Ofra Amir edited In_this_paper_we.tex  over 9 years ago

Commit id: 73cb36ce3c1a9c2ff3e7e978f975966e232e5723

deletions | additions      

       

In this paper we map the MAPF problem to a {\em combinatorial auction} problem. Combinatorial auctions are auctions in which participants can place bids on ``bundles'' of items rather than individual items. Formally, a CA problem is defined by a set of participating agents $K$ and a set of items $M$ that are available for purchase. Agents place bids on bundles of items $B \subset M$. The auctioneer then determines an allocation of bundles based on agents' bids, such that no item is sold to more than one agent.  CAs are useful when agents have non-additive valuations of items. For example, an agent might have a high valuation for a bundle that includes a flight to a vacation destination and a hotel, but will not want to buy the hotel if it cannot get a good price for the flight. CAs have been used in many applications, including including airport time slot allocation~\cite{rassenti1982combinatorial} and transportation service procurement~\cite{song2003combinatorial}. Formally, in a CA the auctioneer puts up for sale a set of items $M$. Agents place bids on bundles of items $B \subset M$. The auctioneer then determines an allocation of bundles based on agents' bids, such that no item is sold to more than one agent.