Next: , Previous: , Up: Residual optimization   [Index]


2.3 Default backend lm_svd_feasible of residual minimization

Levenberg/Marquardt algorithm using singular value decomposition. Constraints must be met by the initial parameters and are attempted to be kept met throughout the optimization.

Returned value cvg will be 0, 1, or 2. Returned structure outp will have the fields niter and user_interaction.

Backend-specific defaults are: MaxIter: 20, fract_prec: zeros (size (parameters)), max_fract_change: Inf for all parameters. The setting TolX is not honoured.

Interpretation of Display: if set to "iter", currently some diagnostics are printed.

Specific option: lm_svd_feasible_alt_s: if falling back to nearly gradient descent, do it more like original Levenberg/Marquardt method, with descent in each gradient component; for testing only.