Multi-carrier track capacity allocation in forward and spot markets of freight transport

Abstract

This dissertation addresses multi-carrier, multi-line train scheduling problems for forward and spot markets. Schedules resulting from solution of these train scheduling problems enable carriers to make customer commitments for serving forward contracts and to transport one-off-loads arising on spot markets. A train slot selection model based on multicommodity network flow concepts is developed for determining timetables for use in forward markets and a column generation methodology is proposed for its solution. The model considers needs of both shippers and carriers. The solution approach is embedded in a simulation-based iterative framework, where demand elasticity is explicitly treated. A combinatorial auction-based track capacity allocation framework is introduced to allocate residual track capacity among competing carriers seeking to run additional trains on an as-needed basis. Bid set construction techniques are proposed that allow carriers to express their preferences for track usage. A winner determination problem is formulated, solution of which provides the optimal allocation of track capacity among carriers. The potential benefits of collaborative operation among carriers in both markets were recognized. Collaborative decision-making (CDM) strategies are designed for scheduling trains to serve forward markets. Performances of these strategies are assessed in a carrier collaboration simulation-assignment framework. A train space leasing technique and a CA-based train slot creation approach are proposed to accommodate one-off-loads on previously scheduled and newly formed trains, respectively. Required techniques for bid set construction are developed. A WDP is formulated that seeks the optimal allocation of track access rights to bidders for the given bid sets. Implementation of the resulting train schedules will aid in creating efficient and cost-effective rail transport, resulting in a competitive and green alternative to truck transportation. Additionally, collaboration among competing carriers can lead to the formation of profitable trains that might otherwise have been underutilized and an ability of each carrier to serve a greater share of the freight market. The methodologies were specifically intended for solving large, real-world train scheduling problems.