In [103]:
%matplotlib inline
import numpy as np
import matplotlib.pyplot as plt
import math
from basic_artificial_market import ArtificialMarket
In [104]:
import numpy as np
import matplotlib.pyplot as plt
import math
class ArtificialMarket():
def __init__(self, num_player=1000, fdmtl=10000.0, ganma=10000, sigma=0.06, P_sigma=30):
self.num_player = num_player
self.random_state = np.random.RandomState()
self.weight = self.weight()
self.fdmtl = fdmtl
self.ganma = ganma
self.sigma = sigma
self.P_sigma = P_sigma
def weight(self, w_1_max=1, w_2_max=10, w_3_max=1):
num_player = self.num_player
weight_1 = np.zeros(num_player)
weight_2 = np.zeros(num_player)
weight_3 = np.zeros(num_player)
random_state = self.random_state
for i in range(num_player):
weight_1[i] = random_state.uniform()*w_1_max
weight_2[i] = random_state.uniform()*w_2_max
weight_3[i] = random_state.uniform()*w_3_max
weight = [weight_1, weight_2, weight_3]
return weight
def one_market_model(self, w, past_data=None, b_limit=None, s_limit=None):
num_player = self.num_player
sigma = self.sigma
P_sigma = self.P_sigma
P_f = self.fdmtl
if past_data is None:
past_data = [P_f]
if b_limit is None:
b_limit = []
if s_limit is None:
s_limit = []
P_t_1 = past_data[-1]
w_1 = w[0]
w_2 = w[1]
w_3 = w[2]
ganma = self.ganma
r_t_e = np.zeros(num_player)
for i in range(num_player):
P_t_1 = past_data[-1]
if len(past_data) < ganma:
r_t_h = np.log10(P_t_1/np.random.choice(past_data))
else:
past_data_ganma = past_data[-ganma]
r_t_h = np.log10(P_t_1/past_data_ganma)
e_t = np.random.normal(0, sigma)
r_t_e[i] = (w_1[i]*np.log10(P_f/P_t_1) + w_2[i]*r_t_h + w_3[i]*e_t)/(w_1[i] + w_2[i] + w_3[i])
print r_t_h
P_e = P_t_1*math.exp(r_t_e[i])
P_o = np.random.normal(P_e, P_sigma)
if P_e > P_o:
if len(s_limit) > 0 and min(s_limit) < P_o:
P_t = np.min(s_limit)
s_limit = np.delete(s_limit, np.argmin(s_limit))
else:
b_limit = np.append(b_limit, P_o)
P_t = P_t_1
else:
if len(b_limit) > 0 and max(b_limit) > P_o:
P_t = np.max(b_limit)
b_limit = np.delete(b_limit, np.argmax(b_limit))
else:
s_limit = np.append(s_limit, P_o)
P_t = P_t_1
past_data.append(P_t)
print P_t
t = len(past_data)
print b_limit
return t, past_data, b_limit, s_limit
def one_market_simulation(self, t_max=100000):
w = self.weight
t, past_data, b_limit, s_limit = self.one_market_model(w)
while t < t_max:
t, past_data, b_limit, s_limit = self.one_market_model(w, past_data, b_limit, s_limit)
return past_data
def two_market_model(self, w, T_A=0.0, T_B=0.0, past_data=None, b_1=None, s_1=None, b_2=None, s_2=None):
w = self.weight
if b_1 == None and s_1==None and b_2==None and s_2==None:
t1, past_data1, b_limit1, s_limit1 = self.one_market_model(w, past_data=None, b_limit=None, s_limit=None)
t2, past_data2, b_limit2, s_limit2 = self.one_market_model(w, past_data=None, b_limit=None, s_limit=None)
else:
b_limit1, s_limit1, b_limit2, s_limit2 = b_1, s_1, b_2, s_2
num_player = self.num_player
sigma = self.sigma
P_sigma = self.P_sigma
P_f = self.fdmtl
if past_data is None:
past_data = [P_f]
P_t_1 = past_data[-1]
w_1 = w[0]
w_2 = w[1]
w_3 = w[2]
ganma = self.ganma
r_t_e = np.zeros(num_player)
if len(past_data) < ganma:
r_t_h = np.log10(P_t_1/np.random.choice(past_data))
else:
past_data_ganma = past_data[-ganma:]
r_t_h = np.log10(P_t_1/np.random.choice(past_data))
for i in range(num_player):
for limit in [b_limit1, b_limit2]:
if limit == []:
limit.append(0)
for limit in [s_limit1, s_limit2]:
if limit == []:
limit.append(20000)
b_A, s_A = max(b_limit1), min(s_limit1)
b_B, s_B = max(b_limit2), min(s_limit2)
if T_A+T_B == 0:
W_A = 0.5
else:
W_A = T_A/(T_A+T_B)
W_B = 1.0 - W_A
e_t = np.random.normal(0, sigma)
r_t_e[i] = (w_1[i]*np.log10(P_f/P_t_1) + w_2[i]*r_t_h + w_3[i]*e_t)/(w_1[i] + w_2[i] + w_3[i])
P_e = P_t_1*math.exp(r_t_e[i])
P_o = np.random.normal(P_e, P_sigma)
if P_e > P_o:
if s_A < P_o or s_B < P_o:
if s_A < s_B:
P_t = s_A
s_limit1.remove(s_A)
T_A += 1.0
elif s_A == s_B:
market = np.random.choice(['A', 'B'], [W_A, W_B])
if market == 'A':
P_t = s_A
s_limit1.remove(s_A)
T_A += 1.0
else:
P_t = s_B
s_limit2.remove(s_B)
T_B += 1.0
else:
P_t = s_B
s_limit2.remove(s_B)
T_B += 1.0
else:
market = np.random.choice(['A', 'B'], [W_A, W_B])
if market == 'A':
P_t = P_t_1
b_limit1.append(P_o)
else:
P_t = P_t_1
b_limit2.append(P_o)
else:
if b_A > P_o or b_B > P_o:
if b_A > b_B:
P_t = b_A
b_limit1.remove(b_A)
T_A += 1.0
elif s_A == s_B:
market = np.random.choice(['A', 'B'], [W_A, W_B])
if market == 'A':
P_t = b_A
b_limit1.remove(b_A)
T_A += 1.0
else:
P_t = b_B
b_limit2.remove(b_B)
T_B += 1.0
else:
P_t = b_B
b_limit2.remove(b_B)
T_B += 1.0
else:
market = np.random.choice(['A', 'B'], [W_A, W_B])
if market == 'A':
P_t = P_t_1
s_limit1.append(P_o)
else:
P_t = P_t_1
s_limit2.append(P_o)
past_data.append(P_t)
t = len(past_data)
return t, T_A, T_B, past_data, b_limit1, s_limit1, b_limit2, s_limit2
def two_market_simulation(self, t_max=100000):
w = self.weight
t, T_A, T_B, past_data, b_1, s_1, b_2, s_2 = self.two_market_model(w)
share = []
while t < t_max:
t, T_A, T_B, past_data, b_1, s_1, b_2, s_2 = self.two_market_model(w, T_A, T_B, past_data, b_1, s_1, b_2, s_2)
share.append(T_A/(T_A+T_B))
return past_data, T_A, T_B, share
def plot_share(self, share):
share_B = []
for i in range(len(share)):
share_B.append(1.0-share[i])
plt.plot(share)
plt.plot(share_B)
In [105]:
AM = ArtificialMarket(ganma=10000)
In [106]:
w = AM.weight
t, past_data, b_limit, s_limit = AM.one_market_model(w)
In [107]:
plt.plot(past_data)
Out[107]:
In [142]:
past_data = AM.one_market_simulation(50000)
plt.plot(past_data)
Out[142]:
In [114]:
past_data2, T_A, T_B, share = AM.two_market_simulation(2)
In [115]:
plt.plot(past_data2)
Out[115]:
In [14]:
AM.plot_share(share)
In [31]:
past_data2[-1]
Out[31]:
In [ ]: