Hybrid Column Generation and Genetic Algorithm Solutions for the Scaling Problem of Multi-Deposit Electric Vehicles

In recent years, the use of Electric Vehicles (EVs) has grown significantly in front
the need to reduce greenhouse gas emissions. This motivates the study of the Problem of
Electric Vehicle Scaling (EVSP), which seeks to optimize the operation of electric fleets. East
This paper proposes to investigate the Multi-Depot variant of EVSP (MD-EVSP), which considers multiple
garages with limited capacities. The objective is to compare three criteria: efficiency (times of
processing of the algorithms), efficacy (final values of the objective functions) and quality (occurrence
of violations of operational restrictions). Hybrid approaches will be implemented and tested: AG-
GC, which applies a Genetic Algorithm to generate initial solutions followed by Column Generation;
and GC-AG, which reverses this order. Both techniques will be evaluated in adapted instances of
previous studies. It is expected that the findings will contribute to economic and operational optimization
of electric fleets in urban scenarios and advance state-of-the-art in vehicle scaling