Development of a Dynamic Simulation Filter
Abstract
The Filter Simulation Experimental Concept for reducing the cost of identifying a preferred system alternative by a combat simulation is investigated. When using this concept, partial battles are simulated to screen candidate alternatives. The Filter concept of simulation experiments is an application of the block experimental procedure used in experimentation with physical systems. Increased correlation among alternatives improving the efficiency of the method is achieved by the use of common random number streams for all alternatives within a block. An experimental model was developed to represent effects of blocked experiments on combat simulation results, and this model includes heterogeneous variances and correlated effects. Using this model, estimators for the variance of average system performance and for the variance of the difference between the average performances of two systems were derived. Moreover, give estimates of model parameters, an algorithm was developed for determining the least-cost test plan to estimate average system performance with an upper limit being placed on the variance of the average performance of an individual system. To test the validity of the Filter concept, four blocks of DYNCOM simulation results were analyzed. Results supported the validity of the Filter Model.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1974
- Accession Number
- ADB001641
Entities
People
- Charles Mccartney
- Gordon M. Clark
Organizations
- Ohio State University