| In this thesis, we analyze the effect of incorporating two types of information in the control of production-inventory systems: (1) advance information on the demand, and (2) information on the status of the production process. For the models studied in this thesis, demand arrives dynamically over time with stochastic inter-arrival times and the production process has a finite capacity and stochastic production times. The systems operate in a make-to-stock, fashion and unsatisfied demand is backlogged. In systems with advance demand information, the production facility receives information in the form of advance announcements and subsequent updates of order due dates. This information is not perfect because (a) customers may request an order prior to or later than the announced expected due date, (b) the time between due date updates is random, and (c) announced orders may be canceled prior to becoming due. Given the current inventory level and the number of announced various stages of update, the production facility is faced with the decision of whether or not to produce. We study the second type of information (information on the status of the production process) in a production-inventory system where the production facility consists of two stages in series. Given information on the amount of inventory at each stage and in the buffer of finished items, the, decision maker decides whether or not to release a new item to the production facility. For each model, we identify an optimal policy and also suggest easy-to-implement heuristics that perform almost as well as optimal policies. In numerical studies, we evaluate the benefit of using the information and also perform sensitivity analysis with respect to system parameters. |
