DETERMINISTIC DEMAND MULTIECHELON PRODUCTION AND INVENTORY CONTROL MODELS

This dissertation develops four deterministic demand multistage serial production and inventory models where raw material, work-in-process, and finished product inventory are unified into each model. The models assume that a single product is manufactured through a multistage serial production process where production at each stage is in lot size with one set-up per manufacturing cycle. The partial movement of product from stage to stage in sub-batches of equal size is permitted. Raw materials are assumed to be procured from outside suppliers with zero lead time. The demand rate for finished product is known with certainty and is constant over time. The system variable cost for each model consists of an ordering cost and a carrying cost for each raw material, a set-up cost for the production process, a cost of moving each sub-batch through a production process, a carrying cost for each work-in-process inventory, and a carrying cost for a finished product. In addition there is a penalty cost for shortage of finished product wherever a shortage is permitted in the model. Each of these costs is either a fixed charge per order (or per set-up), or a linear charge over time and over the average quantity carried (or short) with the exception that a penalty cost for shortage of finished product (in the case of lost sales) is a one time charge based on the total amount short per scheduling period.;The original minimization problem of each of the models can be converted into a more tractable form. A branch and bound algorithm has been developed which guarantees convergence to an optimal solution to the transformed problem. A heuristic approach is also adopted as an alternative solution method. Computer programs for the solution algorithms are written. Two numerical hypothetical problems are given and solved. Hypothetical problems for the general J-stage production and inventory system have been randomly generated and solved by the branch and bound and the heuristic algorithm for analyzing the quality of their solutions. Significant production and inventory control models presented in the literature are compared with the general J-stage model in terms of the quality of solutions. An actual production and inventory environment of one of the product lines of a local company is studied and analyzed.;The first model is a two-stage production process with no shortage permitted. In this model a sharing of common raw materials between production stages is allowed. The second model extends the number of production stages in the first model to the general J-stage system where J is a finite number of production stages. However, raw materials that supply more than one production stage are not permitted. The third and the fourth models are the extension of the second model by allowing shortage of finished product with complete backorder in the third model and with lost sales in the fourth model. The objective is to determine the ordering policies, production lot size, the number of sub-batches and the shortage quantity (if any) for each model which minimize (or come close to minimizing) the system variable cost per year over an infinite planning horizon.