Decentralized Combinatorial Auctions for Multi-Unit Resource Allocation

Abstract: Auction has been used to allocate resources or tasks to processes, machines or other autonomous entities in distributed systems. When different bidders have different demands and valuations on different types of resources or tasks, the auction becomes a combinatorial auction (CA), for which finding an optimal auction result that maximizes total winning bid is NP-hard. Many time-efficient approximations to this problem work with a bid ranking function (BRF). However, existing approximations are all centralized and mostly for single-unit resource. In this paper, we propose the first decentralized CA schemes for multi-unit resources. It includes a BRF-based winner determination scheme that enables every agent to locally compute a critical bid value for her to win the CA and accordingly take her best response to other agent's win declaration. It also includes a critical-value-based pricing scheme for each winner to locally compute her payment. We analyze stabilization, correctness, and consistency properties of the proposed approach. Simulation results confirms that the proposed approach identifies exactly the same set of winners as the centralized counterpart regardless of initial bid setting, but at the cost of lower total winning bid and payment.