I am using lsqcurvefit function from optim package in Octave for non-linear LSQ optimization. In documentation of function its written, only 'lm_svd_feasible' back-end algorithm is available. Which algorithm is it ?
That is described in the manual of the optim package. If you type optim_doc at the Octave prompt (with the optim package loaded), you get the optim manual where that algorithm is described. To search on the manual, type s or /. Note that Regexp search []: appears at the bottom of the window. Type lm_svd_feasible and then enter. The first result will be in the menu:
Optimization backends
* lm_svd_feasible:: L/M algorithm with SVD, constraints met
throughout optimization.
That is actually a link. Since your cursor will already be above lm_svd_feasible, press enter again which will take you to the section "Default backend lm_svd_feasible of residual minimization".
Octave and Octave Forge packages use the info format, so see the info manual for instructions. Alternatively, the manual is also online in HTML