Local Search Methods for the Winner Determination Problem in Multi-Unit Combinatorial Auctions

In this paper, we are interested in the winner determination problem in the multi-unit combinatorial auctions (MU-WDP). In this type of auctions, a set of units of goods has to be auctioned. Bidders request the number of desired units of each good and the total price for the complete bid. Each bid has to be discarded or fully accepted. The objective of the auctioneer is to maximize its revenue. The MU-WDP is known to be NP-hard. In this paper, we propose three local search methods: a local search (LS), a tabu search (TS) and a stochastic local search (SLS) for the MU-WDP. The proposed methods are evaluated on some benchmark problems. The experimental study shows that the SLS algorithm is able to find good solutions for the multi-unit winner allocation compared to both LS and TS methods. Further, experiments against the CPLEX 12.5, show the effectiveness of the proposed SLS method in finding good quality solutions in faster time.