计算机工程与应用 ›› 2010, Vol. 46 ›› Issue (4): 219-222.DOI: 10.3778/j.issn.1002-8331.2010.04.069

• 工程与应用 • 上一篇    下一篇

用多Agent系统分配具有启动成本的有限资源

郑 明,王玲娟,蔚承建   

  1. 南京工业大学 信息科学与工程学院,南京 210009
  • 收稿日期:2008-08-01 修回日期:2008-10-20 出版日期:2010-02-01 发布日期:2010-02-01
  • 通讯作者: 郑 明

Allocating limited capacity resources with fixed price using multi-agent systems

ZHENG Ming,WANG Ling-juan,WEI Cheng-jian   

  1. College of Information Science and Engineering,Nanjing University of Technology,Nanjing 210009,China
  • Received:2008-08-01 Revised:2008-10-20 Online:2010-02-01 Published:2010-02-01
  • Contact: ZHENG Ming

摘要: 提出了基于多Agent系统分配具有启动成本的有限资源的两种方法:集中式求解方法和分布式求解方法。在分布式求解方法中,给出了分布式的连续双向拍卖协议算法,其Agent采用零智慧增强学习策略。实验结果表明:相对于集中式求解方法,使用分布式求解方法是以降低部分效率为代价的,但是当市场的总需求逐渐接近所有卖者能提供的最大资源数时,市场平均效率呈现逐渐递增的趋势。

关键词: 多Agent系统, 集中式求解方法, 分布式求解方法, 连续双向拍卖协议, 零智慧增强学习策略, 市场效率

Abstract: Two methods which are centralized solving method and decentralized solving method are proposed to allocate limited capacity resources with fixed price based on multi-agent system.In the decentralized solving method,a distributed Continuous Double Auction(CDA) protocol is given in which agents adopt Zero Intelligence Plus(ZIP) strategy.The experimental results illustrate that,compared with centralized solving method,the usage of the decentralized solving method is at the cost of reducing part of efficiency and the market efficiency has a trend to increase as total demand gradually approaches the maximum capacity of the sellers.

Key words: multi-agent system, centralized solving method, decentralized solving method, Continuous Double Auction(CDA), Zero Intelligence Plus(ZIP) strategy, market efficiency

中图分类号: