#include <ml/optimization/lbfgs.hpp>

Public Member Functions
	lbfgs_solver (std::shared_ptr< first_order_opt_interface > _model)

void	setup (const DenseVector &init_point, const std::map< std::string, flexible_type > &opts, const std::shared_ptr< smooth_regularizer_interface > &reg=nullptr)

bool	next_iteration ()

const solver_status &	status () const

Detailed Description

Solve a first_order_optimization_iterface model with an LBFGS implementation.

The implementation is based on Algorithm 7.4 (pg 178) of [1].

This subroutine solves an unconstrained minimization problem using the limited memory BFGS method. The routine is especially effective on problems involving a large number of variables. In a typical iteration of this method an approximation Hk to the inverse of the Hessian is obtained by applying M BFGS updates to a diagonal matrix Hk0, using information from the previous M steps. The user specifies the number M, which determines the amount of storage required by the routine. The user may also provide the diagonal matrices Hk0 if not satisfied with the default choice. The algorithm is described in [2].

The user is required to calculate the function value and its gradient.

The steplength is determined at each iteration by means of the line search routine MCVSRCH, which is a slight modification of the routine CSRCH written by More' and Thuente.

References:

(1) Wright S.J and J. Nocedal. Numerical optimization. Vol. 2. New York: Springer, 1999.

(2) "On the limited memory BFGS method for large scale optimization", by D. Liu and J. Nocedal, Mathematical Programming B 45 (1989) 503-528.

Parameters

[in]	model	Model with first order optimization interface.
[in]	init_point	Starting point for the solver.
[in]	opts	Solver options.
[in]	reg	Shared ptr to an interface to a smooth regularizer.

Returns: stats Solver return stats.

Template Parameters

Vector Sparse or dense gradient representation.

Definition at line 93 of file lbfgs.hpp.

The documentation for this class was generated from the following file:

ml/optimization/lbfgs.hpp

Public Member Functions

Detailed Description