Hybrid least squares method
A hybrid algorithm is developed which blends two different approximations to the Hessian, the Levenberg--Marquardt approximation and Davidon's Optimally Conditioned Quasi-Newton approximation, through adaptively chosen parameters. The aim is to study how to combine effectively two different models of the function which are deduced from the available information. A particular implementation is discussed. Also test results and comparisons against the Levenberg--Marquardt and Davidon's Quasi-Newton method, which correspond to limiting cases of the hybrid algorithm.
- Research Organization:
- Argonne National Lab., IL (USA)
- DOE Contract Number:
- W-31109-ENG-38
- OSTI ID:
- 7301826
- Report Number(s):
- ANL-AMD-TM-254(Rev.)
- Country of Publication:
- United States
- Language:
- English
Similar Records
Software performance on nonlinear least-squares problems
On collinear scaling algorithms that extend quasi-Newton methods
A hybrid differential evolution/Levenberg-Marquardt method for solving inverse transport problems
Technical Report
·
Thu Sep 01 00:00:00 EDT 1988
·
OSTI ID:6716106
On collinear scaling algorithms that extend quasi-Newton methods
Conference
·
Fri Dec 30 23:00:00 EST 1994
·
OSTI ID:35777
A hybrid differential evolution/Levenberg-Marquardt method for solving inverse transport problems
Conference
·
Thu Dec 31 23:00:00 EST 2009
·
OSTI ID:1024348