Determining the quality of a model is always troublesome for the model builder and/or user. Error analysis is a useful approach when modeling dynamic systems and when measurements or nominal trajectories are available from the physical system. Techniques of error analysis based upon measures of the error in both the time and frequency domains are discussed. Further, an approach to analyzing the propagation of errors based upon dynamic models of the error is presented. Time domain measures and error propagation analysis are demonstrated on a sixth-order river basin econometric model.
Bellman, R.E. ; Kalaba, R.E.; and Wing, G.M. "Invariant Imbedding and the Reduction of Two-Point Boundary Value Problems to Initial-Value Problems ." Proceedings U.S. National Academy of Sciences46 (1960), 1646-1649.
2.
Box, G.E.P. and Jenkins, G.M.Time Series Analysis: Forecasting and Control (revised edition). Holden-Day, San Francisco, Calif. ( 1976).
3.
Damborg, M.J. and Fuller, L.F. "Model Validation Using Time and Frequency Domain Error Measures." Report no. 76-152 Office of Conservation, ERDA (September 1976).
Eykhoff, P.System Identification. John Wiley, New York, N.Y. (1974).
6.
Kheir, N.A. "On The Credibility of Models." Proceedings, 1981 Summer Computer Simulation Conference. Washington, D.C. (July 1981), 545-546.
7.
SCS Technical Committee on Model Credibility "Terminology for model credibility." Simulation32:3 (March 1979), 103-104.
8.
Smith, N.J. and Sage, A.P. "Hierarchical System Identification of Models for Urban Dynamics." Socioeconomic Planning Sciences7 (October 1973), 545-569.
9.
Tawari, P.V. "A Systematic Procedure for Validating and Evaluating Dynamic Models ." PhD dissertation. University of Washington (November 1974).
