experiments repeated

env/env_test.py

@@ -1,6 +1,6 @@
 import numpy as np
 import matplotlib.pyplot as plt
-# import time
+import time
 # from gym.spaces import Box, Dict
 from env.envs import LazyAgentsCentralized # Import your custom environment
 
@@ -27,7 +27,7 @@ def plot_trajectories(agent_positions_history, current_states, fig, ax):
  ax.grid(True)
 
  plt.draw()
- plt.pause(0.001)
+ plt.pause(0.0001)
 
 
 def plot_agents(agents_state, fig, ax):
@@ -64,7 +64,7 @@ def plot_agents(agents_state, fig, ax):
  ax.grid(True)
 
  plt.draw()
- plt.pause(0.001)
+ plt.pause(0.0001)
 
 
 def plot_std_history(std_vel_history, std_pos_history, fig, ax):
@@ -79,75 +79,133 @@ def plot_std_history(std_vel_history, std_pos_history, fig, ax):
  ax.grid(True)
 
  plt.draw()
- plt.pause(0.001)
+ plt.pause(0.0001)
 
 
 if __name__ == '__main__':
  print("Running env_test.py")
 
- num_experiment = 30
+ cases = [10, 15, 20, 25, 30]
+ mum_cases = len(cases)
+
+ num_experiment = 10
+
+ result_average_converged_time = np.zeros((mum_cases, num_experiment))
+
+ for case in range(mum_cases):
+
+ for exp in range(num_experiment):
+
+ num_agent = cases[case]
+ action = np.ones(num_agent, dtype=np.float32)
+ config = {"num_agent_max": num_agent,
+ "num_agent_min": num_agent,
+ #
+ # Optional parameters
+ "speed": 15, # Speed in m/s. Default is 15
+ "predefined_distance": 60, # Predefined distance in meters. Default is 60
+ # "communication_decay_rate": 1 / 3, # Communication decay rate. Default is 1/3
+ # "cost_weight": 1, # Cost weight. Default is 1
+ # "inter_agent_strength": 5, # Inter agent strength. Default is 5
+ # "bonding_strength": 1, # Bonding strength. Default is 1
+ # "k1": 1, # K1 coefficient. Default is 1
+ # "k2": 3, # K2 coefficient. Default is 3
+ "max_turn_rate": 8 / 15, # Maximum turn rate in rad/s. Default is 8/15
+ "initial_position_bound": 250, # Initial position bound in meters. Default is 250
+ "dt": 0.1, # Delta time in seconds. Default is 0.1
+ # "network_topology": "fully_connected", # Network topology. Default is "fully_connected"
+ # Params to tune
+ "std_pos_converged": 45, # Standard position when converged. Default is R/2
+ "std_vel_converged": 0.1, # Standard velocity when converged. Default is 0.1
+ "std_pos_rate_converged": 0.1, # Standard position rate when converged. Default is 0.1
+ "std_vel_rate_converged": 0.2, # Standard velocity rate when converged. Default is 0.01
+ "max_time_step": 1000 # Maximum time steps. Default is 1000,
+ }
+
+ env = LazyAgentsCentralized(config)
+ obs = env.reset()
+ agent_states = obs['agent_embeddings']
+ reward_sum = 0
 
- for exp in range(num_experiment):
+ # std_vel_history = []
+ # std_pos_history = []
+ # agent_positions_history = np.zeros((num_agent, env.max_time_step, 2), dtype=np.float32)
 
- num_agent = 20
- action = np.ones(num_agent, dtype=np.float32)
- config = {"num_agent_max": num_agent,
- "num_agent_min": num_agent,
- #
- "std_pos_converged": 42, # Standard position when converged. Default is R/2
- "std_vel_converged": 0.1, # Standard velocity when converged. Default is 0.1
- "std_pos_rate_converged": 0.1, # Standard position rate when converged. Default is 0.1
- "std_vel_rate_converged": 0.2, # Standard velocity rate when converged. Default is 0.01
- "max_time_step": 1000 # Maximum time steps. Default is 1000,
- }
+ # fig1, ax1 = plt.subplots()
+ # fig2, ax2 = plt.subplots()
+ # fig3, ax3 = plt.subplots()
 
- env = LazyAgentsCentralized(config)
- obs = env.reset()
- agent_states = obs['agent_embeddings']
+  done = False
+  while not done:
+  obs, reward, done, info = env.step(action)
 
- # std_vel_history = []
- # std_pos_history = []
- agent_positions_history = np.zeros((num_agent, env.max_time_step, 2), dtype=np.float32)
+ # Get the reward
+ # reward_sum += reward
 
- fig1, ax1 = plt.subplots()
- fig2, ax2 = plt.subplots()
- fig3, ax3 = plt.subplots()
+ # extract agent_states from the observations
+ # agent_states = obs['agent_embeddings']
 
- done = False
- while not done:
- obs, reward, done, info = env.step(action)
+ # update agent_positions_history with current positions
+ # agent_positions_history[:, env.time_step-1] = agent_states[:, :2]
 
- # extract agent_states from the observations
- agent_states = obs['agent_embeddings']
+ # # plot trajectories and std dev history
+ # std_pos_history = env.std_pos_hist[:env.time_step]
+ # std_vel_history = env.std_vel_hist[:env.time_step]
+ # plot_agents(agent_states, fig1, ax1)
+ # if env.time_step % 50 == 0:
+ # plot_std_history(std_vel_history, std_pos_history, fig2, ax2)
+ # plot_trajectories(agent_positions_history[:, :env.time_step], agent_states, fig3, ax3)
 
- # update agent_positions_history with current positions
- agent_positions_history[:, env.time_step-1] = agent_states[:, :2]
+ # if env.time_step % 100 == 0:
+ # print(f"Time step: {env.time_step}")
+ # time.sleep(0.01)
 
- # # plot trajectories and std dev history
+ # plot trajectories and std dev history
  # std_pos_history = env.std_pos_hist[:env.time_step]
  # std_vel_history = env.std_vel_hist[:env.time_step]
  # plot_agents(agent_states, fig1, ax1)
- # if env.time_step % 50 == 0:
- # plot_std_history(std_vel_history, std_pos_history, fig2, ax2)
- # plot_trajectories(agent_positions_history[:, :env.time_step], agent_states, fig3, ax3)
-
- # if env.time_step % 100 == 0:
- # print(f"Time step: {env.time_step}")
- # time.sleep(0.01)
-
- # plot trajectories and std dev history
- std_pos_history = env.std_pos_hist[:env.time_step]
- std_vel_history = env.std_vel_hist[:env.time_step]
- plot_agents(agent_states, fig1, ax1)
- plot_std_history(std_vel_history, std_pos_history, fig2, ax2)
- plot_trajectories(agent_positions_history[:, :env.time_step], agent_states, fig3, ax3)
-
- print("Done!")
-
- if env.time_step == env.max_time_step:
- print("Max time step reached!")
-
- plt.close(fig1)
- plt.close(fig2)
- plt.close(fig3)
+ # plot_std_history(std_vel_history, std_pos_history, fig2, ax2)
+ # plot_trajectories(agent_positions_history[:, :env.time_step], agent_states, fig3, ax3)
+
+ # std = np.std(agent_states[:, :2], axis=0).sum()
+ # print(f"Final std_mine: {std}")
+ # print(f"Accumulated reward: {reward_sum}")
+
+ print(f"Experiment {exp} of case {case} finished!")
+ print("Done!")
+
+ if env.time_step == env.max_time_step:
+ print("Max time step reached!")
+
+ # plt.close(fig1)
+ # plt.close(fig2)
+ # plt.close(fig3)
+
+ result_average_converged_time[case, exp] = env.time_step
+
+
+ # # plot the average converged time
+ # fig4, ax4 = plt.subplots()
+ # ax4.set_title('Average Converged Time')
+ # ax4.set_xlabel('Number of Agents')
+ # ax4.set_ylabel('Average Converged Time')
+ # ax4.plot(cases, np.mean(result_average_converged_time, axis=1), label='Average Converged Time')
+ # ax4.legend()
+ # ax4.grid(True)
+ # plt.draw()
+ # plt.pause(0.001)
+
+ # plot the average converged time with bar chart
+ plt.ion() # Turn on interactive mode
+ fig4, ax4 = plt.subplots()
+ ax4.set_title('Average Converged Time')
+ ax4.set_xlabel('Number of Agents')
+ ax4.set_ylabel('Average Converged Time')
+ # Use bar function instead of plot for a bar chart
+ ax4.bar(cases, np.mean(result_average_converged_time, axis=1), label='Average Converged Time')
+ ax4.legend()
+ ax4.grid(True)
+ plt.draw()
+ plt.pause(0.001)
 
+ print("Done!")