-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmain.py
182 lines (164 loc) · 7.52 KB
/
main.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
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
import os
import pickle
import random
from itertools import product
import time
import numpy as np
import torch
from heuristic_agent import HeuristicAgent
from process_discovery_environment import ProcessDiscoveryEnvironment
from td3_agent import TD3Agent, gru_type
from utils import Order, Algo, drift_visualization, generate_csv, save_results, generate_summary
seed = 42
# fix random seed for reproducibility
def fix_seed(seed):
os.environ['PYTHONHASHSEED'] =str(seed)
os.environ['CUBLAS_WORKSPACE_CONFIG']=':4096:8'
torch.manual_seed(seed)
torch.cuda.manual_seed(seed)
torch.cuda.manual_seed_all(seed)
np.random.seed(seed)
random.seed(seed)
def agent_test(env, agent):
env.multi_reward = False
env.multi_log = False
state, info = env.reset()
episode_reward = []
# one episode
while True:
with torch.no_grad():
action = agent.make_action(state)
if type(agent).__name__ == 'HeuristicAgent':
state, reward, done, info = env.step(action)
else:
state, reward, done, info = env.step(np.squeeze(agent.scale_action(np.expand_dims(action,0),np.array(env.action_space)),0))
episode_reward.append(reward)
if done:
# print(sum(episode_reward)/len(episode_reward))
print("test log: ", env.log_name,np.array(env.evaluations[1:])[:,2].mean(), len(env.drift_moments)-1)
drift_visualization(env=env, agent_name=type(agent).__name__)
save_results(env=env, agent_name=type(agent).__name__)
generate_summary(agent_name=type(agent).__name__, log_name=log_name)
break
env.multi_reward = True
env.multi_log = True
logs = [
('BPIC2013Incidents', 800),
# ('BPIC2020DomesticDeclarations', 1000),
# ('BPIC2020InternationalDeclarations', 600),
# ('BPIC2020PermitLog', 700),
# ('BPIC2020PrepaidTravelCost', 200),
# ('BPIC2020RequestForPayment', 700),
]
memory_size_settings = [
# 10,
# 50,
# 100,
200, # default
# 300,
# 400,
# 500,
# 600,
# 700,
# 800,
# 900,
# 1000
] # 遗忘窗口长度
sampling_rate_settings = [
# 0.1,
# 0.2,
# 0.3,
# 0.4,
# 0.5,
# 0.6,
# 0.7,
0.8, # default
# 0.9,
# 1.0
] # 频率采样的阈值
orders = [
Order.FRQ,
# Order.MAX,
# Order.MIN
] # 获得最具代表性轨迹的采样方式
algos = [
Algo.IND,
# Algo.ILP
] # 静态流程发现方法
top_settings = [None] # 选择几个最具代表性的轨迹
filtering_settings = [
True,
# False
] # 流程发现方法是否采样预处理过滤
frequency_settings = [
True,
# False
] # 输入流程发现的日志是否有带频数的重复trace
update_settings = [
True,
# False
] # 是否动态重启流程发现
update_param_settings = [
True,
# False
] # 是否动态更新参数
# env 可调超参数
max_memory_size = 500 # default 500
min_memory_size = 10
max_sampling_rate = 1.0
min_sampling_rate = 0.1
history_window_settings = [
# 1,
# 5,
10, # default
# 20,
# 40
] # 指标变化检测窗口
observation_window_settings = [
# 1,
# 5,
10, # default
# 20,
# 50,
# 100
] # 每次step要算未来n个指标的均值
drift_punish = 1 # 对变动参数导致发生漂移的惩罚
memory_size_punish = 0.001 # 对遗忘窗口大小的惩罚，即考虑内存
reward_value = 'absolute' # reward是采用f值的绝对值还是相对值 'relative' / 'absolute'
# agent 可调超参数
learning_rate=1e-4 # 1e-3的reward曲线波动比较大
replay_buffer_size = 1000000
batch_size = 1024
total_episodes = 200
gru_type = gru_type # state的处理方式是直接输出gru还是分成f值序列和动作序列分别输入gru 'full' / 'fusion'
fix_seed(seed)
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
print(f'Using {device} device')
for (log_name, cut), order, algo, top, filtering, frequency, update, update_param, memory_size, sampling_rate, history_window, observation_window in \
product(logs, orders, algos, top_settings, filtering_settings, frequency_settings, update_settings, update_param_settings,
memory_size_settings, sampling_rate_settings, history_window_settings, observation_window_settings):
# generate_csv(log_name)
env = ProcessDiscoveryEnvironment(log_name, cut, algo, order, top, filtering, frequency, update, update_param, memory_size, sampling_rate,
max_memory_size, min_memory_size, max_sampling_rate, min_sampling_rate, history_window, observation_window,
drift_punish, memory_size_punish, reward_value)
print(
f'log:{log_name},cut:{cut},algo:{algo.name},order:{order.name},top:{top},filtering:{filtering},frequency:{frequency},update:{update},update_param:{update_param},memory_size:{memory_size},sampling_rate:{sampling_rate},history_window:{history_window},observation_window:{observation_window}')
start_time=time.strftime("%Y-%m-%d-%H_%M_%S",time.localtime(time.time()))
with open('experiment_record.txt','a+') as file:
file.write(f'start_time:{start_time},log:{log_name},cut:{cut},algo:{algo.name},order:{order.name},top:{top},filtering:{filtering},frequency:{frequency},update:{update},update_param:{update_param}' +'\n' +
f'init_memory_size:{memory_size},init_sampling_rate:{sampling_rate},max_memory_size:{max_memory_size},min_memory_size:{min_memory_size},max_sampling_rate:{max_sampling_rate},min_sampling_rate:{min_sampling_rate}' +'\n' +
f'history_window:{history_window},observation_window:{observation_window},drift_punish:{drift_punish},memory_size_punish:{memory_size_punish},reward_value:{reward_value},multi_log_count:{len(env.log_list)}'+'\n' +
f'learning_rate:{learning_rate},replay_buffer_size:{replay_buffer_size},batch_size:{batch_size},total_episodes:{total_episodes},gru_type:{gru_type}'+'\n\n'
)
heuristic_agent = HeuristicAgent(env=env, device=device, state_dim=env.state_dim, action_dim=env.action_dim)
agent_test(env, heuristic_agent)
td3_agent = TD3Agent(env=env, device=device, state_dim=env.state_dim, action_dim=env.action_dim, learning_rate=learning_rate, start_time=start_time,
replay_buffer_size=replay_buffer_size, batch_size=batch_size, total_episodes=total_episodes)
td3_agent.train()
# td3_agent.load('model_dir/'+start_time+'/td3_agent_best_model.pth')
agent_test(env, td3_agent)
# results={'evaluation':env.evaluations,'memory_size_list':env.memory_size_list,'sampling_rate_list':env.sampling_rate_list,'drift_flag':env.drift_flag}
# with open("fixed_hyperparameters_results.pickle", "wb") as file:
# pickle.dump(results, file)
# 命令行中用如下命令执行进行耗时分析
# pyinstrument --outfile=time_profile.html -r html main.py