The Pollution-Routing Problem with Speed Optimization and Uneven Topography

This paper considers a joint pollution-routing and speed optimization problem (PRP-SO) where fuel costs and CO2e emissions depend on the vehicle speed, arc payloads, and road grades. We present two methods, one approximate and one exact, for solving the PRP-SO. The approximate strategy solves large-scale instances of the problem with a tabu search-based metaheuristic coupled with an efficient fixed-sequence speed optimization algorithm. The second strategy consists of a tailored branch-and-price (BP) algorithm in which speed optimization is managed within the pricing problem. We test both methods on modified Solomon benchmarks and newly constructed real-life instance sets. Our BP algorithm solves most instances with up to 50 customers and many instances with 75 and 100 customers. The heuristic is able to find near-optimal solutions to all instances and requires less than one minute of computational time per instance. Results on real-world instances suggest several managerial insights. First, fuel savings of up to 53% are realized when explicitly taking into account arc payloads and road grades. Second, fuel savings and emissions reduction are also achieved by scheduling uphill customers later along the routes. Lastly, we show that ignoring elevation information when planning routes leads to highly inaccurate fuel consumption estimates.