[pt] É devido a um aspecto de inovação e mudança no comportamento do consumidor que o mercado varejista vêm se modificando ao longo do tempo e criando novos formatos de lojas físicas. Um exemplo, são as lojas de bairro, pequenas e especializadas, que visam o atendimento de conveniência e buscam aproveitar da melhor forma possível os espaços físicos em localidades estratégicas. Porém, para gerenciar o estoque de mercadorias em um espaço reduzido e limitado se faz necessário também focar na melhoria da gestão dos estoques, com o intuito de melhor atender aos clientes e de evitar desperdícios com excessos e perdas de vendas com as rupturas. Dessa forma, o objetivo deste trabalho é analisar o cenário dos estoques existentes de nove lojas de conveniência de uma rede varejista brasileira e, utilizando um modelo tradicional de gestão de estoques presente na literatura, encontrar respostas para as questões de, quando pedir e o quanto pedir de mercadoria. Além disso, esta pesquisa ainda acrescenta a indagação de o quanto deve ser mantido de estoque de segurança para evitar o desbalanceamento do mesmo e, para responder a estas perguntas, é necessário lidar com o equilíbrio entre custos, nível de serviço e espaço físico disponível. Os resultados do estudo agregam os valores encontrados com os cálculos quantitativos juntamente com uma análise qualitativa com relação ao espaço existente nas lojas analisadas. E, as conclusões foram que para o equilíbrio entre custo, nível de serviço e limitação de espaço físico os valores de ponto de reposição e tamanho de lote seriam suficientes para um nível de serviço de 95 por cento. / [en] It is due to an aspect of innovation and change in consumer behavior that the retail market has been changing over time and creating new physical store formats. One example is neighborhood stores, which are small and specialized, which aim to serve convenience and seek to make the best use of physical spaces in strategic locations. However, in order to manage the inventory of goods in a small and limited space, it is also necessary to focus on improving inventory management, in order to serve customers and to avoid excess waste and sales losses due to disruptions. Thus, the objective of this paper is to study the existing inventory scenario of nine convenience stores of a Brazilian retail chain and, using the traditional inventory management model found in the literature, find answers to the questions of when to order and how much to order of merchandise. In addition, this research further raise the question of how much safety stock should been kept to avoid unbalance, and in order to answer these questions, it is necessary to address the balance between costs, service level and available physical space. The study results aggregate the values found with the quantitative calculations together with a qualitative analysis regarding the existing space in the analyzed stores. In addition, the conclusions were that for the balance between cost, service level and physical space limitation the replacement point values and lot size would be sufficient for a 95 per cent service level.

Il est essentiel d’étudier la dynamique non linéaire en mathématiques et nous permet d’interpréter des phénomènes aussi irréguliers et imprévisibles à la suite de processus déterministes, et non de simples erreurs ou chances statistiques.La dynamique du système est une méthodologie ainsi qu’une méthode pour implémenter des mouvements dynamiques non linéaires. Par ces deux méthodes,nous capturons la crise dans les cycles économiques et testons l’accumulation de capital, et les changements techniques dans le cadre du modèle Macro Stock Management et les appliquons à ceux de l’économie sud-coréenne. Divers outils d’analyse des séries chronologiques sont utilisés pour estimer les effets de l’investissement sur les taux de profit et les effets de l’inégalité salariale et de la dette des consommateurs sur la demande intérieure depuis la crise financière mondiale de 2008. Le Structural Vector Auto-Regressive model impose des restrictions à long terme ou à court terme sur le système VAR, il utilise pour distinguer deux variables avec des caractéristiques similaires. Perron (1989) fait valoir que les tests de racine unitaire traditionnels tels que le test Augmenté Dickey-Fuller sont susceptibles de ne pas détecter la stationnarité ou la non stationnarité des données si elles ont des ruptures structurelles. Zivot et Andrews(1992) et Lumsdaine et Papell (1997) proposent le modèle de test racine unitaire avec ruptures structurelles endogènes. Le test Gregory-Hansen fournit des informations sur une rupture structurelle dans un test de co-intégration. Le modèle ARDL (Auto Regressive Distributed Lags) est utilisé pour capturer les relations à long terme entre les variables. / It is essential to study nonlinear dynamics in mathematics, and it gives us room to interpret such irregular and unpredictable phenomena as a result of deterministic processes, not mere statistical errors or chances. System dynamics is a methodology as well as a method to implement nonlinear dynamic movements.By these two methods, we capture the crisis in the business cycles and test capital accumulation, and technical changes within the framework of the Macro Stock Management model and apply them to those of the Korean economy. In the latter part of this study, various time series analysis tools are used to estimate the effects of investment on the rates of profit and the effects of wage inequality and consumers’ debt on domestic demand since the global financial crisis of2008. Because the Structural Vector Auto-Regressive (SVAR) model impose on the long run or short run restrictions on the VAR system, it uses to distinguish two variables with similar characteristics, such as the after-tax rate of profit and the rate of profit after payment of interests and dividends excluding taxes (the after-payment rate of profit). Perron (1989) argues that the traditional unit root tests such as the Augmented Dickey-Fuller test are likely to fail to detect the stationarity or non-stationarity of the data if it has structural breaks. Zivot and Andrews (1992) and Lumsdaine and Papell (1997) propose the unit root test of the time series data with a single or two breaks in the time series. The Gregory-Hansen test provides information about a structural break in a cointegration test. The Auto Regressive Distributed Lags (ARDL) model is used to capture long-run relationships between variables.

[pt] Elevado custo com estoques é um problema enfrentado por muitas empresas brasileiras. Reduzir esse custo mostra-se como uma oportunidade, mas essa redução deve ser acompanhada de uma análise criteriosa, de maneira a não trazer consequências indesejáveis como parada de unidades ou perda de mercado. A presente dissertação teve como objetivos a proposição de um framework de gestão de estoques intermediários baseado em revisão sistemática de literatura seguida de um estudo de caso com entrevistas e observação participante, resultando na análise da gestão de estoques intermediários de uma empresa de óleo e gás. Uma comparação entre as práticas atuais da companhia e o framework foi realizada e recomendações para implantação listadas, através da proposição de um roadmap para implementação do framework. Apontamento de ganho de agilidade e foco na gestão dos estoques intermediários, e redução do inventário são elencados como contribuição desse estudo. As principais recomendações levantadas são referentes à realização da classificação ABC dos estoques, utilização de princípios e ferramentas da filosofia lean, realização de simulações, modelagem e mapeamento de fluxo de valor. Recomendações para estudos futuros incluem técnicas quantitativas de modelagem matemática de estoques e simulações. / [en] High inventory costs is a problem faced by many Brazilian companies. Reducing these costs can be very beneficial. This reduction needs a careful analysis so as not to bring undesirable consequences, including unit downtime or reduced market share. The goal of this dissertation is to propose an intermediate stock management framework. The proposal is based on a systematic literature review followed by a case study with interviews and participant observation. An analysis of the intermediate stocks management in an oil and gas company is produced. A comparison between the current practices of the company and the proposed framework was carried out. Recommendations and a roadmap on how to implement this framework are offered. Agility in managing intermediate stocks and inventory reduction are pointed as contribution of this study. The main recommendations of this study for the company are related to ABC inventory classification, lean principles and tools, simulation, modelling and value stream mapping. Recommendations for future studies include quantitative techniques for mathematical modeling of inventories and simulations.

La première partie de ce travail concerne la production cyclique pour l'optimisation du taux de production dans les flowshops robotisés, où un robot est chargé du transport des pièces. Les cellules robotisées peuvent être disposées de façon linéaire ou circulaire. Les principaux résultats théoriques concernant la disposition linéaire ne peuvent être étendus à la configuration circulaire. En particulier, trouver le meilleur cycle de production de une pièce (1-cycle) est un problème polynomial dans le cas des cellules linéaires additives, mais NP-difficile pour la configuration correspondante circulaire.Nous nous concentrons principalement sur le cas des cellules circulaires équilibrées, où le temps d'usinage est identique sur toutes les machines. Après avoir présentés des outils pour l'analyse cyclique dans les cellules circulaires, nous établissons des propriétés nécessaires des 1-cycles performants, ce qui permet de conclure sur le problème du meilleur 1-cycle jusqu'à 8 machines. Toutefois, nous fournissons un contre-exemple pour 6 machines à la conjecture classique des 1-cycles, toujours ouverte dans cette configuration.Ensuite, nous étudions la structure des 1-cycles performants pour des cellules circulaires équilibrées arbitrairement grandes. Nous définissons et étudions les propriétés d'une nouvelle famille de cycles basée sur cette structure et formulons une conjecture sur sa dominance sur les 1-cycles qui conduirait à un algorithme polynomial pour le problème du meilleur 1-cycle dans ce cas. Cette structure permet de déterminer le meilleur 1-cycle jusqu'à 11 machines.Dans la deuxième partie, nous présentons le travail réalisé sur un problème industriel proposé par la SNCF dans le cadre du challenge ROADEF/EURO. Nous proposons un algorithme glouton pour ce problème combinant divers aspects de la gestion des trains au sein d'une gare. / The first part of this work deals with cyclic production for throughput optimization in robotic flow-shops, where a robot is in charge of the material handling of parts. Robotic cells may have a linear or a circular layout. Most theoretical results for the linear layout do not hold for the circular layout. In particular, the problem of finding the best one part production cycle (1-cycle), which is a polynomial problem for linear additive cells, has been proved NP-hard for the corresponding circular configuration.We mainly focus on a special case of circular balanced cells, where the processing times are identical for all machines. After presenting tools for cyclic analysis in circular cells, we study necessary properties of efficient 1-cycles. These results allow to conclude on the best one part production cycle for any parameters in circular balanced cells up to 8 machines. However, we provide a counter-example to the classical 1-cycle conjecture, still open for this configuration.Then, we study the structure of efficient one part production cycles in arbitrarily large circular balanced cells. We introduce and study a new family of cycles based on this structure, and formulate a conjecture on its dominance over one part-production cycles, which would lead to a polynomial algorithm for finding the best 1-cycle for circular balanced cells. This structure allows to settle the best one part production cycle for cells with up to 11 machines.In a second part, we present work on an industrial problem of railway stock scheduling proposed by the French railway company in the context of the ROADEF/EURO competition. We propose a greedy algorithm for this problem combining the various aspects of trains handling inside a station.

Denna undersökning handlar om hur data för ett lagerhanteringssystem, inom oljebranschen, kan lagras på ett mer effektivt sätt. Med hjälp av en databasmodell bidrar den här rapporten genom att visa nya lagringsmöjligheter som finns för företaget Preem. Förutom det ger undersökningen, bland annat, svar på hur skillnader mellan uppgiven (enligt leverantören) och uppmätt drivmedelsvolym för utförda leveranser kan beräknas, med hjälp av att begära information från en databas. Genom en bättre utförd datalagring kommer det vara lättare att hantera ett lager och kunna upptäcka när ett fel uppstår, som gör att informationen om volymen i lagret ändras felaktigt. Ett vanligt existerande problem är att två källor anger två olika resultat som berör lagermängden. Rapporten innehåller två framtagna databasmodeller samt inkluderar de teoretiska fakta som krävs för att förstå sådana modeller, och att kunna få en uppfattning för hur en sådan modell implementeras som en databas. Exempel på sådan information är vad en databas är, hur modellering av en sådan går till, bland mycket annat. Samtidigt framhåller denna undersökning idéer på vad det finns för fortsatta utvecklingsmöjligheter för Preem. / The purpose of this research is to study how a stock management system within the oil industry can be stored in a more efficient way. With the use of a database model, this study will show new data storage possibilities for Preem (one of the largest fuel companies in Sweden). The research will determine the difference between the fuel volume stated (by the supplier) and the measured volume (in the tank), by requesting information from a new designed database. With more effective data storage it will be easier to manage stock and to detect errors, which can interfere with the information about the volume in stock. Having two different sources of information that can give two different results for the stock quantity, causes these errors. The report contains two developed database models as well as the theoretical facts that is necessary for understanding such models and their implementation as a database. For example gives a definition of a database, and a description of the modelling process. This research is concluded by some recommendations for Preem’s future progress for the system.

[pt] A logística humanitária visa salvar vidas e aliviar o sofrimento em desastres. Além disso, suprimentos humanitários são itens essenciais para a sobrevivência humana (como exemplo, água e alimentos) e sua falta pode causar grande impacto. No entanto, os recursos disponíveis podem não ser suficientes para atender a demanda dos beneficiários. O custo de privação quantifica o sofrimento causado as vítimas de desastres pela falta desses itens. Assim, este trabalho objetiva analisar o efeito da inclusão dele na gestão de estoque de suprimentos de alívio a desastres. Para alcançar o objetivo proposto, a pesquisa analisa o estado da arte da modelagem deste custo por meio de uma revisão sistemática da literatura acadêmica. Um modelo de gestão de estoque considerando custos de privação é então formulado como um modelo estocástico de dois estágios. Ademais, as diversas funções de custo de privação encontradas na literatura são analisadas. Finalmente, o modelo é validado empiricamente através de dois estudos de caso. Os resultados sugerem que a escolha da função de privação pode impactar em até 99,9 por cento de diferença nos custos totais da gestão de estoques. O nível de atendimento gerado também possui uma alta amplitude quando as diferentes funções de privação encontradas na literatura são comparadas. Conclui-se que as funções geram níveis de atendimentos distintos, cabendo aos tomadores de decisão escolher qual se adequa melhor ao problema estudado. Como contribuição prática, destaca-se que o modelo de gestão de estoque considerando o custo de privação proposto pode auxiliar os tomadores de decisão e reduzir o sofrimento das pessoas atingidas por desastres. Como contribuição acadêmica, o estudo analisa e compara as principais funções de privação propostas na literatura, o que pode auxiliar a proposição de novas modelagens. / [en] Humanitarian logistics aims to save lives and alleviate human suffering in disasters. Furthermore, humanitarian supplies are essential items for humanity (such as water and food), and their lack can have a great impact. However, the available resources may not be enough to meet the beneficiaries needs. The deprivation cost quantifies the suffering caused to disaster victims by the lack of these items. Thus, this paper aims to analyze the effect of including deprivation costs in disaster relief supplies inventory management. To reach the proposed goal, the research analyzes the state of the art of its modeling through a systematic review of the academic literature. An inventory management model considering deprivation costs is then formulated as a two-stage stochastic model. Furthermore, the various deprivation cost functions found in the literature are analyzed. Finally, the model is empirically validated through a two case studies. The results suggest that the choice of the deprivation function can impact up to a 99.9 per cent difference in the total costs of inventory management. The service level generated also has a high amplitude when the different functions are compared. It is concluded that the functions generate different service levels, and it is up to the decision-makers to choose which one best suits the studied problem. As a practical contribution, we highlight that the proposed inventory management model considering the cost of deprivation may help decision-makers and reduce the suffering of people affected by disasters. As an academic contribution, the study analyzes and compares the main deprivation functions proposed in the literature, which can help to propose new models.

Dans cette thèse, nous avons étudié les problèmes de décisions séquentielles, avec comme application la gestion de stocks d'énergie. Traditionnellement, ces problèmes sont résolus par programmation dynamique stochastique. Mais la grande dimension, et la non convexité du problème, amènent à faire des simplifications sur le modèle pour pouvoir faire fonctionner ces méthodes.Nous avons donc étudié une méthode alternative, qui ne requiert pas de simplifications du modèle: Monte Carlo Tree Search (MCTS). Nous avons commencé par étendre le MCTS classique (qui s’applique aux domaines finis et déterministes) aux domaines continus et stochastiques. Pour cela, nous avons utilisé la méthode de Double Progressive Widening (DPW), qui permet de gérer le ratio entre largeur et profondeur de l’arbre, à l’aide de deux méta paramètres. Nous avons aussi proposé une heuristique nommée Blind Value (BV) pour améliorer la recherche de nouvelles actions, en utilisant l’information donnée par les simulations passées. D’autre part, nous avons étendu l’heuristique RAVE aux domaines continus. Enfin, nous avons proposé deux nouvelles méthodes pour faire remonter l’information dans l’arbre, qui ont beaucoup amélioré la vitesse de convergence sur deux cas tests.Une part importante de notre travail a été de proposer une façon de mêler MCTS avec des heuristiques rapides pré-existantes. C’est une idée particulièrement intéressante dans le cas de la gestion d’énergie, car ces problèmes sont pour le moment résolus de manière approchée. Nous avons montré comment utiliser Direct Policy Search (DPS) pour rechercher une politique par défaut efficace, qui est ensuite utilisée à l’intérieur de MCTS. Les résultats expérimentaux sont très encourageants.Nous avons aussi appliqué MCTS à des processus markoviens partiellement observables (POMDP), avec comme exemple le jeu de démineur. Dans ce cas, les algorithmes actuels ne sont pas optimaux, et notre approche l’est, en transformant le POMDP en MDP, par un changement de vecteur d’état.Enfin, nous avons utilisé MCTS dans un cadre de méta-bandit, pour résoudre des problèmes d’investissement. Le choix d’investissement est fait par des algorithmes de bandits à bras multiples, tandis que l’évaluation de chaque bras est faite par MCTS.Une des conclusions importantes de ces travaux est que MCTS en continu a besoin de très peu d’hypothèses (uniquement un modèle génératif du problème), converge vers l’optimum, et peut facilement améliorer des méthodes suboptimales existantes. / In this thesis, we study sequential decision making problems, with a focus on the unit commitment problem. Traditionally solved by dynamic programming methods, this problem is still a challenge, due to its high dimension and to the sacrifices made on the accuracy of the model to apply state of the art methods. We investigate on the applicability of Monte Carlo Tree Search methods for this problem, and other problems that are single player, stochastic and continuous sequential decision making problems. We started by extending the traditional finite state MCTS to continuous domains, with a method called Double Progressive Widening (DPW). This method relies on two hyper parameters, and determines the ratio between width and depth in the nodes of the tree. We developed a heuristic called Blind Value (BV) to improve the exploration of new actions, using the information from past simulations. We also extended the RAVE heuristic to continuous domain. Finally, we proposed two new ways of backing up information through the tree, that improved the convergence speed considerably on two test cases.An important part of our work was to propose a way to mix MCTS with existing powerful heuristics, with the application to energy management in mind. We did so by proposing a framework that allows to learn a good default policy by Direct Policy Search (DPS), and to include it in MCTS. The experimental results are very positive.To extend the reach of MCTS, we showed how it could be used to solve Partially Observable Markovian Decision Processes, with an application to game of Mine Sweeper, for which no consistent method had been proposed before.Finally, we used MCTS in a meta-bandit framework to solve energy investment problems: the investment decision was handled by classical bandit algorithms, while the evaluation of each investment was done by MCTS.The most important take away is that continuous MCTS has almost no assumption (besides the need for a generative model), is consistent, and can easily improve existing suboptimal solvers by using a method similar to what we proposed with DPS.

[pt] Apesar de existir na literatura modelos propostos para gestão de estoques, as premissas consideradas por tais modelos podem inviabilizar suas aplicações. Este trabalho propõe uma metodologia de programação estocástica para reposição e controle de estoques de produto único numa rede logística de duas camadas. O enfoque revisão periódica proposto pode considerar tanto atendimentos à demanda em atraso (backorders) como vendas perdidas (lost sales) sem restrição de pedidos pendentes. Além disso, a fim de alcançar um melhor nível de serviço para o cliente, é introduzida uma regra de rateio proporcional a quantidade faltante do item em estoque no centro de distribuição para atender simultaneamente a demanda de todos os varejistas, a qual é capaz de lidar com as alocações negativas da falta. A periodicidade e o nível alvo da posição dos estoques são determinados através de modelos de programação estocástica de dois estágios e de uma técnica baseada em simulação de Monte Carlo, conhecida como Sample Average Approximation, que levam em conta a natureza incerta dos níveis de demanda pelo item por meio da geração de conjuntos finitos de cenários. Os equivalentes determinísticos são apresentados como modelos de programação não-linear inteira mista e em seguida linearizados. Experimentos numéricos com a metodologia proposta para instâncias do problema geradas aleatoriamente demonstram seu potencial ao obter resultados com erros de aproximadamente 1 por cento. / [en] Although several methods for inventory management are proposed in the literature, the required assumptions can hinder their application in practice. This work proposes a methodology for stock replenishment in two-echelon logistic networks through stochastic programming, considering a single item, periodic review and uncertain demands. The proposed approach is flexible enough to consider backlogs and lost sales cases without limitations on the number of outstanding orders. Also, in order to achieve better customer service, we introduce a variable rationing rule for quantities of the item in short at the distribution center to meet simultaneously all the demands of the retailers, dealing with imbalances or negative allocations of quantities of the item in short. The optimal review periodicity and the target level for inventory position are determined through two-stage stochastic programming models and a Monte Carlo simulation based-technique, known as Sample Average Approximation, which takes into account the uncertain nature of the item demand levels through the generation of finite sets of scenarios. The deterministic equivalent models are presented as mixed-integer non-linear programming models, which are then linearized. Numerical experiments with the proposed approach for instances of the problem randomly generated shows its potential, as the errors of the obtained results are around 1 percent.

Household food waste levels pose a considerable problem in terms of sustainability. Food stock management applications for smartphones are interventions that support people in planning and keeping an overview of their food stock in order to reduce food waste. So far mainly their usability and effectiveness for reduction of food waste have been researched in HCI. This study aimed at investigating how these applications are being integrated into people’s food practices, and how their features contributed to the integration. In a three-week long field study seven participants used one of two applications in their daily lives. Through interviews and diary entries it was observed that some people integrated the applications into their food practices to replace other actions like checking one’s food stock. New connections to the food practices were created through expiration reminders and providing means to check the food stock from a distance. Reminders were seen as helpful even if not always necessary and can be seen as an opportunity to further support the integration process. The main issues for the integration were the high-effort adding processes and remembering to update the inventory after consumption.

This thesis work aims to improve the stock management system of Excillum AB, a Stockholm-based company that specializes in high-performance X-ray sources. The current stock management system has limitations, resulting in gaps when managing the inventory. To address this, the study focuses on developing a calculation model that takes into account various factors such as the global supply chain, customer expectations, service offerings, and financial concerns. The research questions revolve around understanding best practices in stock management from other companies, considerations for stock selection and ordering decisions, and additional parameters to ensure accurate safety stock calculations. The research involves carrying out a literature review, benchmarking practices from other organizations, analyzing Excillum's transaction data, and developing a reliable calculation model. However, some limitations exist, such as relying on Excel for calculations rather than changing values within the ERP system, which reduces demand projection accuracy due to high product customization. Human factors and unavoidable supply chain disruptions are not specifically addressed in the recommended approach. The proposed calculation model aims to improve stock management efficiency, ensure product availability, reduce stockouts, and ultimately contribute to Excillum's growth goals while meeting customer demands. Excillum can optimize stock levels and streamline inventory management by incorporating best practices and taking into account critical factors, establishing the way for long-term success and improved customer satisfaction. / Detta examensarbete syftar till att förbättra lagerhanteringssystemet för Excillum AB, ett Stockholmsbaserat företag som är specialiserat på högpresterande röntgenkällor. Det nuvarande lagerhanteringssystemet har begränsningar, vilket resulterar i luckor vid hanteringen av lagret. För att komma till rätta med detta fokuserar studien på att utveckla en beräkningsmodell som tar hänsyn till olika faktorer såsom den globala leveranskedjan, kundernas förväntningar, tjänsteerbjudanden och ekonomiska problem. Forskningsfrågorna kretsar kring att förstå bästa praxis inom lagerhantering från andra företag, överväganden för aktieval och beställningsbeslut och ytterligare parametrar för att säkerställa korrekta beräkningar av säkerhetslager. Forskningen innebär att man genomför en litteraturgenomgång, benchmarking praxis från andra organisationer, analyserar Excillumstransaktionsdata och utvecklar en tillförlitlig beräkningsmodell. Det finns dock vissa begränsningar, som att förlita sig på Excel för beräkningar snarare än att ändra värden inom ERP-systemet, vilket minskar noggrannheten i efterfrågeprojekteringen på grund av hög produktanpassning. Mänskliga faktorer och oundvikliga störningar i försörjningskedjan behandlas inte specifikt i den rekommenderade metoden. Den föreslagna beräkningsmodellen syftar till att förbättra lagerhanteringseffektiviteten, säkerställa produkttillgänglighet, minska lagerutbudet och i slutändan bidra till Excillumstillväxtmål samtidigt som kundernas krav tillgodoses. Excillum kan optimera lagernivåer och effektivisera lagerhantering genom att införliva bästa praxis och ta hänsyn till kritiska faktorer, etablera vägen för långsiktig framgång och förbättrad kundnöjdhet.