Dynamic Auction for Network Resource Allocation

We study network resource allocation under information asymmetry, i.e., users have private information regarding their utility functions. The service provider would like to design an incentive-compatible pricing scheme to maximize the social welfare. It is well known that there is a market mechanism for the network resource allocation achieving the first best under complete information.Under the market mechanism, the service provider only needs to decide on the resource allocation to maximize the aggregate surrogate utility of the users. However, user utility information is typically private in practice. With the information asymmetry, we propose a novel dynamic auction mechanism. Under our dynamic auction, the users must follow the monotone activity rule in the bidding process and the auctioneer updates the prices based on the submitted bids by maximizing the aggregate surrogate utility. Surprisingly, our dynamic auction mechanism is a first-best dominant-strategy incentive compatible mechanism. In particular, the equilibrium outcome is the same as the market equilibrium and budget-balanced. The auctioneer does not need the utility information of the users, i.e., the auction is robust to the user utility information.This is possible by exploiting of the aggregate surrogate utility maximization of the auctioneer and the monotone activity rule we imposed. Thus, our dynamic auction mechanism is practical as it is simple, detail-free and privacy preserving, enriching the mechanism design theory. Finally, our results can be potentially applied to other network resource allocation problems