gym-examples 3.0.42__py3-none-any.whl → 3.0.44__py3-none-any.whl

gym_examples/__init__.py

@@ -6,4 +6,4 @@ register(
      max_episode_steps=50,
 )
 
-__version__ = "3.0.42"
+__version__ = "3.0.44"

gym_examples/envs/wsn_env.py

@@ -10,7 +10,7 @@ import torch.nn as nn
 import torch.nn.functional as F
 
 # Define the network parameters for the final reward function
-input_dim = 4  # lenght of the individual rewards vector
+# input_dim = 7  # length of the individual rewards vector
 output_dim = 1  # final reward
 
 Eelec = 50e-9  # energy consumption per bit in joules
@@ -90,8 +90,6 @@ class WSNRoutingEnv(gym.Env):
         self.steps += 1
         rewards = [0] * self.n_sensors
         dones = [False] * self.n_sensors
-        # if self.episode_count >= 599:
-        #     print(f"Step {self.steps} with actions = {actions}")
         for i, action in enumerate(actions):
             if action not in range(self.n_sensors + 1):
                 raise ValueError("Invalid action!")
@@ -123,7 +121,7 @@ class WSNRoutingEnv(gym.Env):
                 self.total_latency += self.packet_latency[i] + latency_per_hop
                 self.packet_latency[i] = 0
 
-                rewards[i] = np.ones(input_dim) # input_dim should be equal to the number of individual rewards
+                rewards[i] = self.compute_individual_rewards(i, action)
                 dones[i] = True       
             else:
                 distance = np.linalg.norm(self.sensor_positions[i] - self.sensor_positions[action])
@@ -144,11 +142,8 @@ class WSNRoutingEnv(gym.Env):
                 self.packet_latency[action] += self.packet_latency[i] + latency_per_hop
                 self.packet_latency[i] = 0
 
-                # Compute individual rewards
-                rewards[i] = self.compute_individual_rewards(i, action)
-                reward_packet_delivery_ratio = self.compute_reward_packet_delivery_ratio()
-                reward_latency = self.compute_reward_latency()
-                rewards[i] = np.append(rewards[i], [reward_packet_delivery_ratio, reward_latency])
+                rewards[i] = self.compute_individual_rewards(i, action)                                
+
                 # Update the number of packets
                 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
@@ -169,12 +164,6 @@ class WSNRoutingEnv(gym.Env):
 
         self.get_metrics()
 
-        # reward_packet_delivery_ratio = self.compute_reward_packet_delivery_ratio()
-        # reward_latency = self.compute_reward_latency()
-        # rewards_metrics = [reward_packet_delivery_ratio, reward_latency]
-        # rewards_metrics = self.compute_attention_rewards(rewards_metrics)
-        # rewards = [torch.tensor(r, dtype=torch.float64) + rewards_metrics if isinstance(r, int) else r + rewards_metrics for r in rewards]
-
         return self._get_obs(), rewards, dones, {}
 
     def _get_obs(self):
@@ -295,15 +284,24 @@ class WSNRoutingEnv(gym.Env):
         '''
         Compute the individual rewards
         '''
+        #-- rewards related to the energy consumption minimization and energy balance
         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_number_of_packets = self.compute_reward_number_of_packets(action)
 
-        # return [reward_angle, reward_consumption_energy, reward_dispersion_remaining_energy, reward_number_of_packets]
-        return [reward_consumption_energy, reward_dispersion_remaining_energy]
-        
+        rewards_energy = np.array([reward_angle, reward_consumption_energy, reward_dispersion_remaining_energy, reward_number_of_packets])
+
+        #-- 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()
+
+        rewards_performance = np.array([reward_latency, reward_network_throughput, reward_packet_delivery_ratio])
+
+        return np.concatenate((rewards_energy, rewards_performance))
+                
     def network_reward_dispersion_remaining_energy(self):
         '''
         Compute the reward based on the standard deviation of the remaining energy at the network level
@@ -343,6 +341,15 @@ class WSNRoutingEnv(gym.Env):
 
         return np.clip(1 - normalized_latency, 0, 1)
 
+    def compute_reward_network_throughput(self):
+        '''
+        Compute the reward based on the network throughput
+        '''
+        network_throughput = self.packets_delivered / self.steps if self.steps > 0 else 0
+        maximum_throughput = self.n_sensors * initial_number_of_packets
+        normalized_throughput = network_throughput / maximum_throughput
+        return np.clip(normalized_throughput, 0, 1)
+
 
     def compute_sum_rewards(self, rewards):
         '''

{gym_examples-3.0.42.dist-info → gym_examples-3.0.44.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: gym-examples
-Version: 3.0.42
+Version: 3.0.44
 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-3.0.44.dist-info/RECORD

@@ -0,0 +1,7 @@
+gym_examples/__init__.py,sha256=qIKbuc3FY1hkHzkw8FzbDPnI1CoXn6KiYNCe0zG0MQo,193
+gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
+gym_examples/envs/wsn_env.py,sha256=BA0zRUuNy104AHvJIcSV68AxZ9BC-gouJJGWYpbC_Jo,19592
+gym_examples-3.0.44.dist-info/METADATA,sha256=0MoxuoqIDqB_-OPW9qSPtKvzT-r60xF8txnG_tHkxcg,411
+gym_examples-3.0.44.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
+gym_examples-3.0.44.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
+gym_examples-3.0.44.dist-info/RECORD,,

gym_examples-3.0.42.dist-info/RECORD

@@ -1,7 +0,0 @@
-gym_examples/__init__.py,sha256=kLTFJhJe48NKPRenv6nTfwR9FbFVrCMz4hzeIhHTrgI,193
-gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
-gym_examples/envs/wsn_env.py,sha256=P3-MHflhQ29pLwrDb2oY7NjJvbpDXuTz6bPX2_mbqPQ,19522
-gym_examples-3.0.42.dist-info/METADATA,sha256=T4wmhIiKxlhX7Mccah7D-kCaZap2pgJPtrVgyo0rMPI,411
-gym_examples-3.0.42.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
-gym_examples-3.0.42.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
-gym_examples-3.0.42.dist-info/RECORD,,