作者: David C. Parkes , Lyle H. Ungar
DOI:
关键词: Auction theory 、 Auction algorithm 、 Vickrey–Clarke–Groves auction 、 Vickrey auction 、 Mathematical optimization 、 Computer science 、 Generalized second-price auction 、 Combinatorial auction 、 Bidding 、 Resource allocation 、 Unique bid auction 、 Revenue equivalence
摘要: Combinatorial auctions, which allow agents to bid directly for bundles of resources, are necessary optimal auction-based solutions resource allocation problems with that have non-additive values such as distributed scheduling and task assignment problems. We introduce iBundle, the first iterative combinatorial auction is a reasonable agent bidding strategy, in this case myopic best-response bidding. Its optimality proved novel connection primal-dual optimization theory. demonstrate orders magnitude performance improvements over only other known auction, Generalized Vickrey Auction.