RAMASESH, RANGA VENKATESH; PHD

                         THE PENNSYLVANIA STATE UNIVERSITY, 1988
 
                         BUSINESS ADMINISTRATION, MANAGEMENT (0454)
 

                         In inventory management and control, when lead times are stochastic, a dual-sourcing technique in
                         which the order quantity is procured by placing split-orders on two vendors could offer savings in holding
                         and shortage costs, compared to the sole-sourcing technique. Although several studies have examined
                         the costs and benefits associated with these two approaches in a qualitative and empirical way,
                         theoretical research on this issue is scarce. Research in this area is significant because, in the
                         professional literature there is a considerable debate over the relative merits of the two techniques. In
                         this dissertation, we formulate mathematical models of one- and two-vendor inventory systems under
                         stochastic lead times and demand and examine their optimal total cost performance in the framework of
                         single- and multifactor experiments. We first analyze a simplified base-case model assuming a uniform
                         probability distribution for the lead times and a constant rate of demand. For the two-vendor system, we
                         assume that the lead times for both the vendors are independent and identically distributed and that the
                         order quantity is equally split between the two vendors. We then progressively relax the assumptions
                         and develop models of more complicated and realistic inventory systems. Besides the uniform
                         distribution model, we investigate two additional models with exponentially distributed lead times: in one
                         the lead times for the two vendors are identical and in the other they are different. For each model, our
                         investigation covers the development of the mathematical model, formulation of the expression for the
                         total expected cost as a function of the decision variables, derivation of optimal solutions, and
                         experimentation over a wide range of parameter values, under both and dual-sourcing approaches.
                         Finally, the experimental results are compared to evaluate the performance of the two techniques, and to
                         develop guidelines for a cost-effective choice between the two.

Social Systems Simulation Group P.O. Box 6904 San Diego, CA  92166-0904 Roland Werner, Principal Phone/FAX  (619) 660-1603