saberbud/ceres-solver
1
0
Fork 0
mirror of https://github.com/ceres-solver/ceres-solver.git synced 2026-08-30 00:50:37 +08:00
Code Issues Packages Projects Releases Wiki Activity

Expose line search parameters in Solver::Options.

Browse Source 
Change-Id: Ifc52980976e7bac73c8164d80518a5a19db1b79d
This commit is contained in:
Sameer Agarwal
2013-06-30 14:33:23 -07:00
parent 1c70ae9aa6
commit 09244015e3
8 changed files with 153 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files
include/ceres/solver.h
+44
View File
@@ -63,6 +63,12 @@ class Solver {
line_search_type = ARMIJO;
nonlinear_conjugate_gradient_type = FLETCHER_REEVES;
max_lbfgs_rank = 20;
line_search_interpolation_type = CUBIC;
min_line_search_step_size = 1e-9;
armijo_sufficient_decrease = 1e-4;
min_armijo_relative_step_size_change = 1e-3;
max_armijo_relative_step_size_change = 0.6;
trust_region_strategy_type = LEVENBERG_MARQUARDT;
dogleg_type = TRADITIONAL_DOGLEG;
use_nonmonotonic_steps = false;
@@ -172,6 +178,43 @@ class Solver {
// Limited Storage". Mathematics of Computation 35 (151): 773–782.
int max_lbfgs_rank;
// Degree of the polynomial used to approximate the objective
// function. Valid values are BISECTION, QUADRATIC and CUBIC.
//
// BISECTION corresponds to pure backtracking search with no
// interpolation.
LineSearchInterpolationType line_search_interpolation_type;
// If during the line search, the step_size falls below this
// value, it is truncated to zero.
double min_line_search_step_size;
// Armijo line search parameters.
// Solving the line search problem exactly is computationally
// prohibitive. Fortunately, line search based optimization
// algorithms can still guarantee convergence if instead of an
// exact solution, the line search algorithm returns a solution
// which decreases the value of the objective function
// sufficiently. More precisely, we are looking for a step_size
// s.t.
//
// f(step_size) <= f(0) + sufficient_decrease * f'(0) * step_size
//
double armijo_sufficient_decrease;
// In each iteration of the Armijo line search,
//
// new_step_size >= min_relative_step_size_change * step_size
//
double min_armijo_relative_step_size_change;
// In each iteration of the Armijo line search,
//
// new_step_size <= max_relative_step_size_change * step_size
//
double max_armijo_relative_step_size_change;
TrustRegionStrategyType trust_region_strategy_type;
// Type of dogleg strategy to use.
@@ -673,6 +716,7 @@ class Solver {
LineSearchDirectionType line_search_direction_type;
LineSearchType line_search_type;
int max_lbfgs_rank;
};