Enhanced Dynamic Discretization Discovery for the Continuous Time Load Plan Design Problem

Abstract

For less-than-truckload (LTL) freight transportation carriers, as shipment sizes are small relative to trailer capacity, profitability is driven by the ability to load trailers with shipments from multiple customers. However, LTL carriers are facing increased pressure from customers to provide short delivery times, which limits their ability to hold freight, a critical tactic for consolidation. As consolidation requires the coordination of shipment paths in both space and time, shorter delivery times necessitate planning methods that can accurately estimate consolidation opportunities in time. This paper adapts and enhances the dynamic discretization discovery method designed for solving the scheduled service network design problem, defined on an arbitrarily fine discretization of time, to produce exact and heuristic methods for the LTL analog, the continuous time load plan design problem. With instances from the literature and based on data from a U.S. LTL carrier, we provide computational evidence that the algorithm can effectively solve this problem. With an analysis of high-quality solutions to the instances based on LTL carrier data, we derive insights into terminal staffing policies.