Dynamic joint decision-making problems under uncertainty in retail supply chains

Managing how of a product smoothly within a retail supply chain is a challenging task due to various complexities arising at different levels in which a retailer needs to simultaneously make several (operational and/or strategic) decisions. The different decisions such as inventory, production, ordering and pricing of the product are inter-related and collectively impact on the performance of the firm interacting with different parties in the supply chain. The inter-connectivity of multiple decisions (so-called joint decision-making process) increases the complexity for management of the retail supply chain. Moreover, stochastic and dynamic nature of the retail supply chain as well as its underlying network and product characteristics add further complexities into the joint decision-making process. It is crucial to adopt coordinated and combined decision-making approaches to manage retail supply chains. In this thesis, we develop joint decision-making policies to enhance the operations management of retail supply chains under uncertainty. In particular, we are concerned with three different joint decision-making problems: i) production and pricing of a multigeneration product line, ii) ordering and markdown policies for a perishable product, and iii) ordering and inventory allocation strategies for a dual-channel supply chain. In the first problem, the firm releases a new version of a product periodically while its older versions continue to sell in the market whereas the retailer deals with a perishable product of fixed and short age in the second problem. While perishability of the product is analysed in view of demand variation in the first two problems, we consider a non-perishable product for the final problem about the dual-channel supply chain under both demand and supply uncertainties. In the dual-channel supply chain network, the firm procures the product from a regular and/or emergency supplier and distributes it through multiple channels. Stochastic dynamic programming is used to model the underlying decision-making problems of the retailer that aims to maximize the expected profit over a planning horizon. The stochastic dynamic models suffer from the curse of dimensionality because of the increasing sizes of state and action spaces. Thus, solving these problems is computationally intractable by using traditional solution approaches. We propose alternative novel approaches to solve these complex problems efficiently. Computational experiments are designed to illustrate performance of the joint decision-making models and the proposed solution approaches. In addition, we derive managerial insights to show the significance of dynamic joint decision-making process. The numerical results indicate that jointly taking different operational decisions outperform single decisions made in isolation. Moreover, they highlight further benefits of joint decision-making in efficiently tackling uncertainties and improving the overall performance of retail supply chains