gym-examples 2.0.50__py3-none-any.whl → 2.0.52__py3-none-any.whl

gym_examples/__init__.py

@@ -6,4 +6,4 @@ register(
      max_episode_steps=50,
 )
 
-__version__ = "2.0.50"
+__version__ = "2.0.52"

gym_examples/envs/wsn_env.py

@@ -115,10 +115,10 @@ class WSNRoutingEnv(gym.Env):
             # Calculate final reward
             # rewards_individual = torch.tensor(rewards[i], dtype=torch.double)                
             # final_reward = net(rewards_individual)
-            final_reward = np.mean(rewards[i]) 
+            final_reward = np.sum(rewards[i]) 
             rewards[i] = final_reward
 
-        rewards = [r * self.network_reward_consumption_energy() * self.network_reward_dispersion_remaining_energy() for r in rewards]
+        rewards = [0.5 * r + 0.5 * (1/self.n_sensors) * (self.network_reward_consumption_energy() + self.network_reward_dispersion_remaining_energy()) for r in rewards]
 
         for i in range(self.n_sensors):
             if (self.remaining_energy[i] <= 0) or (self.number_of_packets[i] <= 0):
@@ -185,7 +185,7 @@ class WSNRoutingEnv(gym.Env):
         # Normalize the angle
         normalized_angle = abs(angle) / np.pi
 
-        return 1 - normalized_angle
+        return np.clip(1 - normalized_angle, 0, 1)
     
     def compute_reward_distance(self, i, action):
         '''
@@ -198,7 +198,7 @@ class WSNRoutingEnv(gym.Env):
         # Normalize the distance to the next hop
         normalized_distance_to_next_hop = distance / self.coverage_radius
 
-        return 1 - normalized_distance_to_next_hop
+        return np.clip(1 - normalized_distance_to_next_hop, 0, 1)
 
     def compute_reward_consumption_energy(self, i, action):
         '''
@@ -219,7 +219,7 @@ class WSNRoutingEnv(gym.Env):
         max_total_energy = max_transmission_energy + max_reception_energy
         normalized_total_energy = total_energy / max_total_energy
 
-        return 1 - normalized_total_energy
+        return np.clip(1 - normalized_total_energy, 0, 1)
     
     def compute_reward_dispersion_remaining_energy(self):
         '''
@@ -242,7 +242,7 @@ class WSNRoutingEnv(gym.Env):
         else: 
             normalized_number_of_packets = self.number_of_packets[action] / max_number_of_packets
 
-        return 1 - normalized_number_of_packets
+        return np.clip(1 - normalized_number_of_packets, 0, 1)
 
     def compute_individual_rewards(self, i, action):
         '''
@@ -251,10 +251,11 @@ class WSNRoutingEnv(gym.Env):
         reward_angle = self.compute_reward_angle(i, action)
         reward_distance = self.compute_reward_distance(i, action)
         reward_consumption_energy = self.compute_reward_consumption_energy(i, action)
-        reward_dispersion_remaining_energy = self.compute_reward_dispersion_remaining_energy()
+        # reward_dispersion_remaining_energy = self.compute_reward_dispersion_remaining_energy()
         reward_number_of_packets = self.compute_reward_number_of_packets(action)
 
-        return [reward_angle, reward_distance, reward_consumption_energy, reward_dispersion_remaining_energy, reward_number_of_packets]
+        # return [reward_angle, reward_distance, reward_consumption_energy, reward_dispersion_remaining_energy, reward_number_of_packets]
+        return [reward_angle, reward_distance, reward_consumption_energy, reward_number_of_packets]
 
     def network_reward_dispersion_remaining_energy(self):
         '''
@@ -277,7 +278,7 @@ class WSNRoutingEnv(gym.Env):
         max_reception_energy = self.reception_energy(self.n_sensors * initial_number_of_packets)
         normalized_total_energy = total_energy / (max_transmission_energy + max_reception_energy)
 
-        return 1 - normalized_total_energy
+        return np.clip(1 - normalized_total_energy, 0, 1)
 
     def integrate_mobility(self):
         '''

{gym_examples-2.0.50.dist-info → gym_examples-2.0.52.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: gym-examples
-Version: 2.0.50
+Version: 2.0.52
 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-2.0.52.dist-info/RECORD

@@ -0,0 +1,7 @@
+gym_examples/__init__.py,sha256=6aaeH8rkfL_LI7cvWeIw12KjsNXo-mMyDfYWUedc1AU,193
+gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
+gym_examples/envs/wsn_env.py,sha256=0B3NR4ovrgafYv1oIm-EG_tnReB965W5QICY7rQnKVk,14999
+gym_examples-2.0.52.dist-info/METADATA,sha256=kZ8UE_UTc9Qyoo82ZMIyZWBMCRLtBvRcup6mKDBSbcE,411
+gym_examples-2.0.52.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
+gym_examples-2.0.52.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
+gym_examples-2.0.52.dist-info/RECORD,,

gym_examples-2.0.50.dist-info/RECORD

@@ -1,7 +0,0 @@
-gym_examples/__init__.py,sha256=vQ37pkZ5W5E2Poo_7k2_PSIW-D1ekmFKIbmjgp4quqg,193
-gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
-gym_examples/envs/wsn_env.py,sha256=I9UO0beru5MhLxxb8JRMCCLqm6ZvKZ24blYD0cFuUtg,14785
-gym_examples-2.0.50.dist-info/METADATA,sha256=Nd4YwzGfy1O3zWqQ-AmE3WmpUIMEd2vL122tLuVkUyI,411
-gym_examples-2.0.50.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
-gym_examples-2.0.50.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
-gym_examples-2.0.50.dist-info/RECORD,,