Toward Validation of Computer Simulation Models in Operational Test and Evaluation.

Abstract

Validation of complex computer simulations is considered in the context of the operational test and evaluation of the air launched cruise missile. Published literature on validation is reviewed. Validation is described as a problem-dependent process. The goal of that process is an acceptable level of confidence that the actual and simulated data agree closely enough for an inference about the simulation to be a valid inference about the actual system. The cost of accepting a given level of confidence must be balanced against the cost of obtaining a higher level. Experimental design is seen as the key to obtaining the maximum amount of information from a limited number of test runs and is discussed. The following approach to validity is suggested. Use a screening design to identify the few, most-important variables in the simulation. Use a more complete design to specify parameter/level combinations for simulated and actual data comparison. Finally, explicitly incorporate decision analysis in judging the validity of the model based on that comparison and the use that will be made of the simulation results. A fractional factorial design is used to screen out the important factors from the clutter and multipath sub-model in TAC ZINGER. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 1979
Accession Number
ADA083516

Entities

People

  • James M. Arnett

Organizations

  • Air Force Institute of Technology

Tags

Communities of Interest

  • Advanced Electronics
  • Air Platforms
  • C4I
  • Ground and Sea Platforms
  • Space
  • Weapons Technologies

DTIC Thesaurus Topics

  • Air Force
  • Aircrafts
  • Computational Science
  • Computer Programs
  • Computers
  • Data Analysis
  • Data Mining
  • Data Science
  • Experimental Design
  • Information Science
  • Knowledge Management
  • Mathematical Models
  • Operations Research
  • Radar
  • Random Variables
  • Statistical Algorithms
  • Test And Evaluation

Readers

  • Computational Modeling and Simulation
  • Regression Analysis.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference