Daniel O. Martínez-Quezada, Cristián E. Cortés, Antonio Mauttone, Marcela A. Munizaga
Abstract
Public transport plays a vital role in societal well-being and climate resilience. Battery Electric Vehicles (BEVs), while central to decarbonizing transit fleets, introduce operational challenges, including limited range and reliance on charging infrastructure. This study addresses the Electric Transit Network Design and Frequency Setting Problem (E-TNDFSP) by integrating both user and operator perspectives into a multi-objective framework, solved using the ϵ-constraint method and a Non-dominated Sorting Genetic Algorithm II (NSGA-II). NSGA-II performs comparably to the ϵ-constraint method on small instances and demonstrates scalability in real-world applications, such as in the city of Rivera, Uruguay. Hypervolume indicators confirm consistent improvements across generations, with relative gains above 60 % beyond 100 generations in the Mandl instance. Results show that fleet requirements for BEVs increase by up to 71 % under depot charging to maintain the same service level as diesel fleets. However, this increase can be reduced to 42 % through opportunity charging supported by 70 chargers. Energy consumption and network resilience metrics reveal critical trade-offs among fleet size, charging infrastructure, and user cost, offering data-driven insights for infrastructure planning. The findings support the viability of NSGA-II as a robust decision-making tool in electric transit systems.
