CCL header file for learning unconstraint policy. More...

#include <ccl_math.h>
#include <gsl/gsl_linalg.h>
#include <gsl/gsl_blas.h>
#include <gsl/gsl_math.h>
#include <gsl/gsl_eigen.h>
#include <gsl/gsl_sf.h>
#include <stdio.h>
#include <string.h>
#include <stdlib.h>

Data Structures
struct	LEARN_MODEL_PI_WS
	This structure describes workspace for directly learning the policy model parameters using linear regression. More...

struct	LEARN_MODEL_PI
	This structure describes learning a direct linear policy model parameters LEARN_MODEL_PI. More...

struct	LEARN_MODEL_LW_PI
	This structure describes learning a direct locally weighted linear policy model parameters LEARN_MODEL_LW_PI. More...

struct	LEARN_MODEL_LW_PI_WS
	This structure describes workspace for directly learning the policy model parameters using locally weighted linear regression. More...

Macros
#define	NUM_CENTRES 20

Functions
void	ccl_learn_policy_pi (LEARN_MODEL_PI model, const double BX, const double *Y)
	Main computation routine for learning linear policy. More...

int	ccl_learn_model_pi_ws_alloc (LEARN_MODEL_PI model, LEARN_MODEL_PI_WS ws)
	Allocates the workspace memory for directly learning policy pi. More...

int	ccl_learn_model_pi_ws_free (LEARN_MODEL_PI_WS *ws)
	Free the workspace memory for directly learning policy pi. More...

void	ccl_learn_policy_lw_pi (LEARN_MODEL_LW_PI model, const double WX, const double X, const double Y)
	Main computation routine for locally weighted linear policy. More...

int	ccl_learn_policy_lw_pi_model_alloc (LEARN_MODEL_LW_PI *model)
	Allocates the memory for locally weighted linear policy LEARN_MODEL_LW_PI. More...

int	ccl_learn_policy_lw_pi_model_free (LEARN_MODEL_LW_PI *model)
	Free the memory for locally weighted linear policy LEARN_MODEL_LW_PI. More...

int	ccl_learn_model_lw_pi_ws_alloc (LEARN_MODEL_LW_PI model, LEARN_MODEL_LW_PI_WS ws)
	Allocates the workspace memory for locally weighted linear policy LEARN_MODEL_LW_PI_WS. More...

int	ccl_learn_model_lw_pi_ws_free (LEARN_MODEL_LW_PI_WS *ws)
	Free the workspace memory for locally weighted linear policy LEARN_MODEL_LW_PI_WS. More...

void	predict_linear (const double X, const double centres, const double variance, const LEARN_MODEL_PI model, double Yp)
	Predictions based on linear policy model. More...

void	predict_local_linear (const double X, const double centres, const double variance, const LEARN_MODEL_LW_PI model, double Yp)
	Predictions based on locally weighted linear policy model. More...

int	ccl_read_data_from_file (char filename, int dim_x, int dim_n, double mat)
	Read data from .txt file. More...

int	ccl_write_lwmodel_to_file (char filename, LEARN_MODEL_LW_PI model)
	Write locally weighted model to .txt file. More...

Detailed Description

CCL header file for learning unconstraint policy.

Macro Definition Documentation

#define NUM_CENTRES 20

NUM_CENTRES is the number of rbf centers

Function Documentation

int ccl_learn_model_lw_pi_ws_alloc	(	LEARN_MODEL_LW_PI *	model,
		LEARN_MODEL_LW_PI_WS *	ws
	)

Allocates the workspace memory for locally weighted linear policy LEARN_MODEL_LW_PI_WS.

Parameters

[in]	model	Must point to a valid LEARN_MODEL_LW_PI structure
[in]	ws	Must point to a valid LEARN_MODEL_LW_PI_WS structure

Returns

1 in case of succes
0 in case of failure (e.g. memory could not be allocated).

int ccl_learn_model_lw_pi_ws_free ( LEARN_MODEL_LW_PI_WS * ws )

Free the workspace memory for locally weighted linear policy LEARN_MODEL_LW_PI_WS.

Parameters

[in] ws Must point to a valid LEARN_MODEL_LW_PI_WS structure

Returns

1 in case of succes
0 in case of failure (e.g. memory could not be freed).

int ccl_learn_model_pi_ws_alloc	(	LEARN_MODEL_PI *	model,
		LEARN_MODEL_PI_WS *	ws
	)

Allocates the workspace memory for directly learning policy pi.

Parameters

[in]	model	Must point to a valid LEARN_MODEL_PI structure
[in]	ws	Must point to a valid LEARN_MODEL_PI_WS structure

Returns

1 in case of succes
0 in case of failure (e.g. memory could not be allocated).

int ccl_learn_model_pi_ws_free ( LEARN_MODEL_PI_WS * ws )

Free the workspace memory for directly learning policy pi.

Parameters

[in] ws Must point to a valid LEARN_MODEL_PI_WS structure

Returns

1 in case of succes
0 in case of failure (e.g. memory could not be freed).

void ccl_learn_policy_lw_pi	(	LEARN_MODEL_LW_PI *	model,
		const double *	WX,
		const double *	X,
		const double *	Y
	)

Main computation routine for locally weighted linear policy.

Parameters

[in]	model	Must point to a valid LEARN_MODEL_LW_PI structure
[in]	WX	Locally weighted inputs, must point to an array of dim_b dim_n doubles*
[in]	X	Inputs variables, must point to an array of dim_x dim_n doubles*
[in]	Y	Output variables, must point to an array of dim_y dim_n doubles*
[out]	model	Must point to a valid LEARN_MODEL_LW_PI structure

int ccl_learn_policy_lw_pi_model_alloc ( LEARN_MODEL_LW_PI * model )

Allocates the memory for locally weighted linear policy LEARN_MODEL_LW_PI.

Parameters

[in] model Must point to a valid LEARN_MODEL_LW_PI structure

Returns

1 in case of succes
0 in case of failure (e.g. memory could not be allocated).

int ccl_learn_policy_lw_pi_model_free ( LEARN_MODEL_LW_PI * model )

Free the memory for locally weighted linear policy LEARN_MODEL_LW_PI.

Parameters

[in] model Must point to a valid LEARN_MODEL_LW_PI structure

Returns

1 in case of succes
0 in case of failure (e.g. memory could not be freed).

void ccl_learn_policy_pi	(	LEARN_MODEL_PI *	model,
		const double *	BX,
		const double *	Y
	)

Main computation routine for learning linear policy.

Parameters

[in]	model	Must point to a valid LEARN_MODEL_PI structure
[in]	BX	Higher dimensionality of input variable, must point to an array of dim_b dim_n doubles*
[in]	Y	Ouput variable, must point to an array of dim_y dim_n doubles.*
[in]	model	Must point to a valid LEARN_MODEL_PI structure

int ccl_read_data_from_file	(	char *	filename,
		int	dim_x,
		int	dim_n,
		double *	mat
	)

Read data from .txt file.

Parameters

[in]	filename	Name of file
[in]	dim_x	Dimensionality of input variable
[in]	dim_n	Number of data samples
[out]	mat	Returned data matrix

Returns

1 in case of succes
0 in case of failure (e.g. memory could not be freed).

int ccl_write_lwmodel_to_file	(	char *	filename,
		LEARN_MODEL_LW_PI *	model
	)

Write locally weighted model to .txt file.

Parameters

[in]	filename	Name of file
[in]	model	Must point to a valid LEARN_MODEL_LW_PI structure

Returns

1 in case of succes
0 in case of failure (e.g. memory could not be freed).

void predict_linear	(	const double *	X,
		const double *	centres,
		const double	variance,
		const LEARN_MODEL_PI *	model,
		double *	Yp
	)

Predictions based on linear policy model.

Parameters

[in]	X	Inputs variables, must point to an array of dim_x dim_n doubles*
[in]	centres	Center of rbf, must point to an array of 1 dim_b doubles*
[in]	variance	Variance of rbf
[in]	model	Must point to a valid LEARN_MODEL_PI structure
[out]	Yp	Predictions

void predict_local_linear	(	const double *	X,
		const double *	centres,
		const double	variance,
		const LEARN_MODEL_LW_PI *	model,
		double *	Yp
	)

Predictions based on locally weighted linear policy model.

Parameters

[in]	X	Inputs variables, must point to an array of dim_x dim_n doubles*
[in]	centres	Center of rbf, must point to an array of 1 dim_b doubles*
[in]	variance	Variance of rbf
[in]	model	Must point to a valid LEARN_MODEL_LW_PI structure
[out]	Yp	Predictions

Data Structures

Macros

Functions

Detailed Description

Macro Definition Documentation

Function Documentation