Demand Flexibility in Supply Chain Planning

This chapter begins with an introduction to the book’s scope and preliminary concepts applied throughout the book. We then present a set of basic, foundational, and classical models from the operations planning literature that serve as the underpinning of the work presented throughout the book. These models include the economic order quantity (EOQ), the newsvendor problem, the economic lot-sizing problem (ELSP), the knapsack problem (KP), the generalized assignment problem (GAP), and the facility location problem (FLP). The main results presented later in this book generalize these classical models to account for a planner’s ability to influence demands, which have traditionally served as fixed parameters in these foundational models.

This chapter considers the state of prior literature on operations models that account for demand flexibility. In particular, we focus on generalizations of the models discussed in Chap. 1 that treat demands as decision variables. These generalizations typically involve pricing models, and they provide a foundation for the models we will study in subsequent chapters.

This chapter discusses a generalized version of the economic order quantity (EOQ). In particular, we consider a situation in which a single inventory stage must select from among a set of demand streams, those which it will satisfy. Each demand stream carries with it a constant demand rate as well as a constant revenue rate. We consider several problem variants within this class, including problems with lot size and demand rate constraints.

This chapter deals with a generalization of the single-period newsvendor problem. We consider a setting in which a decision maker at a single stocking point must determine the stock level for a single product under uncertain demand. In addition to determining the item’s stock level, the decision maker must select a subset from a set of individual demands, each of which is uncertain and follows a particular probability distribution. Assuming normally distributed and independent demand streams results in a class of problems that are strikingly similar to the problems considered in the previous chapter, although the underlying model assumptions are quite different.

This chapter defines a model for production planning for a single product in a periodic setting, where the planner must select from a number of individual orders for the product. Associated with any order are a demand quantity, delivery period, and revenue. Acceptance of an order implies that it must be met on time and in full. Orders are placed in advance of the planning horizon, and the planner must determine which orders to accept as well as a production plan for meeting these orders on time over a finite horizon. Production in any period carries a fixed order cost as well as variable production costs, and inventory may be held from period to period, incurring an associated holding cost. The planner’s goal is to maximize profit from order acceptance decisions over the planning horizon.

This chapter delves into problems that require assigning customer demands to supply sources. When no fixed cost exists for using a supply source and supply sources are capacitated, we have a generalized assignment problem (GAP). When fixed costs are incurred for using a supply source, then the problem is a classical facility location problem (FLP). We consider both capacitated and uncapacitated facility location problems, as well as the implications of requiring single sourcing constraints that do not allow splitting a demand between supply sources. Within these problem classes, we analyze two different forms of demand flexibility. The first type of flexibility corresponds to what we saw in the last two chapters, i.e., each demand must be either fully accepted or rejected. The second type of flexibility requires satisfying each demand, but the quantity (e.g., size, number of units) at which a demand is satisfied must fall between prespecified lower and upper limits, while revenue is proportional to the level at which the demand is satisfied. This chapter defines several such assignment and location models containing these dimensions of demand flexibility.

This chapter discusses the limitations of the models we have discussed throughout the book, and uses these limitations to characterize challenging future research directions. In doing so, we discuss the relation of the models we have considered to practice, and their potential for use in practical applications. We also consider the potential for expanding the definition of demand flexibility, as well as the technical difficulties inherent in meeting the challenges implied by potential future research avenues.