EVALUATION OF SIMULATION FROM ITS ABILITY TO SIMULATE COMPONENTS (SUBSYSTEMS) OF OVERALL SYSTEM.

Abstract

The report discusses the ability of a random-nature simulation model to represent a system. Operation of the overall system cannot be observed. However, information is available about the operation of some subsystems. These subsystems are such that their operation determines the operation of the overall system. The simulation is considered satisfactory if and only if it has a specified high probability of simulating all of these subsystems to a sufficiently close approximation (during a simulation run). Suppose that, for each subsystem, a criterion has been developed for deciding when the approximation is sufficiently close. This depends on information about the subsystem and its operation. For the evaluation, the simulation model is run a sufficient number of times. A run is considered satisfactory if and only if all the subsystems are simulated sufficiently close. These binomial data are used to investigate whether the model is satisfactory. Homogeneous and independent runs are assumed ordinarily. However, results also can be obtained that apply to heterogeneous runs. These more general results are useful when the simulation is not strictly repetitive (for example, of a man-machine nature). (Author)

Document Details

Document Type
Technical Report
Publication Date
Jan 23, 1970
Accession Number
AD0701688

Entities

People

  • John E. Walsh

Organizations

  • Southern Methodist University

Tags

DTIC Thesaurus Topics

  • Binomials
  • Data Science
  • Information Science
  • Mathematics
  • Probability
  • Simulations
  • Statistics
  • Test And Evaluation

Readers

  • Computer Science.
  • Statistical inference.