University of California Transportation Center

Business to business (B2B) auctions have become a dominant mechanism used by large shippers to procure contracts for transportation services from logistics companies. The bid analysis problem is of critical importance to shippers and determines which contracts are assigned to specific carriers and at what price. In practice this problem is further complicated by the consideration of shipper business rules, such as restrictions on carrier numbers, limits on the number of individual packages awarded and preferences for incumbent carriers. This paper examines the case in which bidding packages are mutually exclusive. This is referred to as a non-combinatorial auction. In practice, this type of auction is preferred to a full combinatorial auction because it allows the auctioneer (the shipper) to maintain control of the packages and creates much less cognitive strain for bidders (trucking companies). A mathematical programming model for the bid analysis problem is presented and heuristic construction algorithms and Lagrangian relaxation based algorithms are developed to solve the problem. Numerical results show that our Lagrangian relaxation based heuristics perform better than other heuristics and that the solutions are very close to optimal.