gym-examples 3.0.176__py3-none-any.whl → 3.0.177__py3-none-any.whl

gym_examples/__init__.py

@@ -6,4 +6,4 @@ register(
      max_episode_steps=50,
 )
 
-__version__ = "3.0.176"
+__version__ = "3.0.177"

gym_examples/envs/wsn_env.py

@@ -27,6 +27,8 @@ latency_per_hop = 1  # latency per hop in seconds
 coef_network_throughput = 2.6  # coefficient for the network throughput reward
 coef_packet_delivery_ratio = 1.6  # coefficient for the packet delivery ratio reward
 
+base_back_up_dir = "results/data/"
+
 # Define the final reward function using an attention mechanism
 class Attention(nn.Module):
     def __init__(self, input_dim, output_dim):
@@ -56,6 +58,18 @@ class WSNRoutingEnv(gym.Env):
         self.log_done = None # Log the done status of the environment
         self.log_action = None # Log the action taken by the agent
         self.log_steps = None # Log the number of steps taken by the agent
+
+        # Initialize list of episode metrics
+        self.number_of_steps = 0 # Total number of steps taken by the agent since the beginning of the training 
+        self.episode_returns = []
+        self.episode_std_remaining_energy = []
+        self.episode_mean_remaining_energy = []
+        self.episode_total_consumption_energy = []        
+        self.episode_network_throughput = []
+        self.episode_packet_delivery_ratio = []
+        self.episode_network_lifetime = []
+        self.episode_average_latency = []
+
         self.n_sensors = n_sensors
         self.n_agents = n_sensors
         self.coverage_radius = coverage_radius
@@ -75,7 +89,18 @@ class WSNRoutingEnv(gym.Env):
         self.reset()        
 
     def reset(self):
+
+        episode_return = 0
+
         if self.episode_count > 1 and os.getenv('PRINT_STATS') == 'True':
+            self.episode_returns.append(episode_return)
+            self.episode_std_remaining_energy.append(np.std(self.remaining_energy))
+            self.episode_mean_remaining_energy.append(np.mean(self.remaining_energy))
+            self.episode_total_consumption_energy.append(np.sum(initial_energy - self.remaining_energy))
+            self.episode_network_throughput.append(self.network_throughput)
+            self.episode_packet_delivery_ratio.append(self.packet_delivery_ratio)
+            self.episode_network_lifetime.append(self.network_lifetime)
+            self.episode_average_latency.append(self.average_latency)
             print(f"Episode: {self.episode_count}")
             print(f"This episode ends with # Steps: {self.log_steps}")
             print(f"This episode ends with Done: {self.log_done}")
@@ -106,6 +131,8 @@ class WSNRoutingEnv(gym.Env):
     def step(self, actions):
         actions = [actions[i] for i in range(self.n_agents)] # We want to go back from the MultiDiscrete action space to a tuple of tuple of Discrete action spaces
         self.steps += 1
+        if os.getenv('PRINT_STATS') == 'True':
+            self.number_of_steps += 1 
         rewards = [0] * self.n_sensors
         dones = [False] * self.n_sensors
         for i, action in enumerate(actions):
@@ -192,6 +219,28 @@ class WSNRoutingEnv(gym.Env):
         self.log_done = dones
         self.log_action = actions
         self.log_steps = self.steps
+        episode_return += rewards
+
+        if os.getenv('PRINT_STATS') == 'True': # We are trying to extract only the statistics for the PPO algorithm
+            if self.number_of_steps >= num_timesteps:
+                self.episode_returns.append(episode_return)
+                self.episode_std_remaining_energy.append(np.std(self.remaining_energy))
+                self.episode_mean_remaining_energy.append(np.mean(self.remaining_energy))
+                self.episode_total_consumption_energy.append(np.sum(initial_energy - self.remaining_energy))
+                self.episode_network_throughput.append(self.network_throughput)
+                self.episode_packet_delivery_ratio.append(self.packet_delivery_ratio)
+                self.episode_network_lifetime.append(self.network_lifetime)
+                self.episode_average_latency.append(self.average_latency)
+
+                np.save(f"{base_back_up_dir}returns_QMIX_{version}.npy", np.array(self.episode_returns))
+                np.save(f"{base_back_up_dir}std_remaining_energy_QMIX_{version}.npy", np.array(self.episode_std_remaining_energy))
+                np.save(f"{base_back_up_dir}total_consumption_energy_QMIX_{version}.npy", np.array(self.episode_total_consumption_energy))
+                np.save(f"{base_back_up_dir}mean_remaining_energy_QMIX_{version}.npy", np.array(self.episode_mean_remaining_energy))
+                np.save(f"{base_back_up_dir}network_throughput_QMIX_{version}.npy", np.array(self.episode_network_throughput))
+                np.save(f"{base_back_up_dir}packet_delivery_ratio_QMIX_{version}.npy", np.array(self.episode_packet_delivery_ratio))
+                np.save(f"{base_back_up_dir}network_lifetime_QMIX_{version}.npy", np.array(self.episode_network_lifetime))
+                np.save(f"{base_back_up_dir}average_latency_QMIX_{version}.npy", np.array(self.episode_average_latency))
+
         return self._get_obs(), rewards, dones, {}
 
     def _get_obs(self):

{gym_examples-3.0.176.dist-info → gym_examples-3.0.177.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: gym-examples
-Version: 3.0.176
+Version: 3.0.177
 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-3.0.177.dist-info/RECORD

@@ -0,0 +1,7 @@
+gym_examples/__init__.py,sha256=j60kGbCc8FuDrR2oC1Aj0dXR1XU7IFCPNOHTRS7pqSw,194
+gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
+gym_examples/envs/wsn_env.py,sha256=iTa-ZOhp1dcQD01KfqB70UbJQI79B9wQeY6fkRpz_rE,26207
+gym_examples-3.0.177.dist-info/METADATA,sha256=m1AYDNC878oKPiGZ4dJcOEaxak6JFdLUvEuvSU31yoc,412
+gym_examples-3.0.177.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
+gym_examples-3.0.177.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
+gym_examples-3.0.177.dist-info/RECORD,,

gym_examples-3.0.176.dist-info/RECORD

@@ -1,7 +0,0 @@
-gym_examples/__init__.py,sha256=i7wmRRD6heJrXozHrhxzTPi1vwiSSWbwl0KImDTLlHI,194
-gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
-gym_examples/envs/wsn_env.py,sha256=YqXpYZXdd5lW3m2Hohog4iirLHORObEECs3QVr341Rg,22968
-gym_examples-3.0.176.dist-info/METADATA,sha256=TTwYP8pX18ClNKCpUqR0oHggJLZgcrzxl5pV7VSn3t0,412
-gym_examples-3.0.176.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
-gym_examples-3.0.176.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
-gym_examples-3.0.176.dist-info/RECORD,,