next up previous contents
Next: Command line procedures Up: Fit routines Previous: INTERFEROMETRYFIT   Contents


This menu contains variables which control the performance of the Marquardt-Levenberg (M-L) non-linear least squares fitting.

ONE One iteration only
FLOAT Allow $V^2$ calibration to float
OPT. Optimize step size for M-L.
LAMBDA Gradient-Brute force parameter
TOLERANCE Ratio of smallest over largest eigenvalue allowed
CONVERGENCE Convergence criterion as ratio of new over previous $\chi^2$

Choosing the OPT. option will cause a one-time delay before starting M-L, including the loading of the interferometric data a second time as the optimization of the step size overwrites the currently loaded visibilities. The step size data are written into a file Z_h_steps.xdr.

Christian Hummel 2017-06-29