A SIMULATION STUDY OF THE COEFFICIENT OF VARIATION AS A MEASURE OF VARIABILITY OF REQUIREMENTS IN A MATERIAL REQUIREMENTS PLANNING ENVIRONMENT

The management of production inventories is a concern of production and operations managers. Among the decisions a manager must make is the selection of a procedure to determine the timing and size of replenishment orders. The selection of a procedure, in turn, is dependent upon the environment in which it will be applied. The research described in this dissertation is concerned with the adequacy of the coefficient of variation as a descriptor of a requirements vector in the selection of a lot sizing procedure within a Material Requirements Planning (MRP) environment.; To study the coefficient of variation, the characteristics of a requirements vector which contribute to "lumpiness" are described and defined. These characteristics are then modeled into requirements vectors developed using three different dynamic generation procedures. A computer simulation program implements MRP logic and determines total inventory cost for these vectors using a selected group of lot-sizing procedures.; Analysis of variance was used to analyze the results of the simulations using relative total inventory cost as the criterion. Relative total inventory cost is defined as the percentage deviation of a selected lot-sizing procedure from a base (Groff) procedure.; A finding of this research is that while the method used to generate requirements vectors and the grouping patterns of those vectors are significant factors when evaluating lot-sizing procedures, they do not change the conclusions reached using only the coefficient of variation.