年第3期60-63,共4页Journal of Chongqing Normal University:Natural Science

基　　金：福建省教育厅科技项目(No.JA09240);武夷学院智能计算网格科研团队资助(2009)

摘　　要：针对在制造网格广域、动态、异构的复杂环境中如何快速准确地发现并调度资源,使QoS达到最佳效果,本文提出了利用移动Agent在制造网格中发现资源的新方法,先用遗传算法对资源信息进行选择,然后将移动Agent作为用户提交任务的载体,携带用户的资源信息在网格环境中利用蚁群算法对资源进行精确查找。设计了资源优选目标,改进了遗传蚁群算法的路径优化策略,在移动Agent查找路径的前半程,采用遗传算法,充分利用遗传算法的快速性、随机性、全局收敛性,求得一个较为精确的解。后半程,将遗传算法过渡到蚁群优化算法,利用蚁群算法的并行性、正反馈性、求精确解效率高的特点,求得最终结果。经仿真实验证明此算法可以获得很好的收敛速度和精确解。In view of how to discover and dispatch the resources quickly and accurately in the wide-area, dynamic and isomerous complex environment of manufacturing grid, enabled QoS to achieve desired effect, this article proposes the new method that uses the mobile Agent to discover the resources in the manufacturing grid. First it uses the genetic algorithm to choose the resources information, then puts mobile Agent as the cattier for users submission to their duties, The resources information to carried users, to use the ant group algorithm, carry on the exact search in the grid environment to the resources. It desigues the resources optimal goal and improves the optimization strategy of heredity ant group algorithm. On the first half range of mobile Agent searching path, it uses the genetic algorithm and its rapidity, randomness and the global convergence characteristics to get a more precise solution. On the latter half range, it transits the genetic algorithm to the ant group optimization algorithm, uses ant group algorithm＇s characteristics of parallelism, the regeneration and high efficiency in precise solution, to get the final result. After the simulation experiment, it proves that this algorithm method can get the very good convergence rate and the exact solution.

关 键 词：多AGENT 制造网格 遗传 蚁群 调度 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]