A hybrid approach of linear programming and logic modeling for the market core of sealed bid auctions

This paper presents a hybrid approach for sealed bid auction that integrates linear programming and logic modeling techniques. A linear programming model for the sealed bid auction considers only prices for transaction arrangements and yields multiple market cores when goods from sellers are equally attractive to many buyers. The hybrid approach takes into account ordinal preferences of traders as well to find the best market core. On the one hand, the hybrid approach uses a linear programming model to maximize total surplus of market participants based on the bid and ask prices. On the other hand, it employs a logical inferencing approach to satisfy traders' ordinal preferences that are not included in a single combination of bid and ask prices. A constraint logic programming scheme, which combines a mathematical program with a logical inferencing technique in an integrated formalism, is introduced to implement the hybrid approach. The hybrid approach is validated through market simulations, where a set of computer-generated trading data is applied to both the hybrid approach and the linear programming model for their market performance comparison.