The industrial gas industry represents a multi-billion dollar global market and provides essential product to manufacturing and service organizations that drive the global economy. In this dissertation, we focus on improving distribution efficiency in the industrial gas industry by addressing the strategic level problem of bulk tank allocation (BTA) while considering the effects of important operational issues. The BTA problem determines the preferred size of bulk tanks to assign to customer sites to minimize recurring gas distribution costs and initial tank installation costs. The BTA problem has a unique structure which includes a resource allocation problem and an underlying vehicle routing problem with split deliveries. In this dissertation, we provide an exact solution approach that solves the BTA problem to optimality and recommends tank allocations, provides a set of delivery routes, and determines delivery amounts to customers on each delivery route within reasonable computational time. The exact solution approach is based on a branch-and-price algorithm that solves problem instances with up to 40 customers in reasonable computational time. Due to the complexity of the problem and the size of industry representative problems, the solution approaches published in the literature rely on heuristics that require a set of potential routes as input. In this research, we investigate and compare three alternative route generation algorithms using data sets from an industry partner. When comparing the routes generation algorithms, a sweep-based heuristic was the preferred heuristic for the data sets evaluated. The existing BTA solution approaches in the literature also assume a single bulk tank can be allocated at each customer site. While this assumption is valid for some customers due to space limitations, other customer sites may have the capability to accommodate multiple tanks. We propose two alternative mathematical models to explore the possibility and potential benefits of allocating multiple tanks at designated customer site that have the capacity to accommodate more than one tank. In a case study with 20 customers, allowing multiple tank allocation yield 13% reduction in total costs. In practice, industrial gas customer demands frequently vary by time period. Thus, it is important to allocate tanks to effectively accommodate time varying demand. Therefore, we develop a bulk tank allocation model for time varying demand (BTATVD) which captures changing demands by period for each customer. Adding this time dimension increases complexity. Therefore, we present three decomposition-based solution approaches. In the first two approaches, the problem is decomposed and a restricted master problem is solved. For the third approach, a two phase periodically restricting heuristic approach is developed. We evaluate the solution approaches using data sets provided by an industrial partner and solve the problem instances with up to 200 customers. The results yield approximately 10% in total savings and 20% in distribution cost savings over a 7 year time horizon. The results of this research provide effective approaches to address a variety of distribution issues faced by the industrial gas industry. The case study results demonstrate the potential improvements for distribution efficiency. / Ph. D.

In recent years, the importance of incorporating uncertainty into planning models for logistics and transportation systems has been widely recognized in the Operations Research and transportation science communities. Maritime transportation, as a major mode of transport in the world, is subject to a wide range of disruptions at the strategic, tactical and operational levels. This thesis is mainly concerned with the development of robustness planning strategies that can mitigate the effects of some major types of disruptions for an important class of optimization problems in the shipping industry. Such problems arise in the creation and negotiation of long-term delivery contracts with customers who require on-time deliveries of high-value goods throughout the year. In this thesis, we consider the disruptions that can increase travel times between ports and ultimately affect one or more scheduled deliveries to the customers. Computational results show that our integrated solution procedure and robustness planning strategies can generate delivery plans that are both economical as well as robust against uncertain disruptions.

Telemetry units can be used to gauge inventory levels at customers. These readings can help prevent both stockouts and unnecessary deliveries. The research problem we address is where to place a limited number of telemetry units in order to reduce routing costs. Modeling this problem involves the consideration of both inventory theory as well as vehicle routing concepts. We model this problem with an integer program but solve with heuristics. Our results demonstrate that significant savings can be found with limited numbers of telemetry units. We then extend our results to consider the impact of correlation of customer usage on the placement of telemetry units and show even greater savings can be obtained. We also present a model that can be used for locomotive fuel planning. It decides where fuel trucks should be located as well as the volume of the fuel that should be delivered to each locomotive.

This research deals with chemical transport Problems involving maritime pick up from and delivery to storage tanks that are continuously filled and drained. More specifically, we developed decision technology to determine the efficient use of multi compartment bulk ships to transport chemical products while ensuring continuous production with no stock-outs, so that the inventory level of chemical products in storage tanks are maintained between prescribed upper and lower stock levels during the planning horizon. Due to the nature of the products, it is impossible to carry more than two products without these being separated into dedicated compartments of the ships. We need to decide how much of each product to carry, on which ship, subject to the conditions that all harbors must have sufficient product to meet demand, and the stock levels of the products cannot exceed the inventory capacity of that harbor. We have formulated this ship-routing problem as a combined multi-ship pickup-delivery problem with inventory constraints. The original problem is a large-scale non-convex mixed-integer programming problem. All non-convexities involved weighted sums of products of two variables, one of which is binary and the other is continuous but bounded. We have shown that the structure gives rise to an equivalent large-scale linear mixed-integer programming problem (MILP). We studied the underlying structure of the MILP and developed a solution strategy by Lagrangian relaxation method for this large scale MILP with special structure. We also devised heuristic methods that are fast and find a good solution and conducted numerical studies that show how good does the heuristic solution compared to the dual bounds.

In this study we consider an inventory routing problem in which a supplier distributes a single product to multiple retailers in a finite planning horizon. Retailers should satisfy the deterministic and dynamic demands of end customers in the planning horizon, but the retailers can backorder the demands of end customers considering the supply chain costs. In each period the supplier decides the retailers to be visited, and the amount of products to be supplied to each retailer by a fleet of vehicles. The decision problems of the supplier are about when, to whom and how much to deliver products, and in which order to visit retailers while minimizing system-wide costs. We propose a mixed integer programming model and a Lagrangian relaxation based solution approach in which both upper and lower bounds are computed. We test our solution approach with test instances taken from the literature and provide our computational results.

Les activités de transport et de gestion des stocks ont un impact important les unes sur les autres. Assurer un niveau de stock idéal peut demander des livraisons fréquentes, ce qui entraîne des coûts logistiques élevés. Pour optimiser les compromis entre les coûts de stock et de transport, des systèmes VMI (Vendor Managed Inventory) ont été développés pour gérer ensemble les opérations de stock et de transport. Pour un ensemble de clients ayant des demandes sur un horizon de temps, le problème de détermination des tournées et des quantités à livrer avec un coût minimum de gestion de stock et de transport est connu sous le nom de Inventory Routing Problem (IRP). Les systèmes à deux échelons ont également été étudiés pour améliorer le flux de véhicules dans les zones urbaines. étant donné que des nouvelles politiques de gestion sont apparues, dans le but de limiter le trafic des gros véhicules et leur vitesse dans les centres urbains, des Centres de Distribution (DC) sont mis en place pour coordonner les flux de marchandises à l'intérieur et à l'extérieur des zones urbaines. Les produits sont donc livrés aux clients par les fournisseurs via les DC.Nous proposons de combiner un système à deux échelons avec le IRP. Nous introduisons un Operational Two-Echelon Inventory Routing Problem (O-2E-IRP), ce qui est une nouvelle extension du IRP à notre connaissance. Dans le O-2E-IRP proposé, les clients doivent être servis par un fournisseur strictement via des DC et les tournées doivent être définis dans les deux échelons sur un horizon de temps donné. Trois politiques de réapprovisionnement et de configurations de routage différentes sont modélisées pour ce problème. Nous développons deux formulations mathématiques, ainsi qu'un algorithme Branch-and-Cut (B&C) combiné à une matheuristique pour résoudre le problème. De plus, nous analysons plusieurs inégalités valides disponibles pour le IRP et nous introduisons de nouvelles inégalités valides inhérentes au IRP à deux échelons. Des expériences de calcul approfondies ont été effectuées sur un ensemble d'instances générées de manière aléatoire. Les résultats obtenus montrent que les performances des méthodes sont liées à la politique de stock et à la configuration de routage.Dans le contexte d'un IRP à deux échelons, deux décisions tactiques importantes doivent être prises en plus des décisions de livraison de routage et de quantité de livraison: à partir de quel DC sera fourni chaque client et en utilisant quels véhicules ? Répondre à ces questions est extrêmement difficile car cela implique de pouvoir minimiser les coûts opérationnels d'un système de livraison VMI à deux échelons à long-terme et avec des demandes incertaines. Pour faire face à cela, nous présentons le Tactical Two-Echelon Inventory Routing Problem (T-2E-IRP) qui optimise les décisions en fonction d'un horizon à long-terme et en tenant compte des demandes stochastiques. Trois politiques de gestion des stocks sont modélisées et appliquées à un ou aux deux échelons. Nous développons une approche de simulation pour résoudre le T-2E-IRP sur un horizon de temps à long-terme. Nous proposons quatre formulations et deux algorithmes B&C pour définir l'affectation des clients et des véhicules aux DC en fonction d'un horizon de temps court. Ensuite, nous évaluons ces décisions d'affectation via un outil de simulation qui résout un sous-problème du T-2E-IRP, qui consiste en les décisions de livraisons du fournisseur aux DC et des DC aux clients, sur un horizon glissant. De nombreuses expériences sont effectuées pour un ensemble d'instances générées aléatoirement. L'impact de plusieurs paramètres utilisés pour déterminer l'affectation des clients et des véhicules aux DC sur le coût total est analysé. Basé sur des expériences, nous définissons la combinaison de paramètres qui fournit généralement les meilleurs résultats sur les instances générées. / Transport and inventory management activities have a great impact on each other. Ensuring an ideal inventory level can require frequent deliveries, leading to high logistics costs. To optimize the trade-offs between inventory and transportation costs, VMI (Vendor Managed Inventory) systems have been developed to manage inventory and transportation operations together. Given a set of customers with demands over a time horizon, the problem of determining routes and delivery quantities at a minimum inventory holding and transportation costs is known as Inventory Routing Problem (IRP). Two-echelon systems have also been studied to improve the freight vehicle flow inside urban areas. As new management policies have emerged, with the goal of limiting the traffic of large vehicles and their speed in urban centers, Distribution Centers (DC) are introduced to coordinate freight flows inside and outside the urban areas. Products are then delivered from the suppliers to the customers through the DC.We propose to combine a two-echelon system with the IRP. We introduce an Operational Two-Echelon Inventory Routing Problem (O-2E-IRP), which is a new extension of the IRP to the best of our knowledge. On the proposed O-2E-IRP, the customers must be served by a supplier strictly through DC and routes must be defined in both echelons over a given time horizon. Three different replenishment policies and routing configurations are modeled for this problem. We develop two mathematical formulations, and a Branch-and-Cut (B&C) algorithm combined with a matheuristic to solve the problem. In addition, we analyze several valid inequalities available for IRP, and we introduce new ones inherent to the IRP within two echelons. Extensive computational experiments have been carried out on a set of randomly generated instances. The obtained results show that the performance of the methods is related to the inventory policy and routing configuration.In the context of a two-echelon IRP, two important tactical decisions have to be taken in addition to route and quantity delivery decisions: from which DC will be supplied each customer and using which vehicles? Answering these questions is extremely difficult as it implies being able to minimize operational costs for a two-echelon VMI delivery system on long-term and with uncertain demands. In order to deal with this, we introduce the Tactical Two-Echelon Inventory Routing Problem (T-2E-IRP) that optimizes the decisions based on a long-term horizon and considering stochastic demands. Three inventory management policies are modeled and applied at one or both echelons. We develop a simulation approach to solve the T-2E-IRP on a long-term time horizon. We propose four formulations and two B&C algorithms to define the assignment of customers and vehicles to the DC based on a short time horizon. Then, we evaluate these assignment decisions through a simulation tool that solves a subproblem of the T-2E-IRP, which consists of the decisions of deliveries from the supplier to the DC and from the DC to the customers, on a rolling-horizon framework. Extensive computational experiments are performed for a set of randomly generated instances. The impact of several parameters used to determine the assignment of customers and vehicles to DC on the total cost is analyzed. Based on the experiments, we define the combination of parameters that generally provides the best results on the generated instances.

In this study, inventory&ndash / routing system of a company operating in Fast Moving Consumer Goods (FMCG) industry is analyzed. The company has decided to redesign distribution system by locating regional warehouses between production plants and customers. The warehouses in the system are all allowed to hold stock without any capacity restriction. The customers are replenished by the warehouse to which they have been assigned. Customer stocks are continuously monitored by the warehouse and deliveries are to be scheduled. In this multi&ndash / item, two-echelon inventory&ndash / distribution system, main problem is synchronizing inventory and distribution decisions. An integrated Mixed Integer Programming optimization model for inventory and distribution planning is proposed with the aim of optimally coordinating inventory management and vehicle routing. The model determines the replenishment periods of items and amount of delivery to each customer / and constructs the delivery routes with the objective of cost minimization. The integrated model is coded in GAMS and solved by CPLEX. The integrated inventory-routing model is simulated with retrospective data of the company. Computational results on test problems are provided to show the effectiveness of the model developed in terms of the performance measures defined. Moreover, the feasible solution obtained for a period is compared to the realized inventory levels and distribution schedules. Computational results seem to indicate a substantial advantage of the integrated inventory-routing system over the existing distribution system.

[pt] O transporte e os estoques correspondem a maior parte dos custos logísticos de uma empresa. Com o avanço da tecnologia, passou-se a analisar em conjunto esses dois componentes e não mais separados, como era feito anteriormente. O Problema de Roteirização de Veículos com Estoque (Inventory Routing Problem – IRP), nasceu dessa análise conjunta e procura encontrar a melhor rota para os veículos, atendendo a um determinado nível de estoque. Este trabalho apresenta um modelo de IRP com múltiplos veículos e produtos, onde existe a possibilidade de transbordo entre os centros de distribuição existentes. O modelo desenvolvido foi elaborado em um estudo de caso real em uma empresa do setor varejista. Após sua elaboração, o modelo foi testado com uma instância menor e comparado a situação atual da empresa, a fim de testar sua eficiência. Em seguida, foi rodado com os dados completos da empresa, e foram analisados os resultados. Na resolução, foi utilizado o software Xpress, o qual utiliza programação inteira como método de resolução. / [en] Transport and inventories account for most of a company s logistics costs. With the advancement of technology, we began to analyze these two components together and no longer separate, as was done previously. The Inventory Routing Problem (IRP) was born from this joint analysis and seeks to find the best route for the vehicles, meeting a certain level of inventory. This work presents an IRP model with multiple vehicles and products, where there is the possibility of transshipment between existing distribution centers. The developed model was elaborated in a real case study in a company of the retail sector. After its elaboration, the model was tested with a smaller instance and compared to the current situation of the company in order to test its efficiency. It was then run with the complete company data, and the results were analyzed. In the resolution, Xpress software was used, which uses integer programming as the resolution method.

Dans le problème de tournées avec gestion de stock ou "Inventory Routing Problem" (IRP), le fournisseur a pour mission de surveiller les niveaux de stock d'un ensemble de clients et gérer leur approvisionnement en prenant simultanément en compte les coûts de transport et de stockage. Etant données les nouvelles exigences de développement durable et de transport écologique, nous étudions l'IRP sous une perspective énergétique, peu de travaux s'étant intéressés à cet aspect. Plus précisément, la thèse identifie les facteurs principaux influençant la consommation d'énergie et évalue les gains potentiels qu'une meilleure planification des approvisionnements permet de réaliser. Un problème relatif à l'approvisionnement en composants de chaînes d'assemblage d'automobiles est tout d'abord considéré pour lequel la masse transportée, la dynamique du véhicule et la distance parcourue sont identifiés comme les principaux facteurs impactant la consommation énergétique. Ce résultat est étendu à l'IRP classique et les gains potentiels en termes d'énergie sont analysés. Un problème industriel de tournées avec gestion de stock est ensuite étudié et résolu, notamment à l'aide d'une méthode de génération de colonnes. Ce problème met en évidence les limitations du modèle IRP classique, ce qui nous a amené à définir un modèle d'IRP plus réaliste. Finalement, une méthode de décomposition basée sur la relaxation lagrangienne est développée pour la résolution de ce problème dans le but de minimiser la consommation énergétique / The thesis studies the Inventory Routing Problem (IRP) with explicit energy consideration. Under the Vendor Managed Inventory (VMI) model, the IRP is an integration of the inventory management and routing, where both inventory storage and transportation costs are taken into account. Under the new sustainability paradigm, green transport and logistics has become an emerging area of study, but few research focus on the ecological aspect of the classical IRP. Since the classical IRP concentrates solely on the economic benefits, it is worth studying under the energy perspective. The thesis gives an estimation of the energetic gain that a better supplying plan can provide. More specifically, this thesis integrates the energy consumption into the decision of the inventory replenishment and routing. It starts with a part supplying problem in car assembly lines, where the transported mass, the vehicle dynamics and the travelled distance are identified as main energy influencing factors. This result is extended to the classical IRP with energy objective to show the potential energy reduction that can be achieved. Then, an industrial challenge of IRP is presented and solved using a column generation approach. This problem put the limitations of the classical IRP model in evidence, which brings us to define a more realistic IRP model on a multigraph. Finally, a Lagrangian relaxation method is presented for solving this new model with the aim of energy minimization.

[pt] No cenário atual da economia mundial, é essencial aumentar a integração entre os diferentes atores da cadeia de suprimentos das empresas, reduzindo custos operacionais e melhorando a eficiência. O roteamento de navios é parte imprescindível dessa integração no comércio marítimo global, sendo objeto de estudo de muitos autores. Neste trabalho, apresentamos diferentes metodologias para resolver variantes do Problema de Roteamento Marítimo com Estoques. Esse problema envolve um grande número de variáveis e é computacionalmente complexo de ser resolvido. Nossa principal motivação é resolver um caso real de roteamento de navios de uma grande empresa do setor de Óleo e Gás, obtendo soluções de alta qualidade em tempos computacionais plausíveis e melhorando os resultados atuais da empresa. Todas as metodologias desenvolvidas são baseadas em uma combinação de uma meta-heurística com um modelo matemático de programação linear. Uma das principais diferenças entre as metodologias está no modelo matemático para resolver o problema de estoque, onde testamos abordagens de tempo discreto e tempo contínuo. As outras diferenças dizem respeito ao número de produtos avaliados (único ou múltiplos produtos) e à meta-heurística usada (heurística de busca local com um fator de probabilidade de Simulated Annealing ou Hybrid Variable Neighborhood Search). Para a metodologia que utiliza um modelo de tempo discreto, os resultados são satisfatórios, com violações baixas e pontuais do estoque em um tempo computacional aceitável. Para a metodologia que utiliza um modelo de tempo contínuo, os resultados são ainda melhores, uma vez que, em reduzido tempo computacional, as violações de estoque permanecem baixas ou inexistentes, dependendo do cenário avaliado e da meta-heurística utilizada. Os resultados obtidos neste trabalho são notáveis e permitem sua aplicação prática em casos reais. / [en] In the current scenario of the world economy, it is essential to increase the integration between the different players in the companies supply chain, reducing operational costs, and improving efficiency. Ship routing is a substantial part of this integration regarding global maritime commerce, being the object of study by many authors. In this work, we present different methodologies to solve variants of the Maritime Inventory Routing Problem. This problem involves a large number of variables and is a computationally complex problem to solve. Our primary motivation is to solve a ship routing real case of a large company in the Oil and Gas sector, achieving high-quality solutions in plausible processing times and improving companies current results. All developed methodologies are based on a metaheuristic combination with a linear mathematical model. One of the main differences between the methodologies lies in the mathematical model to solve the inventory problem, where we tested discrete-time and continuous-time approaches. Other differences concern the number of evaluated products (single or multi-product) and the metaheuristic used (local search heuristics with a Simulated Annealing probability factor or Hybrid Variable Neighborhood Search). For the methodology using the discretetime model, the results are satisfactory, with low and punctual inventory violations in an acceptable computational time. For the methodology using the continuous-time model, the results are better once, in reduced computational time, inventory violations remain low or non-existent, depending on the scenario evaluated and the metaheuristic used. The results obtained in this work are remarkable and allow its practical application for real cases.