Efficient, Verifiable and Privacy-Preserving Combinatorial Auction Design

We propose a construction to efficiently and securely compute a combinatorial auction (also referred as combinational auction) which is able to forbid participants (both auctioneer and the bidders) from learning unnecessary information except those implied in the output of the auction. The auctioneer and bidders are assumed to be untrusted, and they may misbehave throughout the protocol to illegally increase their own benefit. Therefore, we need to either prevent the misbehavior or detect the misbehavior. We achieve this goal by introducing a payment mechanism to control bidders' behaviors game-theoretically, and we further introduce a blind signature scheme to let bidders verify the authenticity of their payment reported by the auctioneer. Although a third-party signer is involved, he only signs a value blindly (i.e. without knowing the value) and is also untrusted. Moreover, our construction requires at most O(mn^2) rounds of the communication between bidders and the auctioneer where m is the total number of goods and n is the total number of bidders, and the extra computation overhead incurred by our design is very efficient.