Dynamic routing and service network design for less-than-truckload (LTL) motor carriers

This research tries to address the dynamic priority shipment routing problem and dynamic service network design problem for the less-than-truckload (LTL) carriers. First, described is a decision support tool to assist LTL managers in studying, analyzing and planning LTL operations so that scarce resources are used more effectively and efficiently. The decision support tool helps to understand the complicated interactions between the shipment route, closing rules, cost, and service level. Numerical experiments are done using the decision support tool to analyze the existing rules of LTL carriers and to understand their effect on the total cost of the system and the service level provided.; Currently, LTL carriers route both regular and priority shipments through their service networks using some fixed route patterns known as load plans. In this research, an alternative routing strategy for routing priority shipments in LTL networks is proposed. This strategy exploits the stochasticity and dynamism embedded in the routing process and utilizes the real time information about terminals to determine the shipment routes adaptively. The research shows that this strategy can be formulated as the problem of finding a dynamic shortest path problem over a network with random arc costs. An efficient algorithm is developed that can solve this optimization problem in real-time. Numerical testing using real data sets suggests that the proposed strategy can improve the level of service for priority shipments.; LTL carriers currently use ad hoc rules in deciding when there is enough capacity to close a trailer. In order to reduce the fixed and penalty costs incurred by the LTL carriers and to increase the service level provided by LTL carriers to customers, the decision should be optimized over time. In other words, the decision to dispatch a trailer should not only depend on the current shipment level, but should also vary dynamically based on time of day, day of the week, and seasonal effects. A dynamic control policy for dispatching a trailer over a single link is proposed in this research. This research provides an approach to estimate the shape of the recourse function. The dynamic control policy exploits the linearity of the recourse function estimated in solving the trailer dispatching problem efficiently. The algorithm is easy to implement and computationally fast and hence can be extended to solve large LTL networks. Experiments with the dynamic control policy show that the solutions obtained are very close to the optimal.