第2期 蒋伟进，等：MAS动态协作任务求解模型与算法 ·167· 1.2 plications,2001,15(3):258281 1.0 [3]SUBRAMONIAM K,MAHESWARAN M,TOULOUSE M. 0.8 Towards a micro economic model for resource allocation in 6 grid computing system[C]//The 2002 IEEE Canadian Conf 0.4 ◆竞价 on Electrical Computer Engineering.Manitoba,Canada, 0.2 ■蛛网 2002:782-785. 0 30 [4]BUYYA R.Economic-based distributed resource manage- 资源数量 ment and scheduling for grid computing D].Melbourne, Australia:Monash University,2002. 图2资源的利用率对比 [5]CHENG J Q,WELLMAN M P.The WALRAS algorithm: Fig.2 Comparison of resource utilization convergent distributed implementation of general equilibrant 图3是蛛网模型和本文模型达到均衡价格的迭 outcomes[J].Computational Economics,1998,12 (1):1- 代次数.从图中可知，本文模型的均衡价格为6，资 24. 源的总供给为35.改变资源的初始报价，蛛网模型 [6]YGGE F.Market-oriented programming and its application 中初始报价离均衡价格越远，迭代次数越多.本文模 power load management[D].Lund,Sweeten:Lund Univer 型每次迭代根据供需差调整：供需差越大，调价幅度 sity,1998. 越大，很快达到均衡状态 [7]WENG C L,LU X D.A double auction method for resource 50 allocation on computational grids[J].Chinese Joumal of ◆蛛网 Computers,2006,43(6):1004-1009, 40 ■竞价 [8]BUYYA R,ABRAMSON D,VENUGOPAL S.The grid e- 30 conomy[J].Special Issue on Grid Computing,2005,93 (3):698-714. [9]BUYYA R,VAZHKUDAI S.Compute power market:To- 10 wards a market-oriented grid C].Washington.CCGRID 00 23 4 5 6 8 2001:574-581 报价 [10]BUYYA R,ABRAMSON D,GIDDY J.A case for econo- 图3迭代次数 my grid architecture for service-oriented grid computing Fig.3 Iterations [C]//Proc of the 10th IEEE Int'1 Heterogeneous Compu- ting Workshop.Washington:IEEE Computer Society, 3结束语 2001:776-790. [11 BUYYA R,MURSHED M.GRIDSIM:a toolkit for model- 本文提出了一种基于博弈经济机制和多Agent ing and simulation of grid resource management and sched- 动态协作的计算网格资源分配任务求解方法.它通 uling[J].Joumal of Concurrency and Computation:Prac- 过效用函数刻画用户需求的异构性，以一般均衡理 tice and Experience,2002,14(13/15):1175-1220. 论为基础，依靠市场机制实现资源的优化分配.采用 [12]ABRAMSON D,BUYYA R,GIDDY J.A computational MAS博奔经济机制进行资源调度和任务求解有着 economy for grid computing and its implementation in the 许多独特的优点，动态的资源配置提高了系统的自 nimrod-G resource broker[J].Future Generation Comput- 适应性，采用博弈经济原则能够鼓励资源拥有者贡 erSy9tems,2002,18(8):1061-1074. 献他们的空闲资源并从中获利，有助于建立大规模 [13]SUBRAMONIAM K.MAHESWARAN M,TOULOUSE M. 的网格应用系统 Towards a microeconomic model for resource allocation in grid computing system[C]//The 2002 IEEE Canadian 参考文献： Conf.on Electrical Computer Engineering.Manitoba, Canada,2002:373-391. [1]BARUAH S K,COHEN N K.Plaxton in resource allocation [14]CAO H Q,XIAO N,LU X C,et al.A market-based ap- [J].Algorithmica,1996,15(6):600-625. proach to allocate resources for computational grids[J]. [2]WOLSKI R,PLANK JS,BREVIK J,et al.Analyzing mar- Joural of Computer Research and Development,2002,39 ket-based resource allocation strategies for the computational (8):913-916 grid[J].Intemational of High Performance Computing Ap- [15]WOLSKI R,PLANK J S,BREVIK J,BRYAN T.Analy-