作者: Craig Boutilier , Bikash Sabata , Moisés Goldszmidt
DOI:
关键词:
摘要: Market-based mechanisms such as auctions are being studied an appropriate means for resource allocation in distributed and inultiagenl decision problems. When agents value resources combination rather than isolation, one generally relies on combinatorial where bid tor bundles. or simultaneous all resources. We develop a different model, required reources sequentially. This model has the advantage that it can be applied settings models infeasible (e.g.. when made available at points time by parties), well certain benefits applicable. dynamic programming to compute bidding policies based estimated distributions over prices. also describe how these updated provide learning behavior.