We consider a single-part, two-echelon supply chain problem for spare parts. The network consists of a single manufacturing plant, a set of service centers (SCs) and a set of customers. Both echelons keep spare parts using the base-stock replenishment policy. The plant behaves as an M/M/1 queueing system and has limited production and storage capacity. Demand faced by each SC follows an independent Poisson process. The problem is to determine optimal location-allocation and optimal base-stock levels at both echelons while satisfying the target service levels and customer preferences of SCs. We develop a mixed integer non-linear programming model and use cutting-plane method to optimize the inventory-location decisions. We present an exact solution procedure for the inventory stocking problem and demonstrate the limitations of using traditional inventory models like METRIC-like and Approximate in case of high utilization rates. We show the effectiveness of our proposed cutting-plane algorithm and provide important managerial insights for spare parts management.

We consider a single-part, two-echelon supply chain problem for spare parts. The network consists of a single manufacturing plant, a set of service centers (SCs) and a set of customers. Both echelons keep spare parts using the base-stock replenishment policy. The plant behaves as an M/M/1 queueing system and has limited production and storage capacity. Demand faced by each SC follows an independent Poisson process. The problem is to determine optimal location-allocation and optimal base-stock levels at both echelons while satisfying the target service levels and customer preferences of SCs. We develop a mixed integer non-linear programming model and use cutting-plane method to optimize the inventory-location decisions. We present an exact solution procedure for the inventory stocking problem and demonstrate the limitations of using traditional inventory models like METRIC-like and Approximate in case of high utilization rates. We show the effectiveness of our proposed cutting-plane algorithm and provide important managerial insights for spare parts management.

Les problèmes d’optimisation discrète sont pour beaucoup difficiles à résoudre, de par leur nature combinatoire. Citons par exemple les problèmes de programmation linéaire en nombres entiers. Une approche couramment employée pour les résoudre exactement est l’approche de Séparation et Évaluation Progressive. Une approche différente appelée « Resolution Search » a été proposée par Chvátal en 1997 pour résoudre exactement des problèmes d’optimisation à variables 0-1, mais elle reste mal connue et n’a été que peu appliquée depuis. Cette thèse tente de remédier à cela, avec un succès partiel. Une première contribution consiste en la généralisation de Resolution Search à tout problème d’optimisation discrète, tout en introduisant de nouveaux concepts et définitions. Ensuite, afin de confirmer l’intérêt de cette approche, nous avons essayé de l’appliquer en pratique pour résoudre efficacement des problèmes bien connus. Bien que notre recherche n’ait pas abouti sur ce point, elle nous a amené à de nouvelles méthodes pour résoudre exactement les problèmes d’affectation généralisée et de localisation simple. Après avoir présenté ces méthodes, la thèse conclut avec un bilan et des perspectives sur l’application pratique de Resolution Search. / The combinatorial nature of discrete optimization problems often makes them diffi- cult to solve. Consider for instance integer linear programming problems, which are commonly solved using a Branch-and-Bound approach. An alternative approach, Resolution Search, was proposed by Chvátal in 1997 for solving 0-1 optimization problems, but remains little known to this day and as such has seen few practical applications. This thesis attempts to remedy this state of affairs, with partial success. Its first contribution consists in the generalization of Resolution Search to any discrete optimization problem, while introducing new definitions and concepts. Next, we tried to validate this approach by attempting to solve well-known problems efficiently with it. Although our research did not succeed in this respect, it lead us to new methods for solving the generalized assignment and uncapacitated facility location problems. After presenting these methods, this thesis concludes with a summary of our attempts at practical application of Resolution Search, along with further perspectives on this matter. / Thèse réalisée en cotutelle avec l'Université d'Avignon.

Le but du travail réalisé dans cette thèse est de fournir aux fournisseurs de services une solution qui est capable de non seulement déployer les services dans le cloud de façon économique, automatique, mais aussi à grande échelle. La première contribution traite le problème de la construction d'un nouveau service demandé par le client à partir de services disponibles dans le cloud de manière à satisfaire les exigences en termes de qualité de service ainsi qu'en termes de coût. Nous présentons ce problème par un modèle analytique et proposons un algorithme heuristique dont la performance est améliorée de 20-30% par rapport aux autres approches. La seconde contribution est une solution pour déployer les services dans le cloud en considérant les demandes des utilisateurs finaux. Pour assurer qualité de services dans une grande échelle, le service demandé est dupliqué et distribué dans le réseau; chacun de ses réplicas servira les utilisateurs à proximité. Le plan d'approvisionnement selon lequel le service est dupliqué dépend de sa demande, ce qui ne cesse pas de changer en quantité ainsi qu'en distribution, ce qui rend le problème plus compliqué. Nous proposons une solution qui est capable de s'adapter aux changements dans le réseau, y compris ceux des demandes de ses utilisateurs. Enfin, nous proposons un système basé sur OpenStack qui permet de déployer les services complexes dans un cloud qui couvre différente locations (cloud multi-site). A partir d'une demande du client, le système automatiquement calcule le plan d'approvisionnement optimal et le déploie en respectant les contraintes du client.

In this study, the design of a three-level distribution system is considered in which a single supplier ships a number of items to the retailers via a set of distribution centers (DC) and stochastic demand is observed at the retailers. The problem is to specify the number and location of the DCs, and the assignment of the retailers to the DCs in such a way that total facility, transportation, safety stock, and joint ordering and average inventory costs are minimized, and customer service requirements are satisfied. Single source constraints are imposed on the assignment of the retailers to the DCs. The integrated location/inventory model incorporates the inventory management decisions into the strategic location/allocation decisions by considering the benefits of risk pooling and the savings that result in the joint replenishment of a group of items. We develop two heuristic methods to solve the non-linear integer-programming model in an integrated way: (1) Improvement type heuristic, (2) Constructive type heuristic. The heuristic algorithms are tested on a number of problem instances with 81 demand points (retailers) and 4 different types of items. Both of the heuristics are able to generate solutions in very reasonable times. The results are compared to the results of the p-median problem and found that the total cost and the number of DCs can be lowered using our integrated model instead of the p-median problem. Finally, sensitivity analysis is performed with respect to the changes in inventory, transportation, and ordering cost parameters, and variability of the demand.

Cette thèse porte sur le développement de modèles mathématiques et d’algorithmes d’optimisation pour la conception de chaînes logistiques durables. Nous proposons des modèles mono-périodiques, multi-produits et multi-modes de transport à quatre niveaux (fournisseurs, unités de production, entrepôts et clients) couvrant les piliers économique et environnemental du développement durable. Les variables de décision concernent la localisation des sites logistiques intermédiaires (unités de production et entrepôts), les choix de technologie et de mode de transport, et la détermination des flux de produits. Un premier modèle est basé uniquement sur la minimisation des coûts totaux. Ce modèle est étendu au cas bi-objectif en considérant la minimisation des émissions de CO2. Nous proposons une procédure d’optimisation basée sur la recherche à voisinage large (LNS : Large Neighborhood Search). L’application de cette méthode à un problème à variables mixtes tel que la conception de chaîne logistique est inédite. Notre extension au cas bi-objectif fait intervenir l’algorithme récent de recherche locale multi-directionnelle. Les expérimentations numériques permettent d’évaluer la pertinence de nos modèles et de comparer les performances de nos algorithmes à celles d’un solveur du marché. / This thesis focuses on the development of mathematical models and optimization algorithms for the design of sustainable supply chains. We propose single-period, multi-commodity, multi-mode, four level models (suppliers, production facilities, warehouses and customers) covering economic and environmental pillars of sustainable development. The decision variables are related to the location of the intermediate logistics sites (production units and warehouses), the choice of technology and mode of transport, and the determination of product flow. A first model is based solely on minimizing total costs. This model is extended to bi-objective minimization by considering CO2 emissions. We propose an optimization procedure based on the Large Neighborhood Search (LNS) metaheuristic, which had almost never been applied to problems with mixed variables such as design supply chain. Our extension to the bi-objective case involves the use of the multi-directional local search (MDLS). Extensive numerical experiments assess the relevance of our model and compare the performance of our algorithms to those of a state-of-the-art solver.

Com o aumento do número de desastres e consequente incremento no número de pessoas vitimadas, a preparação para esses eventos é uma necessidade das sociedades modernas. Neste sentido, o planejamento das operações logísticas para atendimento as situações de emergências é uma atividade recente e pouco explorada na produção acadêmica. O objetivo deste trabalho é estabelecer uma metodologia para definir locais para o pré-posicionamento de materiais utilizados no socorro a populações afetadas por desastres através de um modelo de otimização estocástica de dois estágios e análise de decisão multicritério e que considerem parâmetros quantitativos e qualitativos. Com base nos custos de transporte e do não atendimento a demanda, e utilizando informações como mapeamentos de riscos; custos de transporte; histórico de ocorrências de desastres; cobertura geográfica; compras de materiais; capacidades de depósitos e de transporte, um modelo estocástico de programação linear minimiza os custos operacionais para abastecimento às vítimas. Uma análise detalhada sobre como atribuir penalidades para demanda não atendida também é apresentada. Devido à incerteza quanto a severidade de um desastre e a influência da mídia nas fases pós-desastres estes parâmetros são representados na forma de cenários. O resultado do modelo estocástico mostra a quantidade de locais e quais localidades minimizam o custo operacional. Após a obtenção desse resultado, uma nova etapa é utilizada para decisão de escolha do local, com a aplicação de modelo de decisão multicritério que considere, além dos valores obtidos pela modelagem, critérios subjetivos característicos a operações humanitárias. Os resultados finais mostram que modelos estocásticos promovem resultados mais confiáveis que os determinísticos, especialmente, em situações nas quais materiais disponíveis não podem atender toda a demanda e que a consideração de critérios qualitativos e quantitativos proporciona uma decisão mais robusta em operações humanitárias. / The increase in disasters and the consequent increase in the number of victims make it highly necessary to prepare for these events in modern societies. Logistics operations planning to meet emergencies is a recent activity and little explored in academic production. Our aim is to establish a method to locate pre-positioned materials used in disaster relief through a two-stage stochastic optimization model and a multi-criteria decision analysis that consider quantitative and qualitative parameters. Based on transportation and unattended demand costs, and using information such as risk mapping, transportation costs, historical occurrences of disasters, coverage, materials purchase, warehouses and transport capacities, a stochastic linear programming model minimizes the operating costs to supply the victims. A detailed analysis on how to assign penalties for unmet demand is also presented. Due to the uncertainty of the disasters severity and the influence of the media in phases after disasters, these parameters are represented as scenarios. The result of the stochastic model shows the quantity and the locations that minimize the operational cost. After this result, a new phase is applied for site selection, with the application of multi-criteria decision analysis that consider the values provided by the model and subjective criteria characteristic of humanitarian operations. The final results show that stochastic models promote more reliable results than deterministic ones, especially in situations in which the materials available cannot meet all the demand and that the consideration of qualitative and quantitative criteria provides better decisions in humanitarian operations.

O trabalho apresenta o desenvolvimento e a aplicação de um modelo de localização de centros intermediários de consolidação e redistribuição de cargas em um ambiente urbano brasileiro. O método integra o TransCAD e o OpenSolver e é aplicado à cidade de Curitiba, uma das dez mais populosas do Brasil. O método proposto é caracterizado como um modelo de localização-roteirização baseado em agrupamento e subsequente roteirização, identificado na literatura por cluster-first routesecond; a adoção deste ordenamento permite tratar o problema para o atendimento de muitos estabelecimentos, como os até 65 mil em alguns dos cenários no estudo de caso de Curitiba. Cada agrupamento representa os pontos a serem visitados em uma única viagem e o processo inicial tenta minimizar as distâncias entre os estabelecimentos de cada grupo; na fase seguinte o melhor roteiro é computado para cada grupo; a terceira etapa consiste em calcular, para cada grupo e candidato, a distância total percorrida na viagem; por fim, a implantação ou não dos candidatos a centros de distribuição é obtida com a minimização em um modelo de programação linear inteira dos custos de aquisição e de operação dos centros de distribuição e dos custos de transportes. A dissertação também aborda a crescente percepção da importância da logística urbana à qualidade de vida nas cidades onde o adensamento populacional acirra a disputa pelo espaço viário e o conceito de City Logistics, que delineia entre outras medidas o ambiente cooperativo no qual implantação de centros de distribuição urbanos deve ocorrer. / This work presents the development and application of a model for the location of intermediary consolidation and redistribution freight centers in Brazilian cities. The method integrates TransCad and OpenSolver, and its use was evaluated with data from the City of Curitiba one of the ten largest in Brazil. The proposed method is characterized as a location-routing model based on clustering and subsequent tour building known as cluster-first route-second. This enables dealing with problem instances containing as many as 65 thousand customers. Each cluster comprehends the points visited on a single trip and the initial process minimizes the distances between customers; the routes are calculated in the next phase and the third step consists in computing the total distance covered in each trip for every cluster and every candidate; finally, the implementation of each distribution center candidate is decided by minimizing the costs of acquisition, operation and distribution, using an integer linear programming model. The dissertation also highlights the growing realization of the importance of urban freight transport to quality of life, especially in cities where increasing population density intensifies the competition for road space, and City Logistics concepts, that outline among other measures the cooperative environment where implementation of urban distribution centers should occur.

Orientador: Orlando Fontes Lima Junior / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Civil, Arquitetura e Urbanismo / Made available in DSpace on 2018-08-07T10:44:37Z (GMT). No. of bitstreams: 1 Verroni_RodrigoEduardoDias_M.pdf: 1562833 bytes, checksum: 7a98b8ffa353b6165a3b4e06de1a640d (MD5) Previous issue date: 2006 / Resumo: Neste trabalho é apresentada uma aplicação de solução para o problema de implantação de instalações hierárquicas. Este modelo foi aplicado a um caso hipotético de localização de uma rede de mercados na cidade de Campinas, no qual são elaborados cenários distintos com relação aos raios de influência das instalações. A localização dos centros de atendimento será obtida como solução de um problema de máxima cobertura, no qual o objetivo é maximizar a população atendida pelos serviços oferecidos. Na revisão bibliográfica foram identificados os principais métodos matemáticos disponíveis na literatura e selecionado o de localização de instalações hierárquicas de dois níveis apresentado por ESPEJO (2001). A partir dos resultados obtidos é efetuada ainda uma discussão sobre a variação da demanda atendida em relação aos raios de cobertura das instalações. Concluiu-se que o modelo selecionado apresenta resultados consistentes para o problema proposto e que a variação dos raios de influência alteram a localização das instalações / Abstract: This paper proposes a solution for the hierarchical facilities implementation problem. The hypothetical case application of the model is the location of a market network in the city of Campinas, in which different sceneries are elaborated in relation to the facilities rays of influence. The service center location is obtained as a solution for the maximum covering problem, in which the objective is to maximize the population assisted by the supplied services. The bibliographic revision identifies the main mathematical methods available in the literature. The two levels hierarchical facilities location model developed by ESPEJO (2001) was selected as the main method to be implemented in this paper. It is concluded that the selected model presents consistent results for the proposed problem, and that the variation of the values of the rays of influence modifies the localization of the facilities / Mestrado / Transportes / Mestre em Engenharia Civil

LKAB hanterar en av världens största malmkroppar vid en underjordsgruva i Kiruna. I samband med deras gruvbrytning uppstår stora mängder avfall som måste hanteras korrekt. De avfallsanläggningar som finns på området har idag platsbrist, därför finns incitament att bygga nya anläggningar. Inför utformningen av dessa nya anläggningar behöver avfallsflöden inventeras och analyseras. I detta arbete har det skett med diskret händelsesimulering baserad på historisk data. På grund av osäkra parametrar har flera scenarion tagits fram. Utformningen av anläggningarna beror på vilken kapacitet de förväntas ha, vilket är beroende av hur ofta anläggningarna ska tömmas och hur mycket avfall som förväntas uppstå på gruvområdet. För att sedan hitta bästa möjliga placering på de nya anläggningarna har ett modifierat anläggningslokaliseringsproblem tillämpats. Den optimala placeringen är i den norra delen av gruvområdet och minskar den totala körsträckan. Placeringen påverkas inte av större flöden. / In Kiruna the company LKAB has an underground mine with one of the largest iron ore quantities in the world. When producing the iron ore, a lot of waste arise that has to be taken care of in a correct way. Due to lack of space in todays waste facilities, there exists incentives to build new ones. To be able to design these new facilities, the flows of waste has to be investigated and analyzed. This has been done by discrete-event simulation based on historical data. Since some of the collected parameters are uncertain, different scenarios have been tested. The design of the facilities are dependent on what capacity they are expected to store, which is dependent on how often the facilities will be emptied and how large quantities will arise in the mining area. To find an optimal placement for the new facilities, a modified facility location problem has been implemented. The optimal placement is in the north part of LKAB’s mining area and will lower the total driving distance. The optimal placement will not be affected by larger flows.