From Spinach Documentation Wiki
Jump to: navigation, search

Performs a bracket section line search to find an appropriate optimisation step length in the specified direction. Based on fminlbfgs.m code from D. Kroon, University of Twente (Nov 2010). This function assumes cheap gradients.




   cost_function  - objective function handle

   d_0    - proposed step vector

   x_0    - current point

   fx_0   - function value at the current point

   gfx_0  - gradient at the current point

   data   - diagnostic data structure


   alpha     - optimum step length multiplier

   fx_1      - function value at the resulting point

   gfx_1     - gradient at the resulting point

   exitflag  - termination message

   data      - updated disgnostic data structure

See also

fminnewton.m, lbfgs.m, hessreg.m, optimcon.m

Version 2.2, authors: David Goodwin, Ilya Kuprov