gym-examples 3.0.83__py3-none-any.whl → 3.0.85__py3-none-any.whl

gym_examples/__init__.py

@@ -6,4 +6,4 @@ register(
      max_episode_steps=50,
 )
 
-__version__ = "3.0.83"
+__version__ = "3.0.85"

gym_examples/envs/wsn_env.py

@@ -23,6 +23,8 @@ base_station_position = np.array([(upper_bound - lower_bound)/2, (upper_bound -
 initial_number_of_packets = 1  # initial number of packets to transmit
 latency_per_hop = 1  # latency per hop in seconds
 
+coef_network_throughput = 2.6  # coefficient for the network throughput reward
+coef_packet_delivery_ratio = 1.6  # coefficient for the packet delivery ratio reward
 
 # Define the final reward function using an attention mechanism
 class Attention(nn.Module):
@@ -151,8 +153,8 @@ class WSNRoutingEnv(gym.Env):
                 self.number_of_packets[action] += self.number_of_packets[i]
             self.number_of_packets[i] = 0 # Reset the number of packets of the sensor i
             # Calculate final reward
-            rewards[i] = self.compute_attention_rewards(rewards[i])
-            # rewards[i] = np.mean(rewards[i])
+            # rewards[i] = self.compute_attention_rewards(rewards[i])
+            rewards[i] = self.compute_weighted_sum_rewards(rewards[i])
         for i in range(self.n_sensors):
             if (self.remaining_energy[i] <= 0) or (self.number_of_packets[i] <= 0):
                 dones[i] = True
@@ -303,6 +305,7 @@ class WSNRoutingEnv(gym.Env):
 
         #-- rewards related to the performance metrics
         reward_latency = self.compute_reward_latency()
+        
         reward_network_throughput = self.compute_reward_network_throughput()
         reward_packet_delivery_ratio = self.compute_reward_packet_delivery_ratio()
 
@@ -367,11 +370,13 @@ class WSNRoutingEnv(gym.Env):
         return np.clip(normalized_throughput, 0, 1)
 
 
-    def compute_sum_rewards(self, rewards):
+    def compute_weighted_sum_rewards(self, rewards):
         '''
-        Compute the sum of the rewards
+        Compute the weigthed sum of the rewards
         '''
-        return np.sum(rewards)
+        weights = np.array([1, 1, 1, 1, 1, coef_network_throughput, coef_packet_delivery_ratio])
+        return np.dot(rewards, weights) / np.sum(weights)
+        # return np.sum(rewards)
     
     def compute_attention_rewards(self, rewards):
         '''

{gym_examples-3.0.83.dist-info → gym_examples-3.0.85.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: gym-examples
-Version: 3.0.83
+Version: 3.0.85
 Summary: A custom environment for multi-agent reinforcement learning focused on WSN routing.
 Home-page: https://github.com/gedji/CODES.git
 Author: Georges Djimefo

gym_examples-3.0.85.dist-info/RECORD

@@ -0,0 +1,7 @@
+gym_examples/__init__.py,sha256=ByDRt4r5RLWZlLnPaP5i24fqifb-NUcAWGiBrjcAuSw,193
+gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
+gym_examples/envs/wsn_env.py,sha256=aeX-q8c9wDPn2tZDoWPbmeuShAPWLZqMg0BsVOC54NU,21009
+gym_examples-3.0.85.dist-info/METADATA,sha256=umuJFO1kGw8fPQwwElY4EyGxHFmnMuaGas67j2L2qWs,411
+gym_examples-3.0.85.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
+gym_examples-3.0.85.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
+gym_examples-3.0.85.dist-info/RECORD,,

gym_examples-3.0.83.dist-info/RECORD

@@ -1,7 +0,0 @@
-gym_examples/__init__.py,sha256=QDtvnXb2Mx55f1nDFtydZEG4rBrtql9NsdptzX8iERo,193
-gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
-gym_examples/envs/wsn_env.py,sha256=kvQViM19DyTdlmkIAUs_WL2UWIk46EiEV5LrWPCekUA,20630
-gym_examples-3.0.83.dist-info/METADATA,sha256=YEoAiSUMbHaInr_ZMK0Nb6XLEByt9O9RfVefbAv1rJs,411
-gym_examples-3.0.83.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
-gym_examples-3.0.83.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
-gym_examples-3.0.83.dist-info/RECORD,,