Conference Agenda

Connected and Autonomous Transportation (CAT) is expected to greatly improve the safety and efficiency of the transportation network. However, CAT requires, for example, road infrastructure, transfer hubs, and fueling stations to be ready for use by autonomous vehicles. In this paper, we investigate the problem of optimally locating such necessary enablers to facilitate CAT transportation.

We introduce a new optimization problem by integrating the location decisions of multiple enablers into a single optimization problem. We present a mixed-integer linear program and a heuristic approach to solve the problem.

In our computational study, we demonstrate the superiority of integrated decision-making over sequential decision-making within this problem setting and show that each enabler induces a rather different network configuration, particularly under limited investment budgets. Our findings suggest that while increased CAT cost benefits might elevate overall travel distances, the potential fuel efficiency of CAT could lead to reduced emissions. Furthermore, introducing financial incentives like tax cuts for CAT users might not significantly boost its uptake in long-haul transportation.

Less-than-truckload freight transport companies reduce transport costs of palette-sized shipments between customers by consolidating shipments at hubs. To transport shipments between hubs, trucks operate a network of multi-stop line hauls. On a strategic level, companies need to plan where to locate hubs and which customers to assign to each hub. These decisions then impact the tactical design of the line haul network. In the literature these two aspects are usually considered separately or the network design is strongly simplified. We consider an integrated problem, solutions to which lead to more cost-effective networks compared to optimizing both aspects individually.

To solve the problem, we propose a branch-price-and-cut approach to dynamically generate line haul variables and constraints connecting the customer assignments with the network design. A variable neighborhood search and various other algorithmic components speed up the solving process. We present computational results of our algorithm and practical insights from the solutions obtained.