| In traditional settings, companies, assuming that the demand side of the system is uncontrollable, try to meet the demand for their products and services in an attempt to maximize their profits by focusing on decisions only on the supply side. Thus, in order to increase their profits, they put most of their efforts on determining better, and hopefully, the best decisions on capacity planning, inventory control policies and many others that affect the supply availability. However, as a result of advancements in information technology in the last few decades, companies have obtained a better understanding of the demand side through the availability and the easy use of demand data. Consequently, they have realized that demand for products and/or services is directly correlated to the price and the quoted or promised delivery times. Practice and research have shown that companies with a good price and lead time policy have improved their system performance compared to others that have suffered from bad price and lead time policies.; In this dissertation, different dynamic price and lead time quotation strategies are considered for a make-to-order manufacturer as a demand management tool. In the first part, the focus is on a case where a single product is manufactured and the decision maker determines an appropriate price and lead time to quote for the next order request, based on the system congestion level at the time of the order arrival. In the second part of this study, quotation decisions for multiple products that serve the same fundamental need but differ in specific attributes (such as speed, capacity or quality) are considered. In such cases, the manufacturer may decide to make a quotation for a better product than the one requested by the customer, if it is more profitable to do so. This kind of products are referred to as "downward substitutable products". Thus, in addition to the price and lead time decisions, the manufacturer needs to determine the product type to quote in order to maximize his profits, when the products are downward substitutable.; These problems are formulated as Markov Decision Processes with an objective of maximizing the total discounted expected profit in the infinite horizon. Like many other MDP applications, the formulation for the downward substitutable products case suffers from the curse of dimensionality as the problem size gets bigger. The final part of this study addresses this issue by proposing a linear programming based approximation algorithm that yields a near optimal policy in significantly shorter time than the exact algorithm. The performance of this approximation is tested by comparing it to both optimal dynamic and static strategies and some simple heuristics. |
