Implementation of an observer for real-world problems. More...

#include "coco.h"
#include "coco_internal.h"
#include "coco_utilities.c"

Data Structures
struct	observer_rw_data_t
	The real-world observer data type. More...

Enumerations
enum	observer_rw_log_e { LOG_NEVER, LOG_LOW_DIM, LOG_ALWAYS }
	Enum for denoting when the decision variables and constraints are logged. More...

Functions
static coco_problem_t *	logger_rw (coco_observer_t observer, coco_problem_t problem)

static void	logger_rw_free (void *logger)

static void	observer_rw (coco_observer_t observer, const char options, coco_option_keys_t **option_keys)
	Initializes the observer for real-world problems. More...

Detailed Description

Implementation of an observer for real-world problems.

Enumeration Type Documentation

Enum for denoting when the decision variables and constraints are logged.

static coco_problem_t* logger_rw	(	coco_observer_t *	observer,
		coco_problem_t *	problem
	)

static

static void logger_rw_free ( void * logger )

static

static

Initializes the observer for real-world problems.

Possible options:

"log_variables: STRING" determines whether the decision variables are to be logged. STRING can take on the values "none" (don't output decision variables), "low_dim"(output decision variables only for dimensions lower or equal to low_dim_vars) and "all" (output all decision variables). The default value is "all".
"log_constraints: STRING" determines whether the constraints are to be logged. STRING can take on the values "none" (don't output constraints), "low_dim"(output constraints only for dimensions lower or equal to low_dim_cons) and "all" (output all constraints). The default value is "all".
"low_dim_vars: VALUE" determines the value used to define "low_dim" for decision variables. The default value is 10.
"low_dim_cons: VALUE" determines the value used to define "low_dim" for constraints. The default value is 10.
"log_only_better: 0/1" determines whether all solutions are logged (0) or only the ones that are better than previous ones (1). This is applicable only for the single-objective problems, where the default value is 0. For multi-objective problems, all solutions are always logged.
"log_time: 0/1" determines whether the time needed to evaluate each solution is logged (0) or not (1). The default value is 0.