-
Notifications
You must be signed in to change notification settings - Fork 0
/
main.py
245 lines (208 loc) · 9.11 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
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
from __future__ import division
import parameters
import players2 as players
import random
import networkx as nx
import sys
import numpy as np
import scipy.stats
import csv
try:
import zmq
except ImportError:
pass
import json
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
class SeriesInstance(object):
def __init__(self, num_states, num_players):
self.numStates = num_states
self.numPlayers = num_players
self.proportion_Players = parameters.proportion_Players
self.epsilon = parameters.epsilon
self.proportion = 1 - self.epsilon
self.players_every_round=0
self.networkType = 'scaleFree'
self.playersList = []
self.playerNetwork = 0
self.meanDegree = parameters.meanDegree
self.systemState_measure_frequency = parameters.measureSystem_state_frequency
self.timeSteps = parameters.timeSteps
self.numGames = parameters.numGames
self.convergence_sequence = []
self.timeSteps_to_convergence = []
self.converged = False
# proportion of players playing each state
self.states_dynamics=dict((i,[]) for i in range(self.numStates))
self.number_of_non_convergences = 0
def createPlayers_list(self):
""" creates players """
self.playersList = [players.player() for count in xrange(self.numPlayers)]
def assignAttributes(self):
""" every agent gets a number of options, to choose his nash strategy from """
for player in self.playersList:
player.numStates = self.numStates
player.state = random.choice(range(self.numStates))
def createNetwork(self):
""" self.playerNetwork, is a network of players, which refer to Player instances
every agent gets the number of his agents """
mapping = dict(enumerate(self.playersList))
#G = nx.barabasi_albert_graph(self.numPlayers, self.meanDegree)
G = nx.watts_strogatz_graph(self.numPlayers, 2 * self.meanDegree, 0)
#G = nx.grid_2d_graph(self.numPlayers, self.numPlayers, periodic=True)
#G = nx.gnm_random_graph(self.numPlayers, 2 * self.numPlayers)
print 'mean degree', np.mean(G.degree().values())
print 'std degree', np.std(G.degree().values())
self.playerNetwork = nx.relabel_nodes(G, mapping)
def update_players_every_round(self):
a = int(self.numPlayers * self.proportion_Players)
if a % 2 == 0:
return a
else:
return a - 1
def sample_players(self):
return random.sample(self.playersList, self.players_every_round)
def collect_neighbor_states(self, players):
""" each of the two players collects the states of his neighbors """
for player in players:
neighbors_states = [neighbor.state for neighbor in self.playerNetwork.neighbors(player)]
player.update_neighbors_states(neighbors_states)
def update_state(self, players):
for player in players:
player.update_state()
def return_system_convergence(self):
states = np.zeros(self.numStates, dtype=int)
for player in self.playersList:
states[player.state] += 1
states = states / self.numPlayers
return states.max()
def update_convergence_sequence(self):
a = self.return_system_convergence()
self.convergence_sequence.append(a)
def is_converged(self):
return self.return_system_convergence() > self.proportion
def update_states_dynamics(self):
states = [0] * self.numStates
for i in self.playersList:
states[i.state] += 1
states = [x / self.numPlayers for x in states]
for i in range(self.numStates):
self.states_dynamics[i].append(states[i])
def draw_network(self, players, time):
plt.figure(1, figsize=(32, 18))
# layout graphs with positions using graphviz neato
pos = nx.graphviz_layout(self.playerNetwork, prog="neato")
# color nodes the same in each connected subgraph
nx.draw(self.playerNetwork,
pos,
node_size=[len(self.playerNetwork.neighbors(player)) * 100
for player in self.playerNetwork],
node_color=[self.color[player.state] for player in self.playerNetwork],
vmin=0.0,
vmax=1.0,
with_labels=False
)
plt.savefig("movie_{0:05d}.png".format(time), dpi=75)
def round(self, time):
players=self.sample_players()
self.collect_neighbor_states(players)
self.update_state(players)
if self.movie:
print time
self.draw_network(players, time)
def game(self):
self.createPlayers_list()
self.assignAttributes()
self.createNetwork()
self.players_every_round = self.update_players_every_round()
for step in range(self.timeSteps):
self.round(step)
if step % self.systemState_measure_frequency == 0:
if self.is_converged():
self.timeSteps_to_convergence.append(step)
return
self.number_of_non_convergences += 1
def run(self):
for _ in range(self.numGames):
self.game()
def simulate_one_parameter_set(name, num_states, num_players, movie=False, color=None):
series_instance = SeriesInstance(num_states, num_players)
series_instance.movie = movie
series_instance.color = color
series_instance.run()
mean = np.mean(series_instance.timeSteps_to_convergence)
var = np.std(series_instance.timeSteps_to_convergence)
print('players: %d, strategies: %d, mean: %.0f, non convergences: %d' % (
num_players, num_states, mean, series_instance.number_of_non_convergences))
with open(name, 'ab') as csvfile:
writer = csv.writer(csvfile, delimiter=',')
writer.writerow([num_players, num_states, mean, var,
series_instance.number_of_non_convergences,
parameters.timeSteps,
parameters.measureSystem_state_frequency,
parameters.meanDegree,
parameters.numGames,
parameters.epsilon,
parameters.proportion_Players])
def parameter_sweep():
if sys.argv[1] == 'm':
name = 'movie%s_%06d_%06d.csv' % (
sys.argv[1], int(sys.argv[2]), int(sys.argv[3]))
simulate_one_parameter_set(name, num_states=int(sys.argv[2]), num_players=int(sys.argv[3]), movie=True, color=sys.argv[4])
if sys.argv[1] == 's':
assert sys.argv[5] == 'p'
range_strategies = np.arange(int(sys.argv[2]), int(sys.argv[3]), int(sys.argv[4]))
range_players = [int(sys.argv[6])]
elif sys.argv[1] == 'p':
assert sys.argv[5] == 's'
range_players = np.arange(int(sys.argv[2]), int(sys.argv[3]), int(sys.argv[4]))
range_strategies = [int(sys.argv[6])]
elif sys.argv[1] == 'z':
pass
else:
print("python main.py s 1 100 10 p 1000")
print("python main.py s from strategy to step p number of players")
print("python main.py p 1 10000 100 s 100")
print("python main.py p from number of players to step s number of strategies")
print("python main.py m 3 10000 rgb")
print("python main.py m number_of_states number_of_players one_color_per_state")
return
if sys.argv[1] == 's' or sys.argv[1] == 'p':
name = 'final%s_%06d_%06d_%06d_%s_%06d.csv' % (
sys.argv[1], int(sys.argv[2]), int(sys.argv[3]), int(sys.argv[4]),
sys.argv[5], int(sys.argv[6]))
for num_players in range_players:
for num_states in range_strategies:
simulate_one_parameter_set(name, num_states, num_players)
elif sys.argv[1] == 'z':
offset = int(sys.argv[3])
address_task = 5557 + offset
address_result = 5558 + offset
address_kill = 5559 + offset
address_prefix = sys.argv[2]
print("address_prefix: %s, address_task: %d, address_result %d, address_kill: %d" % (
address_prefix, address_task, address_result, address_kill))
context = zmq.Context(1)
receiver = context.socket(zmq.REQ)
print("tcp://%s:%i" % (address_prefix, address_task))
receiver.connect("tcp://%s:%i" % (address_prefix, address_task))
sender = context.socket(zmq.PUSH)
sender.connect("tcp://%s:%i" % (address_prefix, address_result))
while True:
print("Receiving")
receiver.send("ready")
message = receiver.recv()
try:
param = json.loads(message)
except:
sender.send("message not parsed")
name = param['name']
num_players = int(param['num_players'])
num_states = int(param['num_states'])
print("Received and Working," , message)
simulate_one_parameter_set(name, num_states, num_players)
print("Worked and Send:")
sender.send(json.dumps({'finished': True}))
print('done')
parameter_sweep()