基于软件人的在线多重拍卖研究  

A Study of Multiple Online Auction Based on SoftMan System

作　　者：曾广平[1] 胡文浩[1] 王洪泊[1] 涂序彦[1] 

机构地区：[1]北京科技大学信息工程学院,北京100083

出　　处：《计算机仿真》2009年第11期304-307,共4页Computer Simulation

基　　金：北京市自然科学基金(4072018)

摘　　要：随着拍卖机构的增加,针对客户需要监视各个拍卖网站,选择一个拍卖参加,做出正确的投标,来保证他们能得到自己想要的东西。由于监视、选择、投标的过程既复杂又耗时,尤其是当各拍卖的开始和结束时间、协议都不相同时,问题更加复杂。提出了一种基于"软件人"的在线多重拍卖选择策略,通过对各种不同拍卖环境的加权分析,使用遗传算法选出较优的拍卖队列及竞标策略。实验表明,策略能够有效提高竞标成功率。As the number of auction houses increases, customers need to watch lots of the auction Web sites, choose to participate in an auction, make a right bid to ensure that they can get what they want. These processes of monitoring, selecting and making bids are both complex and time consuming. The task becomes even more complicated when there are different start and end times and when the auctions employ different protocols. A new strategy based on SoftMan system was proposed. Through analysing a variety of auction environments, SoftMan can use genetic algorithms to elect the auction queue and bidding strategies. The experiment showed that the strategy could increase the bidding success rate.

关 键 词：多重拍卖 软件人 遗传算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]