Difference between revisions of "Lbfgs.m"

Revision as of 17:31, 27 August 2016

This function calculates the quasi-Newton update with the gradient history, giving the search direction. Based on the BFGS algorithm, this limited-memory BFGS (L-BFGS) algorithm stores only few vectors that implicitly represent the approximation of the BFGS algorithm (which stores a matrix equal to the number of optimisation variables)

Syntax

    direction=lbfgs(x_hist,df_hist,grad,N)

Description

This function is the implementation from section 4 of http://dx.doi.org/10.1090/S0025-5718-1980-0572855-7

Arguments

    x_hist         - vector array of waveform history, num_var x size_store

    df_hist        - vector array of gradient history, num_var x size_store

    grad           - vector of the current gradient, (num_vars x 1)

    N              - number of waveform/gradient vectors to store (default=20)

Returns

    direction      - the vector giving the BFGS approximation to the search direction

Notes

The L-BFGS algorithm is the default of fminnewton.m, and should be considered a good mix of computational efficiency and fast convergence.

See also

quasinewton.m, fminnewton.m, optim_tols.m

Revision 3374, authors: Ilya Kuprov David Goodwin