-
Notifications
You must be signed in to change notification settings - Fork 13
/
Copy pathtraining_with_rome_traces.py
157 lines (137 loc) · 8.61 KB
/
training_with_rome_traces.py
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
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
from environment.batch_migration_env import EnvironmentParameters
from environment.batch_migration_env import BatchMigrationEnv
from baselines.linear_baseline import LinearTimeBaseline
from baselines.rnn_critic_network_baseline import RNNCriticNetworkBaseline
from policies.rnn_policy_with_action_input import RNNPolicy
from policies.rnn_policy_with_action_input import RNNValueNet
from policies.rnn_policy_with_action_input import RNNPolicyWithValue
from policies.optimal_solution import optimal_solution_for_batch_system_infos
from policies.fc_categorical_policy import FCCategoricalPolicy
from policies.fc_categorical_policy import FCCategoricalPolicyWithValue
from policies.fc_categorical_policy import FCValueNetwork
from policies.rnn_critic_network import RNNValueNetwork
from policies.random_migrate_policy import RandomMigratePolicy
from policies.always_migrate_policy import AlwaysMigratePolicy
from sampler.migration_sampler import MigrationSamplerProcess
from sampler.migration_sampler import MigrationSampler
from sampler.migration_sampler import EvaluationSampler
from algorithms.dracm import DRACM
from dracm_trainer import Trainer
import itertools
import numpy as np
import tensorflow as tf
import os
os.environ["CUDA_VISIBLE_DEVICES"]="-1"
from utils import logger
if __name__ == "__main__":
number_of_base_state = 64
x_base_state = 8
y_base_state = 8
# possion_rate_vector = np.random.randint(10, 31, size=number_of_base_state)
# print("possion_rate_vector is: ", repr(possion_rate_vector))
# 40.0, 36.0, 32.0, 28.0, 24.0,
logger.configure(dir="./log/ppo-rome-with-optimal-100-bs-64-new", format_strs=['stdout', 'log', 'csv'])
# bs number = 64
possion_rate_vector = [7, 10, 8, 14, 15, 6, 20, 18, 11, 17, 20, 9, 8, 14, 9, 15, 8, 17, 9, 9, 10, 7, 17, 10,
13, 12, 5, 8, 10, 13, 19, 15, 10, 9, 10, 18, 12, 13, 5, 11, 7, 8, 8, 19, 15, 15, 6, 10,
5, 20, 17, 5, 5, 16, 5, 19, 19, 19, 9, 20, 17, 14, 17, 17]
env_default_parameters = EnvironmentParameters(trace_start_index=0,
num_traces=100,
server_frequency=128.0, # GHz
num_base_station=number_of_base_state,
optical_fiber_trans_rate=500.0,
backhaul_coefficient=0.02,
migration_coefficient_low=1.0,
migration_coefficient_high =3.0,
server_poisson_rate=possion_rate_vector, client_poisson_rate=2,
server_task_data_lower_bound=(0.05 * 1000.0 * 1000.0 * 8),
server_task_data_higher_bound=(5 * 1000.0 * 1000.0 * 8),
client_task_data_lower_bound=(0.05 * 1000.0 * 1000.0 * 8),
client_task_data_higher_bound=(5 * 1000.0 * 1000.0 * 8),
migration_size_low=0.5,
migration_size_high=100.0,
ratio_lower_bound=200.0,
ratio_higher_bound=10000.0,
map_width=8000.0, map_height=8000.0,
num_horizon_servers=x_base_state, num_vertical_servers=y_base_state,
traces_file_path='./environment/rome_traces_coordinate.txt',
transmission_rates=[60.0, 48.0, 36.0, 24.0, 12.0], # Mbps
trace_length=100,
trace_interval=12,
is_full_observation=False,
is_full_action=True)
env_eval_parameters = EnvironmentParameters(trace_start_index=120,
num_traces=30,
server_frequency=128.0, # GHz
num_base_station=number_of_base_state,
optical_fiber_trans_rate=500.0,
backhaul_coefficient=0.02,
migration_coefficient_low=1.0,
migration_coefficient_high=3.0,
server_poisson_rate=possion_rate_vector,
client_poisson_rate=2,
server_task_data_lower_bound=(0.05 * 1000.0 * 1000.0 * 8),
server_task_data_higher_bound=(5 * 1000.0 * 1000.0 * 8),
client_task_data_lower_bound=(0.05 * 1000.0 * 1000.0 * 8),
client_task_data_higher_bound=(5 * 1000.0 * 1000.0 * 8),
migration_size_low=0.5,
migration_size_high=100.0,
ratio_lower_bound=200.0,
ratio_higher_bound=10000.0,
map_width=8000.0, map_height=8000.0,
num_horizon_servers=x_base_state, num_vertical_servers=y_base_state,
traces_file_path='./environment/rome_traces_coordinate.txt',
transmission_rates=[60.0, 48.0, 36.0, 24.0, 12.0], # Mbps
trace_length=100,
trace_interval=12,
is_full_observation=False,
is_full_action=True)
env = BatchMigrationEnv(env_default_parameters)
eval_env = BatchMigrationEnv(env_eval_parameters)
print("action dim of the environment: ", env._action_dim)
rnn_policy = RNNPolicyWithValue(observation_dim=env._state_dim,
action_dim=env._action_dim,
rnn_parameter=256,
embbeding_size=2)
vf_baseline = RNNCriticNetworkBaseline(rnn_policy)
sampler = MigrationSampler(env,
policy=rnn_policy,
batch_size=4800,
num_environment_per_core=2,
max_path_length=100,
parallel=True,
num_process=8,
is_norm_reward=True) # 2 * 4 * 30
eval_sampler = EvaluationSampler(eval_env,
policy=rnn_policy,
batch_size=10,
max_path_length=100)
sampler_process = MigrationSamplerProcess(baseline=vf_baseline,
discount=0.99,
gae_lambda=0.95,
normalize_adv=True,
positive_adv=False)
algo = DRACM(policy=rnn_policy,
value_function=rnn_policy,
policy_optimizer=tf.keras.optimizers.Adam(1e-3),
value_optimizer=tf.keras.optimizers.Adam(1e-3),
is_rnn=True,
is_shared_critic_net=True,
num_inner_grad_steps=4,
clip_value=0.2,
vf_coef=0.5,
max_grad_norm=1.0,
entropy_coef=0.03)
trainer = Trainer(train_env=env,
eval_env=eval_env,
algo=algo,
sampler=sampler,
sample_processor=sampler_process,
update_batch_size=480,
policy=rnn_policy,
n_itr=120,
save_interval=5,
eval_sampler=eval_sampler,
test_interval=10,
save_path = './checkpoints_ppo_64-bs-new-2/model_checkpoint_epoch_')
trainer.train(rnn_policy=True, is_test=False)