A Technique for Simulating Mixed Normal and Uniform Distributions.
Abstract
Because of the multivariate nature of personnel problems, simulaton of manpower systems requires techniques for generating sets of variables with prespecified joint distribution functions. For example, Sorenson, while studying the impact of manpower allocation policies, has used the joint normal distribution to simulate the enlisted input. Techniques used to generate these scores have been described by Boldt, Wiskoff, and Fitch and by Hillier. For many problems, the normal approximation will be quite adequate. For certain problems, however, a procedure which allows the generating of both uniform and normal variables is desirable. For example, when the Armed Forces Qualification Test (AFQT) score is included with Army Classification Battery (ACB) scores, the uniform distribution is introduced since the AFQT, being a percentile measure, is uniformly distributed. Conceivably, other percentile measures could be introduced also. Hence the problem of generating mixed normal and uniform distributions can have application in personnel simulations. Procedures for generating such distributions are presented and their derivations outlined.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1965
- Accession Number
- ADA079247
Entities
People
- Robert F. Boldt