modify COMPARE_N in tool, modify attribute of vehicle by using tensor

deepQlearn_0/src/deepQlearning/learn_to_drive.h

@@ -17,15 +17,21 @@
 
 struct qlearning_params {
   double learning_rate;
+  double factor_update_learning_rate;
+  double minimum_threshold_learning_rate;
   double discount_factor;
   double exploration_factor;
+  double factor_update_exploration_factor;
+  double minimum_threshold_exploration_factor;
+  long int nb_training_before_update_weight_in_target;
 };
 
 
-struct reward_lists {
+struct status_qlearning {
   struct main_list_TYPE_L_INT * list_main_cumul;
   struct main_list_TYPE_L_INT * list_target_cumul;
   struct main_list_TYPE_L_INT * progress_best_cumul;
+  long int nb_training_after_updated_weight_in_target;
 };
 
 struct delay_params {
@@ -43,7 +49,7 @@ struct networks_qlearning {
 struct RL_agent {
   struct networks_qlearning * networks;
   struct vehicle * car;
-  struct reward_lists * rewards;
+  struct status_qlearning * status;
   struct delay_params * delay;
   struct qlearning_params *qlearnParams;
 
@@ -53,7 +59,7 @@ struct networks_qlearning * create_nework_qlearning(
   struct config_layers * config,
   bool randomize, float minR, float maxR,  int randomRange
 );
-struct reward_lists * create_reward_lists ();
+struct status_qlearning * create_status_qlearning ();
 struct delay_params * create_delay_params (
   size_t delay_between_episodes,
   size_t delay_between_games
@@ -62,19 +68,20 @@ struct delay_params * create_delay_params (
 struct qlearning_params * create_qlearning_params (
   double learning_rate,
   double discount_factor,
-  double exploration_factor
+  double exploration_factor,
+  long int nb_training_before_update_weight_in_target
 );
 
 struct RL_agent * create_RL_agent (
   struct networks_qlearning * networks,
   struct vehicle * car,
-  struct reward_lists * rewards,
+  struct status_qlearning * status,
   struct delay_params * delay,
   struct qlearning_params *qlearnParams
 );
 
 void free_networks_qlearning (struct networks_qlearning * networks);
-void free_reward_lists(struct reward_lists *rwd_l);
+void free_status_qlearning(struct status_qlearning *status_ql);
 void free_delay_params (struct delay_params *dly_p);
 void free_qlearning_params(struct qlearning_params *q_params);
 void free_RL_agent(struct RL_agent *rlAgent);
@@ -82,5 +89,10 @@ void free_RL_agent(struct RL_agent *rlAgent);
 void copy_weight_in_networks_from_main_to_target(struct networks_qlearning * networks);
 void copy_weight_in_networks_from_main_to_best(struct networks_qlearning * networks);
 
+void train_qlearning(struct RL_agent * rlAgent,
+  int action ,
+  tensor_TYPE_FLOAT * new_state /*input*/,
+  tensor_TYPE_FLOAT * state /*input*/,
+  long reward);
 
 #endif /* __LEARNING_VEHICLE__C_H____ */