We study the effect of strategic consumer behavior on pricing, inventory decisions, and inventory release policies of a monopoly retailer selling a single product over two periods facing uncertain demand. We consider the following three-stage two-period dynamic pricing game. In the first stage the retailer sets his inventory level and inventory release policy; in the second stage the retailer faces uncertain demand that consists of both myopic and strategic consumers. The former type of consumers purchase the good if their valuations exceed the posted price, while the latter type of consumers consider future realizations of prices, and hence their future surplus, before deciding when to purchase the good; in the third stage, the retailer releases its remaining inventory according to the release policy chosen in the first stage. Game theory is employed to model strategic decisions in this setting. Each of the strategies available to the players in this setting (the consumers and the retailer) are solved backward to yield the subgame perfect Nash equilibrium, which allows us to derive the equilibrium pricing policies. This work provides three primary contributions to the fields of dynamic pricing and revenue management. First, if, in the third stage, inventory is released to clear the market, then the presence of strategic consumers may be beneficial for the retailer. Second, we find the optimal inventory release strategy when retailers have capacity limitation. Lastly, we numerically demonstrate the retailer's optimal decisions of both inventory level and the inventory release strategy. We find that market clearance mechanism and intermediate supply strategy may emerge as the retailers optimal choice.

We study the effect of strategic consumer behavior on pricing, inventory decisions, and inventory release policies of a monopoly retailer selling a single product over two periods facing uncertain demand. We consider the following three-stage two-period dynamic pricing game. In the first stage the retailer sets his inventory level and inventory release policy; in the second stage the retailer faces uncertain demand that consists of both myopic and strategic consumers. The former type of consumers purchase the good if their valuations exceed the posted price, while the latter type of consumers consider future realizations of prices, and hence their future surplus, before deciding when to purchase the good; in the third stage, the retailer releases its remaining inventory according to the release policy chosen in the first stage. Game theory is employed to model strategic decisions in this setting. Each of the strategies available to the players in this setting (the consumers and the retailer) are solved backward to yield the subgame perfect Nash equilibrium, which allows us to derive the equilibrium pricing policies. This work provides three primary contributions to the fields of dynamic pricing and revenue management. First, if, in the third stage, inventory is released to clear the market, then the presence of strategic consumers may be beneficial for the retailer. Second, we find the optimal inventory release strategy when retailers have capacity limitation. Lastly, we numerically demonstrate the retailer's optimal decisions of both inventory level and the inventory release strategy. We find that market clearance mechanism and intermediate supply strategy may emerge as the retailers optimal choice.

No description available.

This thesis investigates different aspects of competition under uncertainty using the tools of game theory. In Chapter 1, I consider a quantity oligopoly game. One of the firms is presented with an opportunity to commit to some output before the demand becomes known, but may add to it afterwards, then moving simultaneously with the rivals. I show that the more cost-efficient firm is more likely to behave like a Stackelberg leader, i.e. to produce the optimal Stackelberg leader quantity ex-ante and refrain from adding to it later, letting the rivals respond to its ex-ante output in the manner of Stackelberg followers. In Chapter 2, I study a model of an electoral contest. Two symmetric parties allocate their endowments to building platforms on various issues before the start of a campaign. Next, one of the issues becomes decisive in the course of the campaign with a commonly known probability. The outcome of the election depends on the difference in competence in this issue. I show that if the payoff functions are convex in this difference-the case of 'increasing returns to power'-parties differentiate each other by selecting different campaign issues. On the contrary, when the payoff functions are concave in this difference-the case of 'decreasing returns to power'-parties mimic each other by investing the same amounts into the same issues. Thus, incentives for selecting campaign issues depend critically on the shape of the payoff functions, which might be determined by (1) a non-linear technology transforming parties' investment in various topics into voters' perception of their competence, (2) or parties' inherent motivation for winning by a big margin due to parties' ideological convictions or rent-seeking, (3) or an electoral system giving winners or big parties a disproportionate advantage in the assigned number of seats, (4) or a relatively high extent of power given to the winning party once in office.

Cette thèse étudie la conception d’optimisation de la flexibilité de la politique de logistique et de la capacité de logistique avec une demande incertaine. Ces flexibilités jouent un rôle crucial dans la performance en garantissant la qualité du produit et en maximisant le profit. Nous nous concentrons sur la conception de la capacité de politique de logistique et de la capacité de logistique. Cette recherche étudie trois problèmes d’optimisation: une conception de la politique de commande dans le contrat quantité-flexibilité et la conception de la capacité dans deux entrepôts.Nous examinons d’abord la chaîne logistique à deux échelons entre le constructeur automobile et le détaillant, où le détaillant achète des automobiles écologiques au constructeur automobile et reçoit la demande de ses clients respectueux de l’environnement. Nous établissons une politique de commande dans un contrat à flexibilité de quantité qui prend en compte les exigences écologiques. La politique considère que le constructeur automobiledétermine le niveau de verdissement et que le détaillant détermine le prix de vente au détail avant la signature du contrat. Nous construisons les modèles pour décrire ce contrat dans les chaînes logistiques décisionnelles décentralisées et centralisées. Nous optimisons le niveau de verdissement pour maximiser les profits du constructeur automobile et le prix de vente au détail avec une demande sensible au respect de l’environnement visant à maximiser lesprofits du détaillant. Nous considérons en outre la décision d’équilibre entre le niveau de verdissement et le prix de détail pour maximiser les bénéfices de la chaîne logistique.Nous concevons ensuite les capacité s des robots et des cueilleuses dans un système RMFS (Robotic Mobile Fulfillment System), présenté par plusieurs robots soulevant et transportant des é tagres de stockage de meubles depuis les grilles de stockage jusqu’aux préparateurs de commandes. Nous construisons des modèles de Markov de grande-dimension pour décrire ce système avec des classes de clients, calculons le débit de ce système en fonction du nombre de robots et fournissons des règles de conception permettant de déterminer le nombre optimal de robots et leurs capacités, en tenant compte du compromis entre des robots. Nous vérifions les résultats analytiques des modèles de Markov avec des simulations. Nous consiérons en outre RMFS à sélecteur multiple et étudions sa conception optimale. Un autre objectif de la conception des capacités dans RMFS est considéré. Nous construisons des modèles de réseau de files d’attente pour décrire le système RMFS à l’aide de deux protocoles de partage de robots pour les sélecteurs, proposons les algorithmes correspondants, effectuons des analyses numériques et évaluons les performances du système RMFS en calculant letemps de traitement. Nous calculons ensuite le nombre et la vitesse optimaux des robots et fournissons les règles de conception efficaces pour RMFS.Enfin, nous concevons le stockage public. La conception des entrepôts de stockage publics doit s’adapter aux segments du marché afin d’augmenter le revenu moyen dans un environnement de forte demande. Cet article présente un modèle de revenu intégré aux théories de la file d’attente et de la demande de prix afin de résoudre le problème de conception et de tarification des entrepôts de stockage publics. Nous considérons deux cas de demande dans le modèle, à savoir une demande exponentielle et une demande linéaire par morceaux.Nous développons également une solution basée sur des techniques de programmation dynamiques pour résoudre le problème. En utilisant les données d’un entrepôt, nous menons des expériences numériques. Les résultats montrent que notre approche peut améliorer les revenus attendus des entrepôts publics à forte demande de 16,6 % en moyenne. Nous réalisons en outre une analyse de sensibilité du prix et étudions la relation entre le revenu et le prix. / This thesis studies optimal models for the flexibility of supply chain policies and capacities with uncertain demand. This thesis investigates three optimization problems: An order policy design in the quantity-flexibility contract and the capacity design in robotic warehouses and self-storage warehouses.We first consider the two-echelon supply chain between the automobile manufacturer and the retailer, where the retailer purchases green automobiles from the automobile manufacturer and receives the green sensitive customer demand.We make an order policy in a quantity-flexibility contract that considers green sensitive demands. The policy considers that the automobile manufacturer determines the greening level and the retailer determines the retail price before establishing the contract. We build the models to describe this contract in both decentralized and centralized decision-making supply chains. We apply Stackelberg game to optimize the greening level for maximizing the automobile manufacturer's profit and optimize the retail price with green sensitive demand for maximizing the retailer's profit. We further consider the equilibrium decision between the greening level and the retail price for maximizing the profit of the supply chain.We then study capacities of robots and pickers in a Robotic Mobile Fulfillment System (RMFS), featured byseveral robots lifting and transporting movables storage shelves from storage grids to order pickers. We build high-dimension Markov models to describe this system with customer classes, calculate throughput of this system given the number of robots and provide design rules to determine the optimal number of robots and their capacities considering the trade-off between capacities of picker stations and robots. We verify the analytic results of Markov models with simulations. We further consider multiple-picker RMFS and study its optimal design. We consider another objective of designing capacities in RMFS. We build queue network models to describe the RMFS with two protocols in sharingrobots for pickers, propose the corresponding algorithms, conduct numerical analyses, and evaluate the performance of the RMFS by calculating throughput time. We then calculate the optimal number and velocity of robots and provide the effective design rules for RMFS.Finally, we study the self-storage warehouses. The design of self-storage warehouses needs to fit market segments to increase the average revenue in an environment of high demand. This thesis presents a revenue model integrated with queuing and price-demand theories to solve the design and pricing problem for self-storage warehouses. We consider two demand cases in the model, which are exponential demand and piecewise linear demand. We also develop a solution based on dynamic programming techniques to solve the problem. Using data from a warehouse, we conduct numerical experiments. Results show that our approach can improve the expected revenue of public storage warehouses with high demand by 16.6% on average. We further conduct a sensitivity analysis on price and investigate the relationship between revenue and price.

To handle uncertainties and variabilities in production demands, many manufacturing companies have adopted different strategies, such as varying quoted lead time, rejecting orders, increasing stock or inventory levels, and implementing volume flexibility. Make-to-stock (MTS) systems are designed to offer zero lead time by providing an inventory buffer for the organizations, but they are costly and involve risks such as obsolescence and wasted expenditures. The main concern of make-to-order (MTO) systems is eliminating inventories and reducing the non-value-added processes and wastes; however, these systems are based on the assumption that the manufacturing environments and customers' demand are deterministic. Research shows that in MTO systems variability and uncertainty in the demand levels causes instability in the production flow, resulting in congestion in the production flow, long lead times, and low throughput. Neither strategy is wholly satisfactory. A new alternative approach, multi-channel manufacturing (MCM) systems are designed to manage uncertainties and variabilities in demands by first focusing on customers' response time. The products are divided into different product families, each with its own manufacturing stream or sub-factory. MCM also allocates the production capacity needed in each sub-factory to produce each product family. In this research, the performance of an MCM system is studied by implementing MCM in a real case scenario from textile industry modeled via discrete event simulation. MTS and MTO systems are implemented for the same case scenario and the results are studied and compared. The variables of interest for this research are the throughput of products, the level of on-time deliveries, and the inventory level. The results conducted from the simulation experiments favor the simulated MCM system for all mentioned criteria. Further research activities, such as applying MCM to different manufacturing contexts, is highly recommended.

Face à un environnement incertain et une internationalisation croissante de la chaîne logistique, la planification industrielle et commerciale (PIC) permet d’adapter efficacement les capacités industrielles à la demande du marché. Dans cette étude, nous présentons un modèle original de PIC utilisant des contraintes de flexibilité, pour améliorer la coordination entre les fonctions commerciales et logistiques. Un premier modèle de simulation permet d’étudier la dynamique du système ainsi que l’impact des différents paramètres sur les performances en termes de coûts et de satisfaction client. Afin d’étudier les politiques optimales, nous proposons un nouveau modèle de simulation-optimisation multi objectif. Différentes méthodes d’optimisation sont comparées, et plusieurs recommandations sont émises pour l’implémentation pratique de notre solution. Enfin, nous comparons les performances de plusieurs politiques de gestion des stocks lorsqu’elles sont couplées avec notre méthode de PIC flexible. À partir de données réelles du constructeur automobile Renault, nous présentons une étude comparative détaillée. Nous proposons plusieurs préconisations pratiques sur le type de politiques à privilégier selon les caractéristiques du système. Ces travaux de recherche sont particulièrement pertinents et applicables à d’autres industries confrontées à de fortes exigences commerciales,une faible visibilité sur la demande future et des approvisionnements longs. / Face to uncertain environment and growing globalization of the supply chains, the salesand operations planning (S&OP) aims to adapt efficiently the industrial capacities to themarket demand. In this research, we present an original S&OP model that uses flexibilityconstraints to improve the coordination between sales and logistics functions. A first simulationmodel is developed to study the system dynamics and the impact of different parameters onsystem’s performance in terms of costs and customer satisfaction. We introduce a multiobjectivesimulation-optimization model to investigate the optimal policies. Several optimization methodsare compared and recommendations are given for the practical implementation of our solution.Then, we compare the performances of several policies for managing parts inventories whenthey are coupled with our flexible S&OP. Based on real data of the automobile manufacturerRenault, we present a detailed comparative study. We present several managerial insights on thetype of policies to favor depending on the system characteristics. This research is particularlyrelevant for other industries that face strict customer requirements, uncertain demand and longprocurement lead time.

[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.

[pt] Este trabalho apresenta uma proposta de aceleração da decomposição de Benders aplicada a uma versão mais geral e compacta (menos restrições e variáveis) do modelo de gestão de estoques, otimizado via programação estocástica de dois estágios que considera uma camada, um item, demanda incerta e política de controle (R, S). De maneira a ser possível considerar problemas de grande porte, foram aplicados os métodos L-Shaped tradicional com corte único e a sua forma estendida com múltiplos cortes. Resultados computacionais preliminares mostraram um substancial melhor desempenho computacional do método L-Shaped tradicional em relação à sua forma multi-cut L-Shaped, mesmo o primeiro necessitando de mais iterações para convergir na solução ótima. Tal observação motivou o desenvolvimento de uma nova técnica de aceleração da decomposição de Benders e de um conjunto de desigualdades válidas. Experimentos numéricos mostram que a abordagem proposta de combinar a técnica de aceleração elaborada com as desigualdades válidas desenvolvidas provê significativa redução do tempo computacional necessário para a solução de instâncias de grande porte. / [en] This dissertation presents a speed up proposal for the Benders decomposition applied to a more general and compact version (less constraints and variables) of inventory management model, optimized via two-stage stochastic programming, which considers one layer, one item, uncertain demand and control policy (R, S). In order to be possible to consider large scale problems, the L-Shaped traditional method with single cuts and its extended form with multiple cuts were applied. Preliminary computational results showed a substantially better computational performance of the traditional L-Shaped method in comparison to the multi-cut L-Shaped method, even with the first requiring more iterations to converge on optimum solutions. This observation led to the development of a new technique to accelerate the decomposition of Benders and a set of valid inequalities. Numerical experiments show that the proposed approach of combining the elaborate acceleration technique with the developed valid inequalities, provide significant reduction in the computational time required to solve large scale instances.