This rSQP iteration class provides a tablular output suitable for an 80 char wide console. More...

#include <MoochoPack_MoochoTrackerConsoleStd.hpp>

Inheritance diagram for MoochoPack::MoochoTrackerConsoleStd:

Public Member Functions
	MoochoTrackerConsoleStd (const ostream_ptr_t &o, const ostream_ptr_t &journal_out)
	Construct with an output stream (console presumably) More...

void	set_output_stream (const ostream_ptr_t &o)
	Set the output stream for console outputting. More...

const ostream_ptr_t &	get_output_stream () const
	Get the output stream for console outputting. More...

Public Member Functions inherited from IterationPack::AlgorithmTracker
virtual	~AlgorithmTracker ()

	AlgorithmTracker (const ostream_ptr_t &journal_out)
	Construct with an output stream for journal_out. More...

virtual void	output_final (const Algorithm &algo, EAlgoReturn algo_return) const
	Output information about a just completed algorithm. More...

virtual void	set_journal_out (const ostream_ptr_t &journal_out)
	Set a smart pointer to the journal file. More...

const ostream_ptr_t &	get_journal_out () const
	Get the smart pointer to the journal file. More...

virtual std::ostream &	journal_out () const
	Return a reference to a `std::ostream` to be used to output debug information and the like. More...

Protected Member Functions
void	print_top_header (const NLPAlgoState &s, const NLPAlgo &algo) const
	Print the top header to the output. More...

void	print_header (const NLPAlgoState &s, const NLPAlgo &algo) const
	Print the header to the output. More...

std::ostream &	o () const

Private Types
enum	{ NUM_PRINT_LINES = 10 }

Private Member Functions
	MoochoTrackerConsoleStd ()

Private Attributes
ostream_ptr_t	o_

StopWatchPack::stopwatch	timer_

int	printed_lines_

quasi_newton_stats_iq_member	quasi_newton_stats_

Static Private Attributes
static int	w_i2_ = 2

static char	ul_i2_ [] = "--"

static int	w_i4_ = 4

static char	ul_i4_ [] = "----"

static int	p2_ = 1

static int	w_p2_ = 8

static char	ul_p2_ [] = "--------"

static int	p3_ = 2

static int	w_p3_ = 9

static char	ul_p3_ [] = "---------"

Overridden from AlgorithmTracker
void	initialize ()
	Restarts the timer. More...

void	output_iteration (const Algorithm &algo) const

void	output_final (const Algorithm &algo, EAlgoReturn algo_return) const

Detailed Description

This rSQP iteration class provides a tablular output suitable for an 80 char wide console.

Specifically, these object produces a table with the following fields:

k : The SQP iteration counter starting at zero generally.
f : (iteration quantity f_k) The value of the objective function at the current iteration. This value may be scaled and the scaling factor will be printed out before the table is produced.
||c||s : (iteration quantity feas_kkt_err_k) The scaled value of the constraints norm (usually the infinity norm) at the current iteration. This is the value compared to the convergence criteria feas_tol (see the step class CheckConvergenceStd_AddedStep).
||rGL||s : (iteration quantity opt_kkt_err_k) The scaled value of the norm (usually the infinity norm) of the reduced gradient of the Lagrangian at the current iteration. This is the value compared to the convergence criteria opt_tol (see the step class CheckConvergenceStd_AddedStep).
QN : (iteration quantity quasi_newton_stats_k) Information about the quasi-Newton update of the reduced Hessian rHL_k.
- IN : rHL_k was reinitialized (identity?)
- UP : A standard quasi-Newton update was performed on rHL_km1 -> rHL_k
- DU : A dampened quasi-Newton update (BFGS) was performed on rHL_km1 -> rHL_k.
- SK : The quasi-Newton update was skipped because the current iterate was in the wrong region.
- IS : The quasi-Newton update (BFGS) was skipped because it was not positive definite or illdefined.
#act : (iteration quantity nu_k.nz()) The number of active variable bounds at the current iteration.
||Ypy||2 : (iteration quantity Ypy_k) The 2 norm of the Range space (feasibility) contribution to the full step d.
||Zpz||2 : (iteration quantity Zpz_k) The 2 norm of the Null space (optimality) contribution to the full step d.
||d||inf : (iteration quantity d_k) The infinity norm of the step vector for the primal unknowns x_kp1 = x_k + alpha_k * d_k.

The above quantities can tell you a lot about the progress of the SQP algorithm.

ToDo: Finish discussion.

After the algorithm is finished the total solution time will be printed as well as if the solution was found or not. Also, the number of function and gradient evaluations will be printed. Note that the timer is started from the moment this object is created or when set_output_stream(...) is called.

Definition at line 109 of file MoochoPack_MoochoTrackerConsoleStd.hpp.