Reverse double auction mechanism: An efficient algorithm for E-commerce platform operations

Abstract

Transaction efficiency is critical to the success of e-commerce platforms. We propose a Reverse Double Auction Mechanism (RDouMech), an innovative auction-based algorithm, to revolutionize the transaction process in e-commerce operations. It incorporates unique features of reverse auction, multi-bids, and multi-constraints (i.e., market clearing, individual rationality, and incentive compatibility). This operational innovation facilitates market clearing by precisely matching buyers with sellers, leading to a balance between demand and supply, improved matching rates between buyers and sellers, and a significant increase in transaction prices on the platform. We further validate our algorithm using proprietary data from an online automobile transaction platform. Our results show that RDouMech raises the transaction revenues by 19.73% and doubles social welfare compared to the human expert-assisted artificial intelligence (AI) algorithm that is currently employed by the platform.