Consensus-based decentralized auctions for robusttask allocation
IEEE Transactions on Robotics
An immune partheno-genetic algorithm for winner determination in combinatorial auctions
ICNC'05 Proceedings of the First international conference on Advances in Natural Computation - Volume Part III
| Hi-index | 0.00 |
Combinatorial auctions allow bidders to bid their synergistic values. Because of complementarities between different assets, bidders give their preferences not just for particular items but also for sets or bundles of items. This form of auction shows great potentialunder some given circumstances but is still in its infancy. The difficulty lies in finding the optimal price that is the price of a revenue maximizing set of winning bids. Determining the revenue maximizing set of winning bids is NP-complete. In this paper we review some of the existing approaches as detailed by Sandholm in one of his recent studies. We propose a heuristic, which is a lower bound on the maximal revenue, or an inadmissible heuristic. We then present an algorithm called, "Iterative Threshold Search (ITS) -- Hybrid". We show using this algorithm that, although inadmissible, such a heuristic near the optimal can be a better choice in practice than a surely admissible or upper bound heuristic. We establish through experiments that this new heuristic coupled with the proposed search algorithm improves the performance result significantly over the one presented by Sandholm.