File size: 3,138 Bytes
6f6483e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
[
  {
    "name": "HAPPO",
    "type": "On-Policy",
    "family": "HA",
    "policy": "Stochastic",
    "action_space": "Discrete/Continuous",
    "key_feature": "Heterogeneous-Agent PPO with sequential policy update"
  },
  {
    "name": "HATRPO",
    "type": "On-Policy",
    "family": "HA",
    "policy": "Stochastic",
    "action_space": "Discrete/Continuous",
    "key_feature": "Trust region optimization with Krylov subspace"
  },
  {
    "name": "HAA2C",
    "type": "On-Policy",
    "family": "HA",
    "policy": "Stochastic",
    "action_space": "Discrete/Continuous",
    "key_feature": "Advantage Actor-Critic with heterogeneous agents"
  },
  {
    "name": "MAPPO",
    "type": "On-Policy",
    "family": "MA",
    "policy": "Stochastic",
    "action_space": "Discrete/Continuous",
    "key_feature": "Multi-Agent PPO with centralized value function"
  },
  {
    "name": "SN-MAPPO",
    "type": "On-Policy",
    "family": "MA",
    "policy": "Stochastic",
    "action_space": "Discrete/Continuous",
    "key_feature": "MAPPO with stochastic network architecture"
  },
  {
    "name": "DAN-HAPPO",
    "type": "On-Policy",
    "family": "HA",
    "policy": "Stochastic",
    "action_space": "Discrete/Continuous",
    "key_feature": "Dynamic attention network for agent communication"
  },
  {
    "name": "HADDPG",
    "type": "Off-Policy",
    "family": "HA",
    "policy": "Deterministic",
    "action_space": "Continuous",
    "key_feature": "Heterogeneous-Agent DDPG with experience replay"
  },
  {
    "name": "HATD3",
    "type": "Off-Policy",
    "family": "HA",
    "policy": "Deterministic",
    "action_space": "Continuous",
    "key_feature": "Twin delayed DDPG for reduced overestimation"
  },
  {
    "name": "HASAC",
    "type": "Off-Policy",
    "family": "HA",
    "policy": "Stochastic",
    "action_space": "Continuous",
    "key_feature": "Maximum entropy RL with automatic temperature tuning"
  },
  {
    "name": "MADDPG",
    "type": "Off-Policy",
    "family": "MA",
    "policy": "Deterministic",
    "action_space": "Continuous",
    "key_feature": "Centralized critic with decentralized actors"
  },
  {
    "name": "MATD3",
    "type": "Off-Policy",
    "family": "MA",
    "policy": "Deterministic",
    "action_space": "Continuous",
    "key_feature": "Multi-Agent TD3 with clipped double Q-learning"
  },
  {
    "name": "QMIX",
    "type": "Value-Based",
    "family": "MA",
    "policy": "Greedy",
    "action_space": "Discrete",
    "key_feature": "Monotonic value factorization via mixing network"
  },
  {
    "name": "HAD3QN",
    "type": "Value-Based",
    "family": "HA",
    "policy": "Greedy",
    "action_space": "Discrete",
    "key_feature": "Dueling Double DQN for heterogeneous agents"
  },
  {
    "name": "SHOM",
    "type": "Hybrid",
    "family": "MA",
    "policy": "Mixed",
    "action_space": "Hybrid",
    "key_feature": "Shared heterogeneous observation model"
  },
  {
    "name": "2TS-VVC",
    "type": "Two-Timescale",
    "family": "Special",
    "policy": "Mixed",
    "action_space": "Hybrid",
    "key_feature": "Slow SACD + Fast DDPG for VVC coordination"
  }
]