gym-examples 3.0.79__py3-none-any.whl → 3.0.81__py3-none-any.whl

gym_examples/__init__.py

@@ -6,4 +6,4 @@ register(
      max_episode_steps=50,
 )
 
-__version__ = "3.0.79"
+__version__ = "3.0.81"

gym_examples/envs/wsn_env.py

@@ -40,6 +40,9 @@ class Attention(nn.Module):
         x = self.linear2(x)  # Pass the result through another linear layer
         return x
 
+net = Attention(input_dim, output_dim)
+net = net.double()  # Convert the weights to Double
+
 class WSNRoutingEnv(gym.Env):
     def __init__(self, n_sensors = 20, coverage_radius=(upper_bound - lower_bound)/4):
 
@@ -88,7 +91,7 @@ class WSNRoutingEnv(gym.Env):
 
     def step(self, actions):
         self.steps += 1
-        rewards = [-1] * self.n_sensors
+        rewards = [0] * self.n_sensors
         dones = [False] * self.n_sensors
         for i, action in enumerate(actions):
             if action not in range(self.n_sensors + 1):
@@ -121,8 +124,8 @@ class WSNRoutingEnv(gym.Env):
                 self.total_latency += self.packet_latency[i] + latency_per_hop
                 self.packet_latency[i] = 0
 
-                # rewards[i] = self.compute_individual_rewards(i, action)
-                rewards[i] = np.ones(input_dim)  # Reward for transmitting data to the base station
+                rewards[i] = self.compute_individual_rewards(i, action)
+                # rewards[i] = np.ones(input_dim)  # Reward for transmitting data to the base station
                 dones[i] = True       
             else:
                 distance = np.linalg.norm(self.sensor_positions[i] - self.sensor_positions[action])
@@ -149,8 +152,8 @@ 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] = np.mean(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
@@ -222,8 +225,8 @@ class WSNRoutingEnv(gym.Env):
         # Normalize the angle
         normalized_angle = abs(angle) / np.pi
 
-        # return np.clip(1 - normalized_angle, 0, 1)
-        return np.clip(- normalized_angle, -1, 1)
+        return np.clip(1 - normalized_angle, 0, 1)
+        # return np.clip(- normalized_angle, -1, 1)
     
     def compute_reward_distance(self, i, action):
         '''
@@ -236,8 +239,8 @@ class WSNRoutingEnv(gym.Env):
         # Normalize the distance to the next hop
         normalized_distance_to_next_hop = distance / self.coverage_radius
 
-        # return np.clip(1 - normalized_distance_to_next_hop, 0, 1)
-        return np.clip(-normalized_distance_to_next_hop, -1, 1)
+        return np.clip(1 - normalized_distance_to_next_hop, 0, 1)
+        # return np.clip(-normalized_distance_to_next_hop, -1, 1)
 
     def compute_reward_consumption_energy(self, i, action):
         '''
@@ -258,8 +261,8 @@ class WSNRoutingEnv(gym.Env):
         max_total_energy = max_transmission_energy + max_reception_energy
         normalized_total_energy = total_energy / (max_total_energy + self.epsilon)
 
-        # return np.clip(1 - normalized_total_energy, 0, 1)
-        return np.clip(- normalized_total_energy, -1, 1)
+        return np.clip(1 - normalized_total_energy, 0, 1)
+        # return np.clip(- normalized_total_energy, -1, 1)
 
     def compute_reward_dispersion_remaining_energy(self):
         '''
@@ -270,8 +273,8 @@ class WSNRoutingEnv(gym.Env):
         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 + self.epsilon)
 
-        # return np.clip(1 - normalized_dispersion_remaining_energy, 0, 1)
-        return np.clip(- normalized_dispersion_remaining_energy, -1, 1)
+        return np.clip(1 - normalized_dispersion_remaining_energy, 0, 1)
+        # return np.clip(- normalized_dispersion_remaining_energy, -1, 1)
 
     def compute_reward_number_of_packets(self, action):
         '''
@@ -283,8 +286,8 @@ class WSNRoutingEnv(gym.Env):
         else: 
             normalized_number_of_packets = self.number_of_packets[action] / (max_number_of_packets + self.epsilon)
 
-        # return np.clip(1 - normalized_number_of_packets, 0, 1)
-        return np.clip(- normalized_number_of_packets, -1, 1)
+        return np.clip(1 - normalized_number_of_packets, 0, 1)
+        # return np.clip(- normalized_number_of_packets, -1, 1)
 
     def compute_individual_rewards(self, i, action):
         '''
@@ -318,8 +321,8 @@ class WSNRoutingEnv(gym.Env):
         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 + self.epsilon)
 
-        # return np.clip(1 - normalized_dispersion_remaining_energy, 0, 1)
-        return np.clip(- normalized_dispersion_remaining_energy, -1, 1)
+        return np.clip(1 - normalized_dispersion_remaining_energy, 0, 1)
+        # return np.clip(- normalized_dispersion_remaining_energy, -1, 1)
     
     def network_reward_consumption_energy(self):
         '''
@@ -330,8 +333,8 @@ class WSNRoutingEnv(gym.Env):
         max_total_energy = self.n_sensors * initial_energy
         normalized_total_energy = total_energy / (max_total_energy + self.epsilon)
 
-        # return np.clip(1 - normalized_total_energy, 0, 1)
-        return np.clip(- normalized_total_energy, -1, 1)
+        return np.clip(1 - normalized_total_energy, 0, 1)
+        # return np.clip(- normalized_total_energy, -1, 1)
     
     def compute_reward_packet_delivery_ratio(self):
         '''
@@ -348,8 +351,8 @@ class WSNRoutingEnv(gym.Env):
         max_latency = self.n_sensors * self.steps
         normalized_latency = self.total_latency / (max_latency + self.epsilon)
 
-        # return np.clip(1 - normalized_latency, 0, 1)
-        return np.clip(- normalized_latency, -1, 1)
+        return np.clip(1 - normalized_latency, 0, 1)
+        # return np.clip(- normalized_latency, -1, 1)
 
     def compute_reward_network_throughput(self):
         '''
@@ -371,10 +374,7 @@ class WSNRoutingEnv(gym.Env):
         '''
         Compute the attention-based rewards
         '''
-        input_dim = len(rewards)
         rewards = torch.tensor(rewards, dtype=torch.double)
-        net = Attention(input_dim, output_dim)
-        net = net.double()  # Convert the weights to Double
         final_reward = net(rewards)
         return final_reward 
 

{gym_examples-3.0.79.dist-info → gym_examples-3.0.81.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: gym-examples
-Version: 3.0.79
+Version: 3.0.81
 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-3.0.81.dist-info/RECORD

@@ -0,0 +1,7 @@
+gym_examples/__init__.py,sha256=-tNBdDGkUAe9gRz0t2fK_jAloQqZCns0wnEGIXGdB_s,193
+gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
+gym_examples/envs/wsn_env.py,sha256=onqDtWhDb4uXQtcpWZFILFujrH-AbVZXluV-pgYi3nA,20385
+gym_examples-3.0.81.dist-info/METADATA,sha256=Id-4MFw_RH7WUdIcUlhbAsjDQMX05OIilQgWzKJ1h00,411
+gym_examples-3.0.81.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
+gym_examples-3.0.81.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
+gym_examples-3.0.81.dist-info/RECORD,,

gym_examples-3.0.79.dist-info/RECORD

@@ -1,7 +0,0 @@
-gym_examples/__init__.py,sha256=Zx6apdel9zNMkb769ArqkX8c7CG_d8CHIpz4BQoJUnc,193
-gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
-gym_examples/envs/wsn_env.py,sha256=z4l1ctB9RlQmtxwb1Nl-QGtERQ1C_YVVhacx8fzfp7o,20434
-gym_examples-3.0.79.dist-info/METADATA,sha256=Tk9FJq9C_mAWmJoehqotrD_lB16I1t70kybNMWtHaVA,411
-gym_examples-3.0.79.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
-gym_examples-3.0.79.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
-gym_examples-3.0.79.dist-info/RECORD,,