Maximal Matching for Double Auction

We study the problem of mechanism design for a double auction market where multiple buyers and sellers buy and sell a commodity. We design and implement a matching algorithm that maximizes market liquidity, including the number of transactions and buy/sell-volume. We prove that, given the number of matches, the algorithm also maximizes auctioneer's profit. Based on the CAT Tournament (Trading Agent Competition Market Design) platform, we show with experiments that the new matching method not only increases market liquidity but also significantly improves market share and auctioneer's profit in the long term, compared with equilibrium matching, the most commonly used matching method.