-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathparameters.yaml
112 lines (112 loc) · 2.31 KB
/
parameters.yaml
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
agent_definitions:
qlearning:
type: qlearning
hidden_layer_size: 60
n_hidden_layers: 1
gamma: 0.99
dqn-30:
type: DQN
hidden_layer_size: 30
n_hidden_layers: 1
gamma: 0.99
device: cpu
sync_frequency: 100
buffer_size: 1000
sample_size: 100
dqn-60:
type: DQN
hidden_layer_size: 60
n_hidden_layers: 1
gamma: 0.99
device: cpu
sync_frequency: 100
buffer_size: 1000
sample_size: 100
1agent_obst:
map_image: maps/test_4x4.jpg
agents:
dqn-30: 1
obstacles: 1
epsilon_decay_length: 10000 # 1e4
epsilon_start: 1.0
epsilon_final: 0.01
evolution_frequency: 0
steps: 30000 # 3e4
load_model: ''
comment: ''
1agent_obst_9:
map_image: maps/test_9x9.jpg
agents:
dqn-30: 1
obstacles: 1
epsilon_decay_length: 10000 # 1e4
epsilon_start: 1.0
epsilon_final: 0.0
evolution_frequency: 0
steps: 20000 # 2e4
load_model: ''
comment: ''
3agent_obst:
map_image: maps/empty_4x4.jpg
agents:
dqn-30: 3
obstacles: 2
epsilon_decay_length: 40000 # 4e4
epsilon_start: 0.5
epsilon_final: 0.0
evolution_frequency: 0
steps: 100000 # 1e5
load_model: '3agent'
comment: ''
3agent_6:
map_image: maps/test_6x6.jpg
agents:
dqn-30: 3
obstacles: 0
epsilon_decay_length: 40000 # 4e4
epsilon_start: 0.5
epsilon_final: 0.0
evolution_frequency: 0
steps: 100000 # 1e5
load_model: '3agent_obst'
comment: ''
3agent_9:
map_image: maps/test_9x9.jpg
agents:
dqn-30: 3
obstacles: 0
epsilon_decay_length: 10000 # 1e4
epsilon_start: 0.5
epsilon_final: 0.0
evolution_frequency: 1000
steps: 10000 # 1e4
load_model: '3agent_6'
comment: ''
5agent-stage1:
map_image: maps/test_9x9.jpg
agents:
- type: dqn-60
n_agents: 5
policy: basic_policy
obstacles: 0
epsilon_decay_length: 10000 # 1e4
epsilon_start: 0.3
epsilon_final: 0.1 # no 0 epsilon with deterministic policy
goal_closeness: 5
evolution_frequency: 1000 # evolution only in case some agents diverge
steps: 10000 # 1e4
comment: ''
5agent-stage2:
map_image: maps/test_9x9.jpg
agents:
- type: dqn-60
n_agents: 5
load: 5agent-stage1/dqn-60
obstacles: 0
epsilon_decay_length: 10000 # 1e4
epsilon_start: 0.5
epsilon_final: 0.0
goal_closeness: 5
evolution_frequency: 1000 # 1e3
steps: 10000 # 1e4
comment: ''