gym-examples 3.0.283__py3-none-any.whl → 3.0.285__py3-none-any.whl

gym_examples/__init__.py

@@ -5,4 +5,4 @@ register(
      entry_point="gym_examples.envs:WSNRoutingEnv",
 )
 
-__version__ = "3.0.283"
+__version__ = "3.0.285"

gym_examples/envs/wsn_env.py

@@ -74,7 +74,6 @@ class WSNRoutingEnv(gym.Env):
         self.episode_count = 0
         self.scale_displacement = 0.01 * (upper_bound - lower_bound) # scale of the random displacement of the sensors
         self.epsilon = 1e-10 # small value to avoid division by zero
-        self.current_sensor = 0 # Index of the current sensor
 
         # Define observation space
         self.observation_space = Tuple(
@@ -84,7 +83,8 @@ class WSNRoutingEnv(gym.Env):
         # self.action_space = Tuple(tuple([Discrete(self.n_sensors + 1)] * self.n_agents))
         # self.action_space = MultiDiscrete([self.n_sensors + 1] * self.n_agents)
         # self.action_space = MultiDiscrete([self.n_agents, self.n_sensors + 1]) 
-        self.action_space = Discrete(self.n_sensors + 1)  # +1 for the base station
+        # self.action_space = Discrete(self.n_sensors + 1)  # +1 for the base station
+        self.action_space = Discrete((self.n_sensors + 1)**self.n_agents)
 
         self.reset()        
 
@@ -118,11 +118,10 @@ class WSNRoutingEnv(gym.Env):
 
     def step(self, actions):
         self.steps += 1 
-        rewards = - max_reward
-        # rewards = 0
-        dones = False
-        actions = np.array([self.current_sensor, actions])
-        for i, action in [actions]:
+        rewards = [-max_reward] * self.n_sensors
+        dones = [False] * self.n_sensors
+        actions = self.to_base_n(actions, self.n_sensors + 1)
+        for i, action in enumerate(actions):
             if self.remaining_energy[i] <= 0 or self.number_of_packets[i] <= 0:
                 continue  # Skip if sensor has no energy left or no packets to transmit
             
@@ -137,11 +136,6 @@ class WSNRoutingEnv(gym.Env):
                 transmission_energy = self.transmission_energy(self.number_of_packets[i], self.distance_to_base[i])
                 if self.remaining_energy[i] < transmission_energy:
                     self.remaining_energy[i] = 0
-                    next_sensor = self.find_next_sensor()
-                    if next_sensor is None:
-                        dones = True
-                    else:
-                        self.current_sensor = next_sensor
                     continue  # Skip if the sensor does not have enough energy to transmit data to the base station
                 
                 self.update_sensor_energies(i, transmission_energy)
@@ -153,12 +147,8 @@ class WSNRoutingEnv(gym.Env):
                 self.total_latency += self.packet_latency[i] + latency_per_hop
                 self.packet_latency[i] = 0
 
-                rewards = max_reward # Reward for transmitting data to the base station
-                next_sensor = self.find_next_sensor()
-                if next_sensor is None:
-                    dones = True
-                else:
-                    self.current_sensor = next_sensor
+                rewards[i] = max_reward # Reward for transmitting data to the base station
+                dones[i] = True
             else:
                 distance = np.linalg.norm(self.sensor_positions[i] - self.sensor_positions[action])
                 if distance > self.coverage_radius:
@@ -167,21 +157,11 @@ class WSNRoutingEnv(gym.Env):
                 transmission_energy = self.transmission_energy(self.number_of_packets[i], distance)
                 reception_energy = self.reception_energy(self.number_of_packets[i])
                 if self.remaining_energy[i] < transmission_energy:
-                    self.remaining_energy[i] = 0
-                    next_sensor = self.find_next_sensor()
-                    if next_sensor is None:
-                        dones = True
-                    else:
-                        self.current_sensor = next_sensor 
+                    self.remaining_energy[i] = 0 
                     continue  # Skip if the sensor does not have enough energy to transmit data to the next hop
                 if self.remaining_energy[action] < reception_energy:
                     self.number_of_packets[i] = 0
                     self.remaining_energy[action] = 0
-                    next_sensor = self.find_next_sensor()
-                    if next_sensor is None:
-                        dones = True
-                    else:
-                        self.current_sensor = next_sensor
                     continue  # Skip if the next hop does not have enough energy to receive data
 
                 self.update_sensor_energies(i, transmission_energy)  
@@ -193,16 +173,15 @@ class WSNRoutingEnv(gym.Env):
                 self.packet_latency[action] += self.packet_latency[i] + latency_per_hop
                 self.packet_latency[i] = 0
 
-                rewards = self.compute_individual_rewards(i, action)  
+                rewards[i] = self.compute_individual_rewards(i, action) 
 
                 # Update the number of packets
                 self.number_of_packets[action] += self.number_of_packets[i]
                 
-                self.current_sensor = action
             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 = np.mean(rewards)
+            rewards[i] = np.mean(rewards[i])
         
         # Integrate the mobility of the sensors
         # self.integrate_mobility() 
@@ -214,9 +193,13 @@ class WSNRoutingEnv(gym.Env):
 
         self.get_metrics()
 
-        rewards = rewards.item() if isinstance(rewards, torch.Tensor) else rewards # Convert the reward to a float
-        if not dones:
-            dones = all(self.remaining_energy[i] <= 0 or self.number_of_packets[i] == 0 for i in range(self.n_sensors))
+        rewards = [reward.item() if isinstance(reward, torch.Tensor) else reward for reward in rewards] # Convert the reward to a float
+        rewards = np.mean(rewards) # Average the rewards
+        
+        for i in range(self.n_sensors):
+            if not dones[i]:
+                dones[i] = self.remaining_energy[i] <= 0 or self.number_of_packets[i] == 0
+        dones = np.all(dones)
 
         return self._get_obs(), rewards, dones, self.get_metrics()
 
@@ -225,8 +208,7 @@ class WSNRoutingEnv(gym.Env):
         return [{'remaining_energy': np.array([e]), 
                  'consumption_energy': np.array([initial_energy - e]),
                  'sensor_positions': p,
-                 'number_of_packets': np.array([d]),
-                 'curent_sensor': np.array([self.current_sensor])
+                 'number_of_packets': np.array([d])
                 } for e, p, d in zip(self.remaining_energy, self.sensor_positions, self.number_of_packets)]
 
 
@@ -235,8 +217,7 @@ class WSNRoutingEnv(gym.Env):
         ('remaining_energy', Box(low=0, high=initial_energy, shape=(1,), dtype=np.float64)),
         ('consumption_energy', Box(low=0, high=initial_energy, shape=(1,), dtype=np.float64)),
         ('sensor_positions', Box(low=lower_bound, high=upper_bound, shape=(2,), dtype=np.float64)),
-        ('number_of_packets', Box(low=0, high=self.n_sensors * initial_number_of_packets + 1, shape=(1,), dtype=int)),
-        ('current_sensor', Box(low=0, high=self.n_sensors - 1, shape=(1,), dtype=int))
+        ('number_of_packets', Box(low=0, high=self.n_sensors * initial_number_of_packets + 1, shape=(1,), dtype=int))
     ]))
 
 
@@ -498,4 +479,16 @@ class WSNRoutingEnv(gym.Env):
             next_index = (self.current_sensor + offset) % self.n_sensors
             if self.remaining_energy[next_index] > 0 and self.number_of_packets[next_index] > 0:
                 return next_index
-        return None  # If no such sensor is found
+        return None  # If no such sensor is found
+    
+
+    def to_base_n(self, number, base):
+        """Convert a number to a base-n number."""
+        if number == 0:
+            return [0] * (base - 1)
+        
+        digits = []
+        while number:
+            digits.append(number % base)
+            number //= base
+        return digits[::-1]  # Reverse the list to get the correct order

{gym_examples-3.0.283.dist-info → gym_examples-3.0.285.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: gym-examples
-Version: 3.0.283
+Version: 3.0.285
 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-3.0.285.dist-info/RECORD

@@ -0,0 +1,7 @@
+gym_examples/__init__.py,sha256=RG4x72mbBXJgNSIrzig4Fq6xeAVsPtwrVaN8fHLoVHc,166
+gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
+gym_examples/envs/wsn_env.py,sha256=7AKFR_Q4I_GADV3Q78g-JYacUSIwRDzjxLRZ5X9lzFc,23610
+gym_examples-3.0.285.dist-info/METADATA,sha256=fR_FUKKdoAUxYEjQ8KbeVC4quinNkfxfqShtSaWFMlo,412
+gym_examples-3.0.285.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
+gym_examples-3.0.285.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
+gym_examples-3.0.285.dist-info/RECORD,,

gym_examples-3.0.283.dist-info/RECORD

@@ -1,7 +0,0 @@
-gym_examples/__init__.py,sha256=B6nFhjmZ3o9wglL3vYZps18eP8W7b436z2-pBFs_-2w,166
-gym_examples/envs/__init__.py,sha256=lgMe4pyOuUTgTBUddM0iwMlETsYTwFShny6ifm8PGM8,53
-gym_examples/envs/wsn_env.py,sha256=pi4-ErjIpxM1jrWjMr8vbUHmp0YSY9YaGSTXsL2k50I,24121
-gym_examples-3.0.283.dist-info/METADATA,sha256=0J7nhDUnmHi_7HwHNYycQqRikH_nlbSNNGmVVf6ujm0,412
-gym_examples-3.0.283.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
-gym_examples-3.0.283.dist-info/top_level.txt,sha256=rJRksoAF32M6lTLBEwYzRdo4PgtejceaNnnZ3HeY_Rk,13
-gym_examples-3.0.283.dist-info/RECORD,,