gym-examples 3.0.48__py3-none-any.whl → 3.0.50__py3-none-any.whl

gym_examples/__init__.py

@@ -6,4 +6,4 @@ register(
      max_episode_steps=50,
 )
 
-__version__ = "3.0.48"
+__version__ = "3.0.50"

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 = 7  # length of the individual rewards vector
+input_dim = 4  # length of the individual rewards vector
 output_dim = 1  # final reward
 
 Eelec = 50e-9  # energy consumption per bit in joules
@@ -121,7 +121,8 @@ 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] = 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])
@@ -253,7 +254,7 @@ class WSNRoutingEnv(gym.Env):
         max_transmission_energy = self.transmission_energy(self.n_sensors * initial_number_of_packets, self.coverage_radius)
         max_reception_energy = self.reception_energy(self.n_sensors * initial_number_of_packets)
         max_total_energy = max_transmission_energy + max_reception_energy
-        normalized_total_energy = total_energy / max_total_energy
+        normalized_total_energy = total_energy / (max_total_energy + self.epsilon)
 
         return np.clip(1 - normalized_total_energy, 0, 1)
     
@@ -264,7 +265,7 @@ class WSNRoutingEnv(gym.Env):
         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
+        normalized_dispersion_remaining_energy = dispersion_remaining_energy / (max_dispersion_remaining_energy + self.epsilon)
 
         return np.clip(1 - normalized_dispersion_remaining_energy, 0, 1)
     
@@ -276,7 +277,7 @@ class WSNRoutingEnv(gym.Env):
         if action == self.n_sensors:
             normalized_number_of_packets = 0
         else: 
-            normalized_number_of_packets = self.number_of_packets[action] / max_number_of_packets
+            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)
 
@@ -310,7 +311,7 @@ class WSNRoutingEnv(gym.Env):
         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
+        normalized_dispersion_remaining_energy = dispersion_remaining_energy / (max_dispersion_remaining_energy + self.epsilon)
 
         return np.clip(1 - normalized_dispersion_remaining_energy, 0, 1)
     
@@ -321,7 +322,7 @@ class WSNRoutingEnv(gym.Env):
         total_energy = self.n_sensors * initial_energy - np.sum(self.remaining_energy)
         # Normalize the total energy consumption
         max_total_energy = self.n_sensors * initial_energy
-        normalized_total_energy = total_energy / max_total_energy
+        normalized_total_energy = total_energy / (max_total_energy + self.epsilon)
 
         return np.clip(1 - normalized_total_energy, 0, 1)
     
@@ -329,7 +330,7 @@ class WSNRoutingEnv(gym.Env):
         '''
         Compute the reward based on the packet delivery ratio
         '''
-        packet_delivery_ratio = self.packets_delivered / self.total_packets_sent_by_sensors if self.total_packets_sent_by_sensors > 0 else 0
+        packet_delivery_ratio = self.packets_delivered / (self.total_packets_sent_by_sensors + self.epsilon) if self.total_packets_sent_by_sensors > 0 else 0
         return np.clip(packet_delivery_ratio, 0, 1)
     
     def compute_reward_latency(self):
@@ -338,7 +339,7 @@ class WSNRoutingEnv(gym.Env):
         '''
         # Normalize the average latency
         max_latency = self.n_sensors * self.steps
-        normalized_latency = self.total_latency / max_latency
+        normalized_latency = self.total_latency / (max_latency + self.epsilon)
 
         return np.clip(1 - normalized_latency, 0, 1)
 
@@ -346,9 +347,9 @@ class WSNRoutingEnv(gym.Env):
         '''
         Compute the reward based on the network throughput
         '''
-        network_throughput = self.packets_delivered / self.steps if self.steps > 0 else 0
+        network_throughput = self.packets_delivered / (self.steps + self.epsilon) if self.steps > 0 else 0
         maximum_throughput = self.n_sensors * initial_number_of_packets
-        normalized_throughput = network_throughput / maximum_throughput
+        normalized_throughput = network_throughput / (maximum_throughput + self.epsilon)
         return np.clip(normalized_throughput, 0, 1)
 
 
@@ -382,15 +383,15 @@ class WSNRoutingEnv(gym.Env):
     
     def get_metrics(self):
         # Calculate network throughput
-        self.network_throughput = self.packets_delivered / self.steps if self.steps > 0 else 0
+        self.network_throughput = self.packets_delivered / (self.steps + self.epsilon) if self.steps > 0 else 0
         # Calculate energy efficiency
-        self.energy_efficiency = self.packets_delivered / self.total_energy_consumed if self.total_energy_consumed > 0 else 0
+        self.energy_efficiency = self.packets_delivered / (self.total_energy_consumed + self.epsilon) if self.total_energy_consumed > 0 else 0
         # Calculate packet delivery ratio
-        self.packet_delivery_ratio = self.packets_delivered / self.total_packets_sent_by_sensors if self.total_packets_sent_by_sensors > 0 else 0
+        self.packet_delivery_ratio = self.packets_delivered / (self.total_packets_sent_by_sensors + self.epsilon) if self.total_packets_sent_by_sensors > 0 else 0
         # Calculate network lifetime
         self.network_lifetime = self.first_node_dead_time if self.first_node_dead_time is not None else self.steps
         # Calculate average latency
-        self.average_latency = self.total_latency / self.packets_delivered if self.packets_delivered > 0 else 0
+        self.average_latency = self.total_latency / (self.packets_delivered + self.epsilon) if self.packets_delivered > 0 else 0
 
         return {
             "network_throughput": self.network_throughput,

{gym_examples-3.0.48.dist-info → gym_examples-3.0.50.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: gym-examples
-Version: 3.0.48
+Version: 3.0.50
 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-3.0.50.dist-info/RECORD

@@ -0,0 +1,7 @@
+gym_examples/__init__.py,sha256=5bGZE79MM87Lsi9G4pt9pzYriZjrdSe-u9MppmESRFY,193
+gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
+gym_examples/envs/wsn_env.py,sha256=S4HYOGOsORx63BvKGPwYhDlRnVITWU6cY1ZN2lZLzDQ,19947
+gym_examples-3.0.50.dist-info/METADATA,sha256=nUkf1NtLaCSfGMpgdwb7IzO_HIVUrbS5EfFsDFCzUtE,411
+gym_examples-3.0.50.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
+gym_examples-3.0.50.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
+gym_examples-3.0.50.dist-info/RECORD,,

gym_examples-3.0.48.dist-info/RECORD

@@ -1,7 +0,0 @@
-gym_examples/__init__.py,sha256=ZehSlz7OTIZqRjGB9S80rgzaj53-KcgOU7ejr9ugqts,193
-gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
-gym_examples/envs/wsn_env.py,sha256=E8T8JqXFAQDfCrgBwVk4bibgg7v3d5pqNhxCM-aZXDo,19625
-gym_examples-3.0.48.dist-info/METADATA,sha256=_bmDbz-c9noRL-J7USttwAi02WKWq13A-fiq6JpPA2s,411
-gym_examples-3.0.48.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
-gym_examples-3.0.48.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
-gym_examples-3.0.48.dist-info/RECORD,,