-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathvalidate.py
148 lines (115 loc) · 5.27 KB
/
validate.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
from params import config
import numpy as np
import torch
from ppo import PPO
def validate(validation_set, model, ub_num_of_operations_per_job, release_times):
N_JOBS = validation_set[0].shape[0]
N_MACHINES = validation_set[0].shape[2]
import numpy as np
import torch
from fjsp_env.fjsp_env import FJSP
from agent_utils import greedy_select_action
from graph_pool import get_graph_pool_step
from params import device
if release_times != None:
from stochastic_arrival_times.fjsp_env.fjsp_env import StochasticFJSP
FJSP = StochasticFJSP
env = FJSP(n_j=N_JOBS, n_m=N_MACHINES, num_of_operations_ub_per_job=ub_num_of_operations_per_job)
makespans = []
for i, data in enumerate(validation_set):
if release_times != None:
adj, fea, candidate, mask, machine_feat = env.reset(data, ub_num_of_operations_per_job, release_times[i])
else:
adj, fea, candidate, mask, machine_feat = env.reset(data, ub_num_of_operations_per_job)
graph_pool_step = get_graph_pool_step(env.num_of_operations)
rewards = -env.initial_quality
while True:
fea_tensor = torch.from_numpy(np.copy(fea)).to(device)
adj_tensor = torch.from_numpy(np.copy(adj)).to(device).to_sparse()
candidate_tensor = torch.from_numpy(np.copy(candidate)).to(device)
mask_tensor = torch.from_numpy(np.copy(mask)).to(device)
machine_feat_tensor = torch.from_numpy(np.copy(machine_feat)).to(device)
with torch.no_grad():
pi, _ = model(
x=fea_tensor,
adj_matrix=adj_tensor,
candidate=candidate_tensor.unsqueeze(0),
mask=mask_tensor.unsqueeze(0),
graph_pool=graph_pool_step,
machine_feat=machine_feat_tensor.unsqueeze(0)
)
action = greedy_select_action(pi, candidate)
adj, fea, reward, done, candidate, mask, machine_feat = env.step(action)
rewards += reward
if done:
break
makespans.append(rewards - env.positive_rewards)
return np.array(makespans)
def validate_and_get_environment(validation, model, ub_num_of_operations_per_job, release_times):
N_JOBS = validation.shape[0]
N_MACHINES = validation.shape[2]
import numpy as np
import torch
from fjsp_env.fjsp_env import FJSP
from agent_utils import greedy_select_action
from graph_pool import get_graph_pool_step
from params import device
if release_times != None:
from stochastic_arrival_times.fjsp_env.fjsp_env import StochasticFJSP
FJSP = StochasticFJSP
env = FJSP(n_j=N_JOBS, n_m=N_MACHINES, num_of_operations_ub_per_job=ub_num_of_operations_per_job)
makespan = 0
if release_times != None:
adj, fea, candidate, mask, machine_feat = env.reset(validation, ub_num_of_operations_per_job, release_times)
else:
adj, fea, candidate, mask, machine_feat = env.reset(validation, ub_num_of_operations_per_job)
graph_pool_step = get_graph_pool_step(env.num_of_operations)
rewards = -env.initial_quality
while True:
fea_tensor = torch.from_numpy(np.copy(fea)).to(device)
adj_tensor = torch.from_numpy(np.copy(adj)).to(device).to_sparse()
candidate_tensor = torch.from_numpy(np.copy(candidate)).to(device)
mask_tensor = torch.from_numpy(np.copy(mask)).to(device)
machine_feat_tensor = torch.from_numpy(np.copy(machine_feat)).to(device)
with torch.no_grad():
pi, _ = model(
x=fea_tensor,
adj_matrix=adj_tensor,
candidate=candidate_tensor.unsqueeze(0),
mask=mask_tensor.unsqueeze(0),
graph_pool=graph_pool_step,
machine_feat=machine_feat_tensor.unsqueeze(0)
)
action = greedy_select_action(pi, candidate)
adj, fea, reward, done, candidate, mask, machine_feat = env.step(action)
rewards += reward
if done:
break
makespan += (rewards - env.positive_rewards)
return makespan, env
if __name__ == '__main__':
data_loaded = np.load(f'./validation/mk01_validation_set_4.npy')
validation_data = []
for i in range(data_loaded.shape[0]):
validation_data.append(data_loaded[i])
torch.manual_seed(config.torch_seed)
if torch.cuda.is_available():
torch.cuda.manual_seed_all(config.torch_seed)
np.random.seed(200)
ppo = PPO(
lr=config.learning_rate,
gamma=config.gamma,
k_epochs=config.k_epochs,
eps_clip=config.epsilon_clip,
n_j=config.n_j,
n_m=config.n_m,
num_of_layers=config.num_of_layers,
input_dim=config.input_dim,
hidden_dim=config.hidden_dim,
num_of_mlp_layers_feature_extract=config.num_of_mlp_layers_feature_extract,
num_of_mlp_layers_actor=config.num_of_mlp_layers_actor,
hidden_dim_actor=config.num_of_hidden_dim_actor,
num_of_mlp_layers_critic=config.num_of_mlp_layers_critic,
hidden_dim_critic=config.num_of_hidden_dim_critic
)
validate(validation_set=validation_data, model=ppo.policy, ub_num_of_operations_per_job=config.num_of_operations_ub_per_job)