qpa_control_type structure

#include <galahad_qpa.h>

struct qpa_control_type {
    // components

    bool f_indexing;
    ipc_ error;
    ipc_ out;
    ipc_ print_level;
    ipc_ start_print;
    ipc_ stop_print;
    ipc_ maxit;
    ipc_ factor;
    ipc_ max_col;
    ipc_ max_sc;
    ipc_ indmin;
    ipc_ valmin;
    ipc_ itref_max;
    ipc_ infeas_check_interval;
    ipc_ cg_maxit;
    ipc_ precon;
    ipc_ nsemib;
    ipc_ full_max_fill;
    ipc_ deletion_strategy;
    ipc_ restore_problem;
    ipc_ monitor_residuals;
    ipc_ cold_start;
    ipc_ sif_file_device;
    rpc_ infinity;
    rpc_ feas_tol;
    rpc_ obj_unbounded;
    rpc_ increase_rho_g_factor;
    rpc_ infeas_g_improved_by_factor;
    rpc_ increase_rho_b_factor;
    rpc_ infeas_b_improved_by_factor;
    rpc_ pivot_tol;
    rpc_ pivot_tol_for_dependencies;
    rpc_ zero_pivot;
    rpc_ inner_stop_relative;
    rpc_ inner_stop_absolute;
    rpc_ multiplier_tol;
    rpc_ cpu_time_limit;
    rpc_ clock_time_limit;
    bool treat_zero_bounds_as_general;
    bool solve_qp;
    bool solve_within_bounds;
    bool randomize;
    bool array_syntax_worse_than_do_loop;
    bool space_critical;
    bool deallocate_error_fatal;
    bool generate_sif_file;
    char symmetric_linear_solver[31];
    char sif_file_name[31];
    char prefix[31];
    bool each_interval;
    struct sls_control_type sls_control;
};

detailed documentation

control derived type as a C struct

components

bool f_indexing

use C or Fortran sparse matrix indexing

ipc_ error

error and warning diagnostics occur on stream error

ipc_ out

general output occurs on stream out

ipc_ print_level

the level of output required is specified by print_level

ipc_ start_print

any printing will start on this iteration

ipc_ stop_print

any printing will stop on this iteration

ipc_ maxit

at most maxit inner iterations are allowed

ipc_ factor

the factorization to be used. Possible values are 0 automatic 1 Schur-complement factorization 2 augmented-system factorization

ipc_ max_col

the maximum number of nonzeros in a column of A which is permitted with the Schur-complement factorization

ipc_ max_sc

the maximum permitted size of the Schur complement before a refactorization is performed

ipc_ indmin

an initial guess as to the integer workspace required by SLS (OBSOLETE)

ipc_ valmin

an initial guess as to the real workspace required by SLS (OBSOLETE)

ipc_ itref_max

the maximum number of iterative refinements allowed (OBSOLETE)

ipc_ infeas_check_interval

the infeasibility will be checked for improvement every infeas_check_interval iterations (see infeas_g_improved_by_factor and infeas_b_improved_by_factor below)

ipc_ cg_maxit

the maximum number of CG iterations allowed. If cg_maxit < 0, this number will be reset to the dimension of the system + 1

ipc_ precon

the preconditioner to be used for the CG is defined by precon. Possible values are 0 automatic 1 no preconditioner, i.e, the identity within full factorization 2 full factorization 3 band within full factorization 4 diagonal using the barrier terms within full factorization

ipc_ nsemib

the semi-bandwidth of a band preconditioner, if appropriate

ipc_ full_max_fill

if the ratio of the number of nonzeros in the factors of the reference matrix to the number of nonzeros in the matrix itself exceeds full_max_fill, and the preconditioner is being selected automatically (precon = 0), a banded approximation will be used instead

ipc_ deletion_strategy

the constraint deletion strategy to be used. Possible values are:

0 most violated of all 1 LIFO (last in, first out) k LIFO(k) most violated of the last k in LIFO

ipc_ restore_problem

indicate whether and how much of the input problem should be restored on output. Possible values are 0 nothing restored 1 scalar and vector parameters 2 all parameters

ipc_ monitor_residuals

the frequency at which residuals will be monitored

ipc_ cold_start

indicates whether a cold or warm start should be made. Possible values are

0 warm start - the values set in C_stat and B_stat indicate which constraints will be included in the initial working set. 1 cold start from the value set in X; constraints active at X will determine the initial working set. 2 cold start with no active constraints 3 cold start with only equality constraints active 4 cold start with as many active constraints as possible

ipc_ sif_file_device

specifies the unit number to write generated SIF file describing the current problem

rpc_ infinity

any bound larger than infinity in modulus will be regarded as infinite

rpc_ feas_tol

any constraint violated by less than feas_tol will be considered to be satisfied

rpc_ obj_unbounded

if the objective function value is smaller than obj_unbounded, it will be flagged as unbounded from below.

rpc_ increase_rho_g_factor

if the problem is currently infeasible and solve_qp (see below) is .TRUE. the current penalty parameter for the general constraints will be increased by increase_rho_g_factor when needed

rpc_ infeas_g_improved_by_factor

if the infeasibility of the general constraints has not dropped by a fac of infeas_g_improved_by_factor over the previous infeas_check_interval iterations, the current corresponding penalty parameter will be increase

rpc_ increase_rho_b_factor

if the problem is currently infeasible and solve_qp or solve_within_boun (see below) is .TRUE., the current penalty parameter for the simple boun constraints will be increased by increase_rho_b_factor when needed

rpc_ infeas_b_improved_by_factor

if the infeasibility of the simple bounds has not dropped by a factor of infeas_b_improved_by_factor over the previous infeas_check_interval iterations, the current corresponding penalty parameter will be increase

rpc_ pivot_tol

the threshold pivot used by the matrix factorization. See the documentation for SLS for details (OBSOLE

rpc_ pivot_tol_for_dependencies

the threshold pivot used by the matrix factorization when attempting to detect linearly dependent constraints.

rpc_ zero_pivot

any pivots smaller than zero_pivot in absolute value will be regarded to zero when attempting to detect linearly dependent constraints (OBSOLE

rpc_ inner_stop_relative

the search direction is considered as an acceptable approximation to the minimizer of the model if the gradient of the model in the preconditioning(inverse) norm is less than max( inner_stop_relative * initial preconditioning(inverse) gradient norm, inner_stop_absolute )

rpc_ inner_stop_absolute

see inner_stop_relative

rpc_ multiplier_tol

any dual variable or Lagrange multiplier which is less than multiplier_t outside its optimal interval will be regarded as being acceptable when checking for optimality

rpc_ cpu_time_limit

the maximum CPU time allowed (-ve means infinite)

rpc_ clock_time_limit

the maximum elapsed clock time allowed (-ve means infinite)

bool treat_zero_bounds_as_general

any problem bound with the value zero will be treated as if it were a general value if true

bool solve_qp

if solve_qp is .TRUE., the value of prob.rho_g and prob.rho_b will be increased as many times as are needed to ensure that the output solution is feasible, and thus aims to solve the quadratic program (2)-(4)

bool solve_within_bounds

if solve_within_bounds is .TRUE., the value of prob.rho_b will be increased as many times as are needed to ensure that the output solution is feasible with respect to the simple bounds, and thus aims to solve the bound-constrained quadratic program (4)-(5)

bool randomize

if randomize is .TRUE., the constraint bounds will be perturbed by small random quantities during the first stage of the solution process. Any randomization will ultimately be removed. Randomization helps when solving degenerate problems

bool array_syntax_worse_than_do_loop

if .array_syntax_worse_than_do_loop is true, f77-style do loops will be used rather than f90-style array syntax for vector operations (OBSOLETE)

bool space_critical

if .space_critical true, every effort will be made to use as little space as possible. This may result in longer computation time

bool deallocate_error_fatal

if .deallocate_error_fatal is true, any array/pointer deallocation error will terminate execution. Otherwise, computation will continue

bool generate_sif_file

if .generate_sif_file is .true. if a SIF file describing the current problem is to be generated

char symmetric_linear_solver[31]

the name of the symmetric-indefinite linear equation solver used. Possible choices are currently: ‘sils’, ‘ma27’, ‘ma57’, ‘ma77’, ‘ma86’, ‘ma97’, ‘ssids’, ‘mumps’, ‘pardiso’, ‘mkl_pardiso’, ‘pastix’, ‘wsmp’, and ‘sytr’, although only ‘sytr’ and, for OMP 4.0-compliant compilers, ‘ssids’ are installed by default; others are easily installed (see README.external). More details of the capabilities of each solver are provided in the documentation for galahad_sls.

char sif_file_name[31]

definite linear equation solver

name of generated SIF file containing input problem

char prefix[31]

all output lines will be prefixed by .prefix(2:LEN(TRIM(.prefix))-1) where .prefix contains the required string enclosed in quotes, e.g. “string” or ‘string’

bool each_interval

component specifically for parametric problems (not used at present)

struct sls_control_type sls_control

control parameters for SLS