Accession Number:

ADA090420

Title:

Non-Linear Chi-Square Algorithm an Improvement on Non-Linear Least Squares,

Descriptive Note:

Corporate Author:

ARMY ARMAMENT RESEARCH AND DEVELOPMENT COMMAND DOVER NJ

Personal Author(s):

• Moore,Richard L.

Report Date:

1980-06-01

Pagination or Media Count:

15.0

Abstract:

Because of the high cost of testing, many large weapon systems cannot be tested over the full range of possible battlefield parameters. As a result, the developer and the reviewing authorities have come to rely on system simulation to demonstrate the system capability over the range of untested parameters. These simulations also are useful to investigate the change in performance resulting from possible subsystem modifications. In some important programs, the Government relies on simulations of competing systems to indicate the relative performance of these systems in situations for which no tests have been made, although of course, simulations such as these have been validated as much as possible by system tests. In these instances the procurement decision rests heavily on the validity of the system simulations. Consequently the need arises for a generally accepted procedure which is undoubtedly fair to each contractor and provides the maximum amount of objective judgment about the validity of the simulation. In any such procedure the Government must be able to rapidly evaluate simulations furnished from a variety of sources. Author

Descriptors:

• *TEST AND EVALUATION
• *SIMULATION
• *WEAPON SYSTEMS
• ALGORITHMS
• COSTS
• MATHEMATICAL ANALYSIS
• LEAST SQUARES METHOD
• DATA REDUCTION
• NONLINEAR ANALYSIS
• FIGURE OF MERIT
• CHI SQUARE TEST
• EXTRAPOLATION

Subject Categories:

• Statistics and Probability
• Operations Research
• Guns

Distribution Statement:

APPROVED FOR PUBLIC RELEASE