Optimizing Fuel Consumption and Pollutant Emissions in Truck Routing with Parking Availability Prediction and Working Hours Constraints
- Author(s): Vital, Filipe;
- Ioannou, Petros
- et al.
Published Web Locationhttps://doi.org/10.7922/G2S75DPP
The transportation sector is responsible for a significant part of the U.S.’s greenhouse emissions, with a considerable amount being generated by medium-and heavy-duty trucks. However, when it comes to the trucking industry, ‘green’ routing studies do not consider other critical practical factors, like working hours regulations and parking availability. Due to parking shortages, routes and schedules that do not account for parking availability may lead to last-minute changes that make them more polluting than expected. Similarly, working hours regulations influence the timing of required rest stops, which may force drivers to deviate from initially selected routes and schedules with negative consequences to fuel consumption and emissions.
This study addresses a variant of the shortest path and truck driver scheduling problem under parking availability constraints which focuses on optimizing fuel consumption and emissions by controlling the truck's travel speed and accounting for time-dependent traffic conditions. As it is impossible to be absolutely certain about the future parking availability of any location during planning, the case of stochastic parking availability was also studied. When studying the trade-offs between prioritizing emissions reduction or trip duration, it was found that although focusing on emissions reduction can increase trip duration significantly, this impact is greatly reduced when considering scenarios with limited parking availability. The problem formulation was further extended to model drivers’ possible recourse actions when unable to find parking and the ensuing costs. This formulation was used to study how the solutions are affected by the level of information provided to drivers. It was found that ignoring uncertainty in parking availability results in inconsistent performance even when restricting parking to periods when probability of finding parking is high. Furthermore, results might not reflect the intent of the cost function used, e.g., minimizing illegal parking events and/or the priority assigned to emissions reduction. Giving drivers full information about the probability of finding parking at any time/location significantly improves performance and reduces illegal parking-related risks, but also substantially increase problem complexity and computation time. Using full information regarding parking availability but restricting the parking times to high availability time-windows can reduce complexity while maintaining consistent, although reduced, performance.