Ipopt Documentation  
 
Loading...
Searching...
No Matches
IpBacktrackingLineSearch.hpp
Go to the documentation of this file.
1// Copyright (C) 2004, 2009 International Business Machines and others.
2// All Rights Reserved.
3// This code is published under the Eclipse Public License.
4//
5// Authors: Carl Laird, Andreas Waechter IBM 2004-08-13
6// Andreas Waechter IBM 2005-10-13
7// derived file from IpFilterLineSearch.hpp
8
9#ifndef __IPBACKTRACKINGLINESEARCH_HPP__
10#define __IPBACKTRACKINGLINESEARCH_HPP__
11
12#include "IpLineSearch.hpp"
14#include "IpRestoPhase.hpp"
15#include "IpConvCheck.hpp"
16
17namespace Ipopt
18{
19
36{
37public:
40
52 const SmartPtr<RestorationPhase>& resto_phase,
53 const SmartPtr<ConvergenceCheck>& conv_check
54 );
55
59
61 virtual bool InitializeImpl(
62 const OptionsList& options,
63 const std::string& prefix
64 );
65
71
79 virtual void Reset();
80
95 bool rigorous
96 )
97 {
98 rigorous_ = rigorous;
99 }
100
105 {
107 }
108
114
121
124 static void RegisterOptions(
126 );
128
129private:
140
143 );
144
148 );
150
159 bool skip_first_trial_point,
160 Number& alpha_primal,
161 bool& corr_taken,
162 bool& soc_taken,
163 Index& n_steps,
164 bool& evaluation_error,
165 SmartPtr<IteratesVector>& actual_delta
166 );
167
173
179 SmartPtr<IteratesVector>& actual_delta
180 );
181
189 Number alpha_primal
190 );
191
201 Number alpha_primal,
202 Number alpha_dual,
204 );
205
219 SmartPtr<IteratesVector>& actual_delta,
220 bool& satisfies_original_criterion
221 );
222
237 Number alpha_primal_test,
238 Number& alpha_primal,
239 SmartPtr<IteratesVector>& actual_delta
240 );
241
250 Number alpha_primal_test,
251 Number& alpha_primal,
252 SmartPtr<IteratesVector>& actual_delta
253 );
254
262
270
273
279
288
294
297
319
324
336
372
375
384
389
394
398
401
416
423
428
434
439
446
449
454
457
464};
465
466} // namespace Ipopt
467
468#endif
General implementation of a backtracking line search.
SmartPtr< const IteratesVector > watchdog_delta_
Watchdog search direction at reference point.
Index watchdog_shortened_iter_trigger_
Number of shortened iterations that trigger the watchdog.
Number tiny_step_y_tol_
Tolerance for y variables for the tiny step stopping heuristic.
bool DoBacktrackingLineSearch(bool skip_first_trial_point, Number &alpha_primal, bool &corr_taken, bool &soc_taken, Index &n_steps, bool &evaluation_error, SmartPtr< IteratesVector > &actual_delta)
Method performing the backtracking line search.
void operator=(const BacktrackingLineSearch &)
Default Assignment Operator.
Number tiny_step_tol_
Tolerance for detecting tiny steps.
Number alpha_red_factor_
factor by which search direction is to be shortened if trial point is rejected.
bool CheckAcceptabilityOfTrialPoint(Number alpha_primal)
Method for checking if current trial point is acceptable.
Number last_mu_
Barrier parameter value during last line search.
void StoreAcceptablePoint()
Store current iterate as acceptable point.
virtual void FindAcceptableTrialPoint()
Perform the line search.
Number constr_viol_tol_
unscaled constraint violation tolerance
bool skipped_line_search_
Flag indicating whether no acceptable trial point was found during last line search.
bool in_soft_resto_phase_
Flag indicating whether we are currently in the "soft" restoration phase mode, in which steps are acc...
Index watchdog_trial_iter_max_
Number of watch dog trial steps.
bool start_with_resto_
Indicates whether the algorithm should start directly with the restoration phase.
SmartPtr< ConvergenceCheck > conv_check_
bool TrySoftRestoStep(SmartPtr< IteratesVector > &actual_delta, bool &satisfies_original_criterion)
Try a step for the soft restoration phase and check if it is acceptable.
Index accept_after_max_steps_
Maximal number of trial steps before we blindly accept trial point.
bool DetectTinyStep()
Detect if the search direction is too small.
bool TrySecondOrderCorrection(Number alpha_primal_test, Number &alpha_primal, SmartPtr< IteratesVector > &actual_delta)
Try a second order correction for the constraints.
virtual void SetRigorousLineSearch(bool rigorous)
Set flag indicating whether a very rigorous line search should be performed.
BacktrackingLineSearch(const SmartPtr< BacktrackingLSAcceptor > &acceptor, const SmartPtr< RestorationPhase > &resto_phase, const SmartPtr< ConvergenceCheck > &conv_check)
Constructor.
Number alpha_for_y_tol_
Tolerance for primal step to switch to full equality constraint multiplier steps.
AlphaForYEnum alpha_for_y_
Flag indicating whether the dual step size is to be used for the equality constraint multipliers.
Number soft_resto_pderror_reduction_factor_
Reduction factor for the restoration phase that accepts steps reducing the optimality error ("soft re...
Index count_successive_shortened_steps_
Counter for the number of successive iterations in which the full step was not accepted.
bool expect_infeasible_problem_
Indicates whether problem can be expected to be infeasible.
bool accept_every_trial_step_
Flag indicating whether the line search should always accept the full (fraction-to-the-boundary) step...
void StopWatchDog()
Stop watch dog if started and restore iterate from before watchdog started.
void PerformDualStep(Number alpha_primal, Number alpha_dual, SmartPtr< IteratesVector > &delta)
Method for setting the dual variables in the trial fields in IpData, given the search direction.
bool TryCorrector(Number alpha_primal_test, Number &alpha_primal, SmartPtr< IteratesVector > &actual_delta)
Try higher order corrector (for fast local convergence).
SmartPtr< const IteratesVector > acceptable_iterate_
void StopWatchDog(SmartPtr< IteratesVector > &actual_delta)
Method for stopping the watch dog.
Number watchdog_alpha_primal_test_
Step size for Armijo test in watch dog.
virtual ~BacktrackingLineSearch()
Destructor.
bool rigorous_
Flag indicating whether the line search is to be performed robust (usually this is true,...
SmartPtr< BacktrackingLSAcceptor > acceptor_
static void RegisterOptions(SmartPtr< RegisteredOptions > roptions)
Methods for OptionsList.
BacktrackingLineSearch(const BacktrackingLineSearch &)
Copy Constructor.
bool tiny_step_last_iteration_
Flag indicating if a tiny step was detected in previous iteration.
bool RestoreAcceptablePoint()
Restore acceptable point into the current fields of IpData if found.
bool fallback_activated_
Flag indicating whether the algorithm has asked to immediately switch to the fallback mechanism (rest...
void PerformMagicStep()
Perform magic steps.
Number expect_infeasible_problem_ctol_
Tolerance on constraint violation for expect_infeasible_problem heuristic.
bool in_watchdog_
Flag indicating if the watchdog is active.
virtual void Reset()
Reset the line search.
bool magic_steps_
Flag indicating whether magic steps should be used.
virtual bool CheckSkippedLineSearch()
Check if the line search procedure didn't accept a new iterate during the last call of FindAcceptable...
SmartPtr< const IteratesVector > watchdog_iterate_
Watchdog reference iterate.
SmartPtr< RestorationPhase > resto_phase_
void StartWatchDog()
Method for starting the watch dog.
Index watchdog_shortened_iter_
Counter for shortened iterations.
virtual bool ActivateFallbackMechanism()
Activate fallback mechanism.
Index max_soft_resto_iters_
Maximal number of iterations that can be done in the soft iteration phase before the algorithm revert...
Number expect_infeasible_problem_ytol_
Trigger tolerance on constraint multipliers.
Index soft_resto_counter_
Counter for iteration performed in soft restoration phase in a row.
virtual bool InitializeImpl(const OptionsList &options, const std::string &prefix)
InitializeImpl - overloaded from AlgorithmStrategyObject.
AlphaForYEnum
enumeration for the different alpha_for_y_ settings
Index watchdog_trial_iter_
Counter for watch dog iterations.
bool CurrentIsAcceptable()
Method for determining if the current iterate is acceptable (in the sense of the acceptable_tol optio...
Base class for line search objects.
This class stores a list of user set options.
Template class for Smart Pointers.
This file contains a base class for all exceptions and a set of macros to help with exceptions.
ipindex Index
Type of all indices of vectors, matrices etc.
Definition IpTypes.hpp:20
ipnumber Number
Type of all numbers.
Definition IpTypes.hpp:17