def evolutionary_network_variable_psi_transformable_state(
simulation_settings, logger, save=False):
print("evolutionary_network_variable_psi_transformable_state")
# saved in weights_es.npy
filepath = 'C://Users//REUBS_LEN//PycharmProjects//RocketLanding//weights_es_variable_psi_transformable_state.npy'
evo_strategy_parameters = {
'population_size': 100,
'action_size': 3,
'noise_standard_deviation': 0.1,
'number_of_generations': 200,
'learning_rate': 0.00025,
'state_size': 14,
'max_num_actions': 250
}
env = []
for i in range(evo_strategy_parameters['population_size'] + 1):
env.append(RocketLander(simulation_settings))
if save:
save_file = filepath
load_file = None
else:
save_file = None
load_file = filepath
evolutionary_network_transformable_state(env, evo_strategy_parameters,
logger, load_file, save_file)
from environments.rocketlander import RocketLander
from constants import LEFT_GROUND_CONTACT, RIGHT_GROUND_CONTACT
import numpy as np
if __name__ == "__main__":
# Settings holds all the settings for the rocket lander environment.
settings = {
'Side Engines': True,
'Clouds': True,
'Vectorized Nozzle': True,
'Starting Y-Pos Constant': 1,
'Initial Force': 'random'
} # (6000, -10000)}
env = RocketLander(settings)
s = env._reset()
from control_and_ai.pid import PID_Benchmark
# Initialize the PID algorithm
pid = PID_Benchmark()
left_or_right_barge_movement = np.random.randint(0, 2)
epsilon = 0.05
total_reward = 0
episode_number = 5
for episode in range(episode_number):
while (1):
a = pid.pid_algorithm(
last_game_average = -1000
last_best_noisy_game = -1000
max_game_average = -1000
noisy_game_no_longer_valid = False
#Create testing enviroment
settings = {
'Side Engines': True,
'Clouds': True,
'Vectorized Nozzle': True,
'Starting Y-Pos Constant': 1,
'Initial Force': 'random'
} # (6000, -10000)}
env = RocketLander(settings)
env.refresh(render=True)
env.render(mode="human")
env.reset()
print("-- Observations", env.observation_space)
print("-- actionspace", env.action_space)
#initialize training matrix with random states and actions
#dataX = np.random.random(( 5,num_env_variables+num_env_actions ))
dataX = tf.placeholder("float", [None, num_env_variables + num_env_actions])
#Only one output for the total score / reward
#dataY = np.random.random((5,1))
dataY = tf.placeholder("float", [None, 1])
eps = []
eps.append(OUPolicy(0, 0.2, 0.4))
eps.append(OUPolicy(0, 0.2, 0.4))
eps.append(OUPolicy(0, 0.2, 0.4))
simulation_settings = {'Side Engines': True,
'Clouds': True,
'Vectorized Nozzle': True,
'Graph': False,
'Render': False,
'Starting Y-Pos Constant': 1,
'Initial Force': 'random',
'Rows': 1,
'Columns': 2,
'Episodes': 300}
env = RocketLander(simulation_settings)
#env = wrappers.Monitor(env, '/tmp/contlunarlander', force=True, write_upon_reset=True)
FLAGS.retrain = True # Restore weights if False
FLAGS.test = False
FLAGS.num_episodes = 300
model_dir = 'C://Users//REUBS_LEN//PycharmProjects//RocketLanding//DDPG//models_unlimited_episodes_full_normalized_normal_state'
agent = DDPG(
action_bounds,
eps,
env.observation_space.shape[0],
actor_learning_rate=0.0001,
critic_learning_rate=0.001,
retrain=FLAGS.retrain,
log_dir=FLAGS.log_dir,