gym-examples 2.0.77__py3-none-any.whl → 2.0.79__py3-none-any.whl

gym_examples/__init__.py

@@ -6,4 +6,4 @@ register(
      max_episode_steps=50,
 )
 
-__version__ = "2.0.77"
+__version__ = "2.0.79"

gym_examples/envs/wsn_env.py

@@ -130,10 +130,10 @@ class WSNRoutingEnv(gym.Env):
             rewards[i] = final_reward
 
         # 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]
-        rewards = [0.5 * r + 0.5 * (self.network_reward_consumption_energy() + self.network_reward_dispersion_remaining_energy()) for r in rewards]
+        # rewards = [0.5 * r + 0.5 * (self.network_reward_consumption_energy() + self.network_reward_dispersion_remaining_energy()) for r in rewards]
         
         # Only proceed if network consumption energy is not zero to avoid unnecessary list comprehension
-        self.rewards_individual = [r for r in self.rewards_individual if ((r != 0) and (r[3] < 1))]
+        self.rewards_individual = [r for r in self.rewards_individual if ((r != 0) and (r[len(r) -1] < 1))]
 
         # self.rewards_individual = [{"ind": r, "net_consumption_energy": self.network_reward_consumption_energy(), "net_dispersion_energy": self.network_reward_dispersion_remaining_energy()} for r in self.rewards_individual if ((r != 0) and (self.network_reward_consumption_energy() != 0))]
         for i in range(self.n_sensors):
@@ -239,16 +239,26 @@ class WSNRoutingEnv(gym.Env):
 
         return np.clip(1 - normalized_total_energy, 0, 1)
     
-    def compute_reward_dispersion_remaining_energy(self):
+    def compute_reward_dispersion_remaining_energy(self,i):
         '''
-        Compute the reward based on the standard deviation of the remaining energy
+        Compute the reward based on the difference between the remaining energy of the sensor i and the mean remaining energy of all sensors
         '''
-        dispersion_remaining_energy = np.std(self.remaining_energy)
-        # Normalize the standard deviation of the remaining energy
-        max_dispersion_remaining_energy = initial_energy / 2 # maximum standard deviation of the remaining energy if n_sensors is even
-        normalized_dispersion_remaining_energy = dispersion_remaining_energy / max_dispersion_remaining_energy
-
-        return np.clip(1 - normalized_dispersion_remaining_energy, 0, 1)
+        difference = np.abs(self.remaining_energy[i] - np.mean(self.remaining_energy))
+        # Normalize the difference
+        normalized_difference = difference / initial_energy
+
+        return np.clip(1 - normalized_difference, 0, 1)
+
+    # def compute_reward_dispersion_remaining_energy(self):
+    #     '''
+    #     Compute the reward based on the standard deviation of the remaining energy
+    #     '''
+    #     dispersion_remaining_energy = np.std(self.remaining_energy)
+    #     # Normalize the standard deviation of the remaining energy
+    #     max_dispersion_remaining_energy = initial_energy / 2 # maximum standard deviation of the remaining energy if n_sensors is even
+    #     normalized_dispersion_remaining_energy = dispersion_remaining_energy / max_dispersion_remaining_energy
+
+    #     return np.clip(1 - normalized_dispersion_remaining_energy, 0, 1)
     
     def compute_reward_number_of_packets(self, action):
         '''
@@ -268,13 +278,13 @@ 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_consumption_energy = self.compute_reward_consumption_energy(i, action)
+        reward_dispersion_remaining_energy = self.compute_reward_dispersion_remaining_energy(i)
         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_number_of_packets]
-        return [reward_angle, reward_distance, reward_number_of_packets]
+        return [reward_angle, reward_distance, reward_dispersion_remaining_energy, reward_number_of_packets]
 
     def network_reward_dispersion_remaining_energy(self):
         '''

{gym_examples-2.0.77.dist-info → gym_examples-2.0.79.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: gym-examples
-Version: 2.0.77
+Version: 2.0.79
 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-2.0.79.dist-info/RECORD

@@ -0,0 +1,7 @@
+gym_examples/__init__.py,sha256=60kE6HFM4-uZSatHGKbkRoOAaZQGRuOZAcrz7NDyiPI,193
+gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
+gym_examples/envs/wsn_env.py,sha256=SyD3DwXcsuSpLHxZhaVFnL3oe4KbdvQvP-bm8qIquhs,16706
+gym_examples-2.0.79.dist-info/METADATA,sha256=Yq5hVewL-_sXk8r7ATRsBENfobdaCx5qLp6jBvwNkn4,411
+gym_examples-2.0.79.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
+gym_examples-2.0.79.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
+gym_examples-2.0.79.dist-info/RECORD,,

gym_examples-2.0.77.dist-info/RECORD

@@ -1,7 +0,0 @@
-gym_examples/__init__.py,sha256=SnKbKQ-A6nJGr_zxdY0mkzTXHxF2i5hRV_V-co2F0CE,193
-gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
-gym_examples/envs/wsn_env.py,sha256=Fct3VXMPwNUTZyLAAPGGq1RLG282VKnohIds_d1qqv8,16166
-gym_examples-2.0.77.dist-info/METADATA,sha256=bd1lNeat-U4vFn2q_k6u1Ve_RycnBwsCmfdOGKZcCnI,411
-gym_examples-2.0.77.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
-gym_examples-2.0.77.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
-gym_examples-2.0.77.dist-info/RECORD,,