The goal of this dissertation is to understand the market-based mechanisms that enable shippers to allocate lanes in an efficient way for truckload (TL) transportation procurement despite the self-interest of carriers. To understand the market-based mechanisms, we focus on proposing some novel models and mechanisms to enhance the use of combinatorial auction for TL transportation procurement. In this dissertation, our approach to gaining the understanding consists of three parts: 1. We develop a carrier optimal bid generation model for carriers (bidders) to discover the best sets of lanes to bid for at a given round. The optimal bid generation model simultaneously generates alternative tours and selects the most profitable package bid for the carrier under a myopic strategy, which has never been considered before. The simultaneous tour generation and selection significantly lessen the computational complexities of a carrier's optimization problem since it is unnecessary for the carrier to calculate the values for all possible packages. 2. We present an iterative combinatorial auction design that integrates the optimization problems for both the shipper and the bidders where the approximate dual prices derived from the result of a winner determination solution are used by the bidders in identifying profitable lanes. The auctions also allow the bidders to submit exclusive-OR (XOR) bids and are able to deal with some common business considerations. The extension of the concept of active bids enables this mechanism to effectively mitigate the exposure problem, the threshold problem, and the free-riding problem. Furthermore, both the shippers and the carriers are better off compared to multi-round auctions that do not integrate the shippers' and carriers' optimizations. 3. We extend a deterministic winner determination model to a two-stage stochastic winner determination model for TL transportation procurement under shipment volume uncertainty. We demonstrate that the value of the stochastic solution is always at least as good as one obtained by a deterministic model based on using expected shipment volumes. The sWDP model is to the best of our knowledge the first winner determination formulation of any kind that explicitly incorporates demand uncertainty.

The goal of this dissertation is to understand the market-based mechanisms that enable shippers to allocate lanes in an efficient way for truckload (TL) transportation procurement despite the self-interest of carriers. To understand the market-based mechanisms, we focus on proposing some novel models and mechanisms to enhance the use of combinatorial auction for TL transportation procurement. In this dissertation, our approach to gaining the understanding consists of three parts: 1. We develop a carrier optimal bid generation model for carriers (bidders) to discover the best sets of lanes to bid for at a given round. The optimal bid generation model simultaneously generates alternative tours and selects the most profitable package bid for the carrier under a myopic strategy, which has never been considered before. The simultaneous tour generation and selection significantly lessen the computational complexities of a carrier's optimization problem since it is unnecessary for the carrier to calculate the values for all possible packages. 2. We present an iterative combinatorial auction design that integrates the optimization problems for both the shipper and the bidders where the approximate dual prices derived from the result of a winner determination solution are used by the bidders in identifying profitable lanes. The auctions also allow the bidders to submit exclusive-OR (XOR) bids and are able to deal with some common business considerations. The extension of the concept of active bids enables this mechanism to effectively mitigate the exposure problem, the threshold problem, and the free-riding problem. Furthermore, both the shippers and the carriers are better off compared to multi-round auctions that do not integrate the shippers' and carriers' optimizations. 3. We extend a deterministic winner determination model to a two-stage stochastic winner determination model for TL transportation procurement under shipment volume uncertainty. We demonstrate that the value of the stochastic solution is always at least as good as one obtained by a deterministic model based on using expected shipment volumes. The sWDP model is to the best of our knowledge the first winner determination formulation of any kind that explicitly incorporates demand uncertainty.