Le véhicule électrique est actuellement au coeur des alternatives énergétiques qui permettent de faire face à la croissance du coût du carburant et au réchauffement climatique. En revanche, l’autonomie limitée des batteries des véhicules électriques et l’indisponibilité d’un nombre suffisant de bornes de recharge représentent des enjeux majeurs auxquels se trouvent confrontés les utilisateurs. Le déploiement des véhicules électriques doit alors passer par la conception et l’expérimentation des outils d’aide à la décision pour gestion optimisée et adaptée de l’écosystème du véhicule électrique. C’est dans ce contexte que s’inscrit cette thèse qui vise à fournir des outils d’aide à la décision pour l’optimisation des usages, de la recharge et des tournées des véhicules électriques dans le cadre industriel. Dans un premier temps, nous proposons d’étudier le problème d’optimisation conjointe de l’affectation et de la recharge des véhicules électriques. L’objectif de ce problème est de maximiser l’usage des véhicules électriques et minimiser les coûts de recharge tout en prenant en compte les contraintes d’ordre opérationnel et technique. Pour résoudre ce problème, nous proposons une méthode exacte et deux heuristiques. Nous comparons les performances de ces méthodes sur des instances réelles et d’autres aléatoires. Nous exposons ensuite plusieurs extensions au problème de base en intégrant de nouvelles fonctions objectif et de nouvelles contraintes. Nous étudions par la suite notre problème de point de vue ordonnancement et nous proposons une étude de complexité et des algorithmes d’approximation avec garantie de performance pour le problème d’ordonnancement d’intervalles sous contraintes d’énergie. Finalement, nous nous intéressons à un nouveau problème de construction de tournées pour une flotte mixte de véhicules électriques et thermiques. Pour résoudre ce problème, nous proposons des heuristiques et des méta-heuristiques hybrides et nous comparons les performances des différentes méthodes sur des instances généralisées de la littérature / Electric Vehicles may decrease transportation-related emissions and provide for less dependence on foreign oil. However, electric vehicles are still facing many weaknesses related to the high purchase prices, limited battery range and scarce charging infrastructure. The deployment of electric vehicles must then involve the design and the deployment of charging infrastructures. Within this study, the overall objective is to provide enhanced optimization methods and decision tools for electric vehicles assignment, charging and routing that are relevant to different real-world constraints. Firstly, we propose to study the joint scheduling and optimal charging of electric vehicles problem. This problem consists in assigning a set of already constructed routes to the available electric and conventional vehicles and in, simultaneously, optimizing the electric vehicles charging costs while ensuring that all constraints are satisfied. The objective of this problem is to maximize the use of EVs and to minimize charging costs. Secondly, we propose different extensions to our baseline problem and we vary the objective functions and the considered constraints. Moreover, our problem can be seen as a fixed interval scheduling problem with complementary constraints of energy. We propose then to study the complexity and the approximability of many variants of this new problem. Finally, we consider a new variant of the electric vehicle routing problem with a heterogeneous fleet of vehicles and we propose different heuristics and metaheuristics to solve it. We test the different solving methods on benchmark instances and we evaluate the efficiency of each method

Este trabalho foi motivado pela necessidade de otimização nos serviços de entregas urbanas combinados com o alto custo de implantação e customização da maioria dos pacotes disponíveis comercialmente, que muitas empresas se defrontam, na expectativa de obtenção de respostas rápidas e eficientes, numa base diária. Geralmente a utilização destes software requerem grandes investimentos de tempo e recursos. Os custos operacionais do transporte de cargas têm experimentado um aumento expressivo devido a fatores que, de uma forma ou de outra, impedem o fluxo eficiente dos veículos na rede, tomando evidente a necessidade de ferramentas flexíveis e efetivas. Vários estudos encontrados na literatura revelaram que fatores como restrições de circulação e velocidades nos arcos têm contribuído para aumentar distâncias de percursos e a frota de veículos. Geralmente, estas rotas são planejadas de forma simplificada, utilizando um fator de correção, que fornece uma solução aproximada. Um SIG foi objeto de estudo na operação de entregas urbanas, de forma a atingir os objetivos deste trabalho. O estudo de caso abordou os serviços de entregas de bebidas na cidade de São Carlos, onde foi aplicada a heurística de economias de Clarke & Wright implementadas no software TransCAD 3.2. Foram feitas diversas simulações, comparando os resultados das distâncias em rede com os valores das distâncias estimadas, como também com as distâncias percorridas pela empresa distribuidora. / This research was motivated by the necessity of optimization of urban delivery services and the high cost of implementation and customization of most routing packages commercially available. Moreover, the companies expect to obtain fast and efficient answers on a daily base. The use of some routing software generally requires significant investments of time and other resources. The operational costs of freight transport have had a remarkable increase due to factors which somehow restraint the efficient flow of the vehicles in a network, leading to the need of flexible and effective tools. Several studies in the literature have revealed that factors such as restrictions of circulation and speed on network contribute to increase the travel distances and the fleet size. Generally, these routes are planned in a simplified way, using a correction factor to get an approximated solution. This research has considered the use of Geographical Information Systems as a tool to achieve better results for routing delivery services. The case study was the delivery service of beverages in the city of São Carlos. The Clarke & Wright\'s heuristic of economies was irnplemented in the TransCAD 3.2 software. Several simulations were carried out, comparing the results of route length, considering network and estimated distances, as well as the real one traveled by the delivery company\'s vehicles.

Este trabalho foi motivado pela necessidade de otimização nos serviços de entregas urbanas combinados com o alto custo de implantação e customização da maioria dos pacotes disponíveis comercialmente, que muitas empresas se defrontam, na expectativa de obtenção de respostas rápidas e eficientes, numa base diária. Geralmente a utilização destes software requerem grandes investimentos de tempo e recursos. Os custos operacionais do transporte de cargas têm experimentado um aumento expressivo devido a fatores que, de uma forma ou de outra, impedem o fluxo eficiente dos veículos na rede, tomando evidente a necessidade de ferramentas flexíveis e efetivas. Vários estudos encontrados na literatura revelaram que fatores como restrições de circulação e velocidades nos arcos têm contribuído para aumentar distâncias de percursos e a frota de veículos. Geralmente, estas rotas são planejadas de forma simplificada, utilizando um fator de correção, que fornece uma solução aproximada. Um SIG foi objeto de estudo na operação de entregas urbanas, de forma a atingir os objetivos deste trabalho. O estudo de caso abordou os serviços de entregas de bebidas na cidade de São Carlos, onde foi aplicada a heurística de economias de Clarke & Wright implementadas no software TransCAD 3.2. Foram feitas diversas simulações, comparando os resultados das distâncias em rede com os valores das distâncias estimadas, como também com as distâncias percorridas pela empresa distribuidora. / This research was motivated by the necessity of optimization of urban delivery services and the high cost of implementation and customization of most routing packages commercially available. Moreover, the companies expect to obtain fast and efficient answers on a daily base. The use of some routing software generally requires significant investments of time and other resources. The operational costs of freight transport have had a remarkable increase due to factors which somehow restraint the efficient flow of the vehicles in a network, leading to the need of flexible and effective tools. Several studies in the literature have revealed that factors such as restrictions of circulation and speed on network contribute to increase the travel distances and the fleet size. Generally, these routes are planned in a simplified way, using a correction factor to get an approximated solution. This research has considered the use of Geographical Information Systems as a tool to achieve better results for routing delivery services. The case study was the delivery service of beverages in the city of São Carlos. The Clarke & Wright\'s heuristic of economies was irnplemented in the TransCAD 3.2 software. Several simulations were carried out, comparing the results of route length, considering network and estimated distances, as well as the real one traveled by the delivery company\'s vehicles.