Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Optimal weighting function for the invariant-imbedding estimator

Technical Report ·
OSTI ID:6308945
One of the problems in using the invariant imbedding estimator is to find the optimal or near optimal initial conditions for the weighting functions. Computational experiences indicate that these initial conditions influence the convergence rate tremendously. This problem is further complicated by the fact that the number of weighting functions increases quadratically with the number of parameters or variables to be estimated. It is not a simple matter to estimate the initial conditions to be used for a large number of interconnected weighting functions. In this work, least squares criterion combined with various optimization schemes is used to obtain the optimal initial conditions. It is shown that the convergence rate can be improved tremendously. These improved convergence rates should be very useful for off-line estimations with a limited number of experimental data.
Research Organization:
University of Southern California, Los Angeles (USA). Dept. of Electrical Engineering
DOE Contract Number:
AT03-76ER70019
OSTI ID:
6308945
Report Number(s):
USC-113P-48; ON: DE81026837
Country of Publication:
United States
Language:
English