Researchers Yue Su, Sophie N. Parragh, Nicolas Dupin, and Jakob Puchinger, affiliated with the Johannes Kepler University Linz and the Austrian Institute of Technology, have recently tackled a complex optimization problem in the realm of autonomous electric vehicle (AEV) fleet management. Their work, titled “The Bi-objective Electric Autonomous Dial-a-Ride Problem,” aims to improve the efficiency of AEV-based transportation services, such as ride-sharing or on-demand shuttles.
The classic dial-a-ride problem (DARP) involves optimizing routes for vehicles to transport people between pickup and drop-off locations. The researchers have introduced electric, autonomously driving vehicles into this problem, creating the electric autonomous dial-a-ride problem (E-ADARP). Unlike previous studies that combined routing costs and user satisfaction into a single objective, the researchers treated these as two separate objectives to be optimized concurrently. The first objective focuses on minimizing routing costs, while the second aims to minimize the total excess user ride time, ensuring passenger satisfaction.
To address this bi-objective E-ADARP, the team developed a novel exact framework called the “fragment-based checker.” This framework uses a smart “select-and-check” algorithm that iteratively constructs feasible solutions using fragments of routes. The researchers also proposed several enhancements to improve the computational efficiency of their method. They evaluated different variants of their checker algorithm using a previously developed branch-and-price algorithm and benchmarked their checker-based framework against state-of-the-art criterion space frameworks and a generalized branch-and-price algorithm.
The computational experiments demonstrated the effectiveness of the proposed framework. Out of 38 instances, 21 were solved optimally, with small-to-medium-sized instances resolved within seconds. For larger-scale instances, particularly those requiring high battery end levels, the approaches provided high-quality approximations of the Pareto frontiers, which represent the trade-offs between the two objectives. The researchers also compared efficient solutions with varying energy restrictions, offering valuable managerial insights for different kinds of service providers.
This research is particularly relevant to the energy industry as it provides a robust method for optimizing the use of electric autonomous vehicles in transportation services. By minimizing routing costs and excess user ride time, the framework can help reduce energy consumption and improve the overall efficiency of AEV fleets. The insights gained from this study can aid energy companies and transportation service providers in making informed decisions about fleet management and energy usage.
The research was published in the journal “Transportation Science,” a leading publication in the field of transportation research. The study’s findings contribute to the growing body of knowledge on optimizing electric autonomous vehicle fleets and offer practical applications for the energy sector.
This article is based on research available at arXiv.