Productivity Estimates of the Strategic Airlift System by the Use of Simulation
Abstract
Although the strategic airlift system is under continuous analysis, C-5A problems provided impetus to analyze the airlift system productivity function by using a large-scale simulation model. Since limited verification and validation tests had been performed on the simulation model, the design of experiments was of critical importance. The experimental design had to be flexible enough to salvage the maximum amount of information possible upon the discovery of either a verification or validation error. In addition, the experimental design was required to accommodate the estimation of a large number of possibly changing independent variables. The experimental design developed for the analysis was full factorial design sets for a finite number of factors. Initial analysis began with aggregated sets of factors at two levels, and information gained from experiment execution was used to parse the sets. The process was sequential and parsing continued until the major explanatory independent variables were identified or enough information was obtained to eliminate the factor from further direct analysis. This design permitted the overlapping of simulation runs to fill out the factorial design sets.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1972
- Accession Number
- AD0756780
Entities
People
- Richard L. Nolan
- Rocci Mastroberti
Organizations
- Harvard University