forked from liangyiting/DDPG-for-process-optimization-
-
Notifications
You must be signed in to change notification settings - Fork 1
/
replay_buffer.py
34 lines (27 loc) · 987 Bytes
/
replay_buffer.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
from collections import deque
import random
class ReplayBuffer(object):
def __init__(self, buffer_size):
self.buffer_size = buffer_size
self.num_experiences = 0
self.buffer = deque()
def get_batch(self, batch_size):
# Randomly sample batch_size examples
return random.sample(self.buffer, batch_size)
def size(self):
return self.buffer_size
def add(self, state, action, reward, new_state, done):
experience = (state, action, reward, new_state, done)
if self.num_experiences < self.buffer_size:
self.buffer.append(experience)
self.num_experiences += 1
else:
self.buffer.popleft()
self.buffer.append(experience)
def count(self):
# if buffer is full, return buffer size
# otherwise, return experience counter
return self.num_experiences
def erase(self):
self.buffer = deque()
self.num_experiences = 0