- set time horizon in learning_role

assume/reinforcement_learning/learning_role.py

@@ -17,6 +17,7 @@
     get_schedule_fn,
     linear_schedule,
 )
+from assume.common.utils import datetime2timestamp
 
 logger = logging.getLogger(__name__)
 
@@ -124,6 +125,14 @@ def __init__(
         # list of avg_changes
         self.avg_rewards = []
 
+    def set_simulation_horizon_to_learning_role(self, start, end):
+        """
+        Add simulation horizon to learning role for calculation of decay.
+        """
+
+        self.start = datetime2timestamp(start)
+        self.end = datetime2timestamp(end)
+
     def load_inter_episodic_data(self, inter_episodic_data):
         """
         Load the inter-episodic data from the dict stored across simulation runs.

assume/world.py

@@ -262,7 +262,7 @@ def setup_learning(self) -> None:
             from assume.reinforcement_learning.learning_role import Learning
 
             self.learning_role = Learning(self.learning_config)
-            self._add_simulation_horizon_to_learning_role()
+            self.learning_role.set_simulation_horizon_to_learning_role(start=self.start, end=self.end)
             # separate process does not support buffer and learning
             self.learning_agent_addr = addr(self.addr, "learning_agent")
             rl_agent = agent_composed_of(
@@ -480,13 +480,7 @@ def create_unit(
             **unit_params,
         )
 
-    def _add_simulation_horizon_to_learning_role(self):
-        """
-        Add simulation horizon to learning role for calculation of decay.
-        """
 
-        self.learning_role.start = datetime2timestamp(self.start)
-        self.learning_role.end = datetime2timestamp(self.end)
 
     def _add_bidding_strategies_to_learning_role(self, unit_id, bidding_strategies):
         """