ann.h

Go to the documentation of this file.

/* $Header: /var/cvs/mbdyn/mbdyn/mbdyn-1.0/libraries/libann/ann.h,v 1.16 2017/01/12 14:43:23 masarati Exp $ */
/* 
 * MBDyn (C) is a multibody analysis code. 
 * http://www.mbdyn.org
 *
 * Copyright (C) 1996-2017
 *
 * Pierangelo Masarati  <masarati@aero.polimi.it>
 * Paolo Mantegazza     <mantegazza@aero.polimi.it>
 *
 * Dipartimento di Ingegneria Aerospaziale - Politecnico di Milano
 * via La Masa, 34 - 20156 Milano, Italy
 * http://www.aero.polimi.it
 *
 * Changing this copyright notice is forbidden.
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation (version 2 of the License).
 * 
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */
/*
 * Copyright (C) 2008
 *
 * Mattia Mattaboni     <mattaboni@aero.polimi.it>
 */

#ifndef ANN_H
#define ANN_H

#ifdef __cplusplus
extern "C" {
#endif /* __cplusplus */

#include "ActivationFunction.h"
#include "matrix.h"

#define ANN_W_A_NONE            (0x0U)
#define ANN_W_A_TEXT            (0x1U)
#define ANN_W_A_BIN             (0x2U)

#define ANN_FEEDBACK_NONE       (0x0U)
#define ANN_FEEDBACK_UPDATE     (0x1U)

/* diagnostics */
typedef enum { 
        ANN_OK = 0,
        ANN_NO_MEMORY,
        ANN_NO_FILE,
        ANN_DATA_ERROR,
        ANN_MATRIX_ERROR,
        ANN_GEN_ERROR
} ann_res_t;

typedef enum {
        ANN_TM_BATCH,
        ANN_TM_SEQUENTIAL
} ann_training_mode_t;

typedef matrix*  ANN_vector_matrix;
typedef vector*  ANN_vector_vector;


/* Artificial Neural Network structure*/
typedef struct ANN {
        /* NEURON'S ACTIVATION FUNCTION */
        w_init_f                w_init;
        w_destroy_f             w_destroy;
        w_read_f                w_read;
        w_write_f               w_write;
        w_eval_f                w_eval;
        void                    *w_priv;

        /* NETWORK ARCHITECTURE */
        int                     N_input;        /* network input number*/
        int                     N_output;       /* network (visible) output number*/
        int                     N_layer;        /* network layer number*/
        int                     *N_neuron;      /* neuron number*/
        int                     r;              /* number of time delay*/

        /* TRAINING PARAMETERS */
        double                  eta;            /* learning rate*/
        double                  rho;            /* momentum term*/

        /* SYNAPTIC WEIGHTS */
        ANN_vector_matrix       W;              /* network synaptic weights*/

        /* JCOBIAN MATRIX */
        matrix                  jacobian;

        /* SCALE FACTORS */
        matrix                  input_scale;    /* input scale factors */
        matrix                  output_scale;   /* output scale factors */

        /* SIMULATION DATA */
        ANN_vector_vector       v;              /* neurons' internal activity*/
        ANN_vector_vector       Y_neuron;       /* neurons' output*/
        vector                  yD;             /* output */

        /* TRAINING DATA */
        ANN_vector_matrix       dEdW;           /* error gradient*/
        ANN_vector_matrix       **dy;           /* output network gradient*/

        /* PRIVATE TRAINING DATA */
        ANN_vector_matrix       *dydW, dW;
        ANN_vector_vector       dXdW, temp, dydV, dEdV, dXdu;
        vector                  input, output, error;
        
          
} ANN;


/* FUNCTIONS' PROTOTYPES  */
ann_res_t ANN_init( ANN *, const char * );
ann_res_t ANN_destroy( ANN * );
ann_res_t ANN_write( ANN *, FILE *, unsigned );
ann_res_t ANN_sim( ANN *, vector *, vector *, unsigned );
ann_res_t ANN_DataRead( matrix *, int *, char * );
ann_res_t ANN_DataWrite( matrix *, char * );
double ANN_InternalFunction( double , ANN * );
double ANN_InternalFunctionDer( double , ANN * );
ann_res_t ANN_vector_matrix_init( ANN_vector_matrix *, int *, int );
ann_res_t ANN_vector_vector_init( ANN_vector_vector *, int *, int );
ann_res_t ANN_dEdW( ANN *, vector *);
ann_res_t ANN_dXdW( ANN *, int , int , int );
ann_res_t ANN_WeightUpdate( ANN *, ANN_vector_matrix, double );
ann_res_t ANN_TrainingEpoch( ANN *, matrix *, matrix *, matrix *, int, ann_training_mode_t);
ann_res_t ANN_reset( ANN * );
ann_res_t ANN_TotalError( matrix *, matrix *, double *);
ann_res_t ANN_vector_matrix_ass( ANN_vector_matrix *, ANN_vector_matrix *, int* , int, double );
ann_res_t ANN_jacobian_matrix( ANN *, matrix * );

void ANN_error( ann_res_t, char * );

#ifdef __cplusplus
}
#endif /* __cplusplus */

#endif /* ANN_H */