An Approximation Approach for Large-Scale Multi-unit Combinatorial Auctions with Reserve-Price Biddings

In this paper, an approach on an auction-based pricing mechanism is proposed that can be modeled as a variant of reserve price biddings on combinatorial auctions. In there, the combinatorial auction is extended to cover multi-unit scenario, which allows bidding for indistinguishable items to cover the case, for example, to assign an allocation of aggregated electricity in a day, considering electricity generation costs on the power suppliers. Although such a mechanism could be naively applied for dynamic electricity auctions and other various purposes, it is difficult to be applied for large-scale auction problems due to its computational intractability and theoretical limitations. In this paper, first a naive mechanism with reserve price bidding cannot be applied since it might violate the reserve price condition. Then, an extended pricing mechanism is introduced that employs an approximate allocation and pricing algorithm that is capable to handle multi-unit auctions with reserve price biddings, guaranteeing the reserve price condition. The algorithm is expected to efficiently produce approximation allocations that are necessary in pricing and it also behaves as an approximation of VCG(Vickrey-Clarke-Groves) mechanism satisfying budget balance condition and bidders' individual rationality without having single-minded bidders assumption.