摘要
组合拍卖是多A gen t系统中进行资源分配和任务分配的重要方法。递增叫价组合拍卖可减少买方对物品估价值的暴露。然而在现有的递增叫价组合拍卖中,效用随着时间复杂度的降低迅速减少。该文提出一种多活性级的递增叫价组合拍卖方法,并给出了理性买方策略。该方法是对iBund le拍卖方法的扩展。买方可对物品组合在不同的活性级上加价,使得拍卖兼顾了时间复杂度和效用。实验表明,在时间复杂度不超过iBund le的情况下,该方法拍卖的效用最大率、总效用率、卖方效用率三项指标都明显高于iBund le。
A combinatorial auction for the Combinatorial Allocation Problem can greatly simplify multl-agent resource and task allocations. The ascending price combinatorial auction reduces the valuation information revealing to the buyers. However, in current methods, the profit is reduced sharply when the time is limited. A new method was developing using a simple strategy for rational buyers as an expansion of the iBundle method. Buyers in the new method may change their activity level and adjust their biding price according to the activity level, which greatly reduces the temporal complexity of the method. Test results demonstrate that the method is more efficient in less time than iBundle.
出处
《清华大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2006年第4期547-550,共4页
Journal of Tsinghua University(Science and Technology)
基金
国家自然科学基金资助项目(60373079
60496323)
关键词
多AGENT系统
组合拍卖
资源分配
multi-agent system
combinatorial auction
resource allocation