Algorithms for infrastructure design for electric vehicles with long-term planning

• Ano de defesa: 2021
• Autor(a) principal: Corrêa, Victor Hugo Vidigal
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Universidade Federal de Viçosa
• Palavras-chave em Português: Veículos elétricos; Logística; Metaheurística; Ciência da Computação
• Link de acesso: https://locus.ufv.br//handle/123456789/28235

The modernization of our lifestyle, originating from the Industrial Revolution in the XVIII century, has led to successive changes in the consumption patterns of the society, raising them to levels never seen before by facilitating the population’s access to various goods in increasingly distant markets from the production centers. With the surge of vehicles powered by fossil fuels, it was possible to dispose of products more quickly to markets all over the world, but with the consequence of the emission, over the years, of tons of CO 2 into the atmosphere, accumulating and causing the phenomenon called global warming which, if not tackled, could cause several negative impacts on the planet’s ecosystem. As one of several measures proposed to combat global warming, the use of electric vehicles has gained a lot of attention in recent years. Its use is particularly interesting for the transport sector, which is responsible for approximately 25% of global pollutant gas emissions. However, due to the particularities of the sector, the adoption of this means of transportation is still economically unattractive, especially due to the absence of an infrastructure network to efficiently operate its use. To optimize the construction of this infrastructure, this study aimed to present an algorithm aimed at reducing costs, considering a long-term planning of deliveries from a logistics company. Two algorithms are presented, one solves the short-term problem considering factors such as battery replacement, partial recharging, time windows and capacitated vehicles while the second, composed of 3 steps and based on the first, solves the problem considering long-term planning. The algorithms are tested with instances created based on real data from the United Kingdom and the State of Minas Gerais and, by the results obtained in the computational experiments, they were able to reduce the average costs in the long-term operational planning of deliveries by electric vehicles by 5.25% if compared to the algorithm for the problem that does not consider long-term planning. Keywords: Electric Vehicles. Logistics Infrastructure. Metaheuristics.