THE SIZES OF SIMULATION SAMPLES REQUIRED TO COMPUTE CERTAIN INVENTORY CHARACTERISTICS WITH STATED PRECISION AND CONFIDENCE
Abstract
The design and operation of simulation models for studying management policies and other problems that involve complex systems of random variables is presented. A major use of such models lies in measuring parameters which help to identify and select among preferred management policies. The estimation of these parameters is a statistical problem, and one of the significant statistical decisions is the size of sample to be drawn from the si ulation model for making such estimates. Also studied are calculations of sample sizes, measured in number of simulated time periods, required to estimate certain parameters of selected inventory models with specified precision and confidence. In particular, the inventory models used include the classical zero procurement lead-time case, plus selected nonzero procurement lead-time cases having 2-, 5-, and 10-period lead-times. The parameters estimated are the mean number of shortages and of overages per time period, an overage being the amount of stock on hand at the end of the period, with negative amounts (or shortages) equated to zero. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1962
- Accession Number
- AD0286793
Entities
People
- Murray A. Geisler
Organizations
- RAND Corporation