Qualificação de doutorado do discente André Luyde, dia 26/04/2022 as 14:00.
Qualificação de doutorado do discente André Luyde, dia 26/04/2022 as 14:00.
Título: META-HEURISTIC ALGORITHM FOR A CLASS OF ELECTRIC VEHICLE ROUTING PROBLEM
Discente: André Luyde da Silva Souza
Orientador: Puca Huachi Vaz Penna
Link meet: https://meet.google.com/hoq-kfjx-ynb
Data: 26/04/2022
Horário: 14:00
Resumo:
This work regards a class of the Electric Vehicle Routing Problem composed of different variants of the problems with some restrictions. The problems described in the class are extensions of the Vehicle Routing Problem (VRP) added by constraints regarding electric vehicles and their charging restrictions. The problems presented are the Electric Vehicle Routing Problem (E-VRP), the Electric Vehicle Routing Problem with Time Windows (E-VRPTW), and the Electric Vehicle Routing Problem with Battery Swap Stations (E-VRP-BSS). To solve the class of problems, we propose an extension of the Variable Neighborhood Search (VNS) heuristic. Our algorithm differs from the classical VNS in the stage of the local search. In our algorithm, we replace the local search with the Randomized Variable Neighborhood Descent (RVND). To prove the efficiency of our algorithm, we select five benchmarks sets of instances, one of the E-VRP, subdivided into four scenarios, two of the E- VRPTW, a small-size, and a large-size benchmark, subdivided into three classes, and two for the E-VRP-BSS. We compare our results with the current best-known solutions available in the literature and other authors. Our algorithm can solve almost all small instances from the three problems of the class EVRP in the optimality. The VNS algorithm also showed efficiency in finding high-quality solutions to the problems from the class.