MIST-Bullet/runner.py at classicalControl · rubendsa/MIST-Bullet · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
import tensorflow as tf
import time

from enum import Enum

from policy_f import PolicyFunction
from value_f import ValueFunction
from ag_tree import run_ag_tree
from environment import Environment
from nn_utils import restore_from_lowest_cost, restore_from_highest_cost, save_model, save_most_recent, restore_most_recent

class LoadType(Enum):
    CREATE_NEW_MODEL = 1
    LOAD_HIGHEST_COST = 2
    LOAD_LOWEST_COST = 3
    LOAD_MOST_RECENT = 4

def run_experiment(load_type, max_time, save_dir):
    start_time = time.time()
    OBS_SIZE = 18
    ACTION_SIZE = 4
    value_f = ValueFunction(input_dim=OBS_SIZE)
    policy_f = PolicyFunction(input_dim=OBS_SIZE, output_dim=ACTION_SIZE)

    sess = tf.Session()
    init=tf.global_variables_initializer()
    sess.run(init)
    saver = tf.train.Saver()

    env = Environment(policy_f, num_threads=1)

    best_cost = 100000000
    if load_type == LoadType.CREATE_NEW_MODEL:
        pass
    elif load_type == LoadType.LOAD_HIGHEST_COST:
        best_cost = restore_from_highest_cost(save_dir, saver, sess)
    elif load_type == LoadType.LOAD_LOWEST_COST:
        best_cost = restore_from_lowest_cost(save_dir, saver, sess)
    elif load_type == LoadType.LOAD_MOST_RECENT:
        restore_most_recent(save_dir, saver, sess)

    # i = 0
    while(True):
        # i += 1
        # print("loop {}".format(i))
        current_cost = run_ag_tree(sess, value_f, policy_f, env, rollout_len=500, num_initials_trajs=100, num_branches=200, noise_depth=2, discount_factor=0.99)

        if current_cost < best_cost:
            best_cost = current_cost
            save_model(save_dir, saver, sess, best_cost)

        #always save most recent model
        save_most_recent(save_dir, saver, sess)

        current_time = time.time()
        duration = current_time - start_time
        if duration > max_time:
            print("End of experiment, time duration reached")
            break


if __name__ == "__main__":
    save_dir = "C:/Users/user/Transformation/MIST_Bullet/models"
    time_in_seconds = 6 * 3600
    run_experiment(LoadType.CREATE_NEW_MODEL, time_in_seconds, save_dir)