def solve(self):
try:
# We need to transofrm our GUI entries to variables !
po00 = (int(self.payoffEntryL00.get()),
int(self.payoffEntryR00.get()))
po01 = (int(self.payoffEntryL01.get()),
int(self.payoffEntryR01.get()))
po10 = (int(self.payoffEntryL10.get()),
int(self.payoffEntryR10.get()))
po11 = (int(self.payoffEntryL11.get()),
int(self.payoffEntryR11.get()))
payoff = [[po00, po01], [po10, po11]]
M = np.array(payoff)
s = Solver(payoff=M)
result = s.solve(1, 0)
strategies = []
for r in result:
if (r[0] == 0 and r[1] == 0):
strategies.append("(A,C)")
elif (r[0] == 0 and r[1] == 1):
strategies.append("(A,D)")
elif (r[0] == 1 and r[1] == 0):
strategies.append("(B,C)")
else:
strategies.append("(B,D)")
ne = " , ".join(strategies)
ne_msg = f'Pure Nash Equiliberia Strategies :\n{ne}\n'
print(ne_msg)
self.resultLabel.config(text=ne_msg)
po00 = (self.payoffFrame00, self.payoffLP00, self.payoffV00,
self.payoffRP00, self.payoffEntryL00, self.payoffEntryR00)
po01 = (self.payoffFrame01, self.payoffLP01, self.payoffV01,
self.payoffRP01, self.payoffEntryL01, self.payoffEntryR01)
po10 = (self.payoffFrame10, self.payoffLP10, self.payoffV10,
self.payoffRP10, self.payoffEntryL10, self.payoffEntryR10)
po11 = (self.payoffFrame11, self.payoffLP11, self.payoffV11,
self.payoffRP11, self.payoffEntryL11, self.payoffEntryR11)
show_ne = [po00, po01, po10, po11]
print(result)
res2Show = []
for r in result:
bin_num = str(r[0]) + str(r[1])
idx = int(bin_num, 2)
res2Show.append(idx)
for i in res2Show:
self.showNE(show_ne[i][0], show_ne[i][1], show_ne[i][2],
show_ne[i][3], show_ne[i][4], show_ne[i][5])
except ValueError as verr:
showerror(title=" Invalid PayOff",
message=" All entries need to be Integers!")
return
def solve(self):
try:
# We need to transofrm our GUI entries to variables !
po00=(int(self.payoffEntryL00.get()),int(self.payoffEntryR00.get()))
po01=(int(self.payoffEntryL01.get()),int(self.payoffEntryR01.get()))
po10=(int(self.payoffEntryL10.get()),int(self.payoffEntryR10.get()))
po11=(int(self.payoffEntryL11.get()),int(self.payoffEntryR11.get()))
payoff=[[po00,po01],[po10,po11]]
M=np.array(payoff)
s = Solver(payoff=M)
result=s.solve(0, 1)
problem=0
if(result[0] is None or result[1] is None):
if (result[0] is not None):
msg = f'Nash Equiliberia Mixed Strategies : \np={Fraction(result[0]).limit_denominator()}\nq∈[0,1]'
elif (result[1] is not None ):
msg = f'Nash Equiliberia Mixed Strategies : \np∈[0,1]\nq={Fraction(result[1]).limit_denominator()}'
else:
msg = f'Nash Equiliberia Mixed Strategies : \np∈[0,1]\nq∈[0,1]'
else:
if (result[0] > 1 or result[1] > 1):
msg = f'No Nash Equilebria found in Mixed Strategies'
problem=1
elif (result[0] < 0):
if (result[1] < 0):
msg = f'Nash Equiliberia Mixed Strategies : \np∈[0,1]\nq∈[0,1]'
else:
msg = f'Nash Equiliberia Mixed Strategies : \np∈[0,1]\nq={Fraction(result[1]).limit_denominator()}'
elif (result[1] < 0):
if (result[0] < 0):
msg = f'Nash Equiliberia Mixed Strategies : \np∈[0,1]\nq∈[0,1]'
else:
msg = f'Nash Equiliberia Mixed Strategies : \np={Fraction(result[0]).limit_denominator()}\nq∈[0,1]'
else:
msg=f'Nash Equiliberia Mixed Strategies : \np = {result[0]} \nq = {result[1]}'
# if(result[0]>1 and result[1]>1):
# msg = f'No Nash Equilebria found in Mixed Strategies'
# elif(result[0]<0 or result[0]>1):
# msg = f'Nash Equiliberia Mixed Strategies: \np∈[0,1] \nq = 0'
# elif(result[1]<0 or result[1]>1):
# msg = f'Nash Equiliberia Mixed Strategies : \np=0 \nq∈[0,1]'
# else:
# msg=f'Nash Equiliberia Mixed Strategies : \np = {result[0]} \nq = {result[1]}'
# messagebox.showinfo(title="Mixed Strategies Nash Equilibria", message=msg)
self.resultLabel.configure(text=msg)
if(not problem):
self.PQgraph(result[0],result[1])
plt.show()
except ValueError as verr:
showerror(title=" Invalid PayOff", message=" All entries need to be Integers!")
return
alphas = [[1]] #np.linspace(0.6, 0.6, 1) # mixing coefficients
colors = ['red', 'pink', 'brown']
for mu_iter in mus:
mu = np.array([[55]]) #np.array([[-mu_iter], [mu_iter]])
errorss_daem = []
alphass_daem = []
muss_daem = []
errorss_em = []
alphass_em = []
muss_em = []
bs = []
for alpha in alphas:
solver = Solver(alpha=np.array(alpha), mu=mu, sigma=sigma)
# Xi is 1 dimensional. N data points
# need to generate data properly as required
def getsample(k):
return np.random.normal(mu[k], sigma[k]**0.5)
X = np.array([getsample(toss(alpha)) for j in range(N)]).reshape(1, N)
print('Actual:')
print('alpha: {}\nmu:\n{}\nsigma:\n{}\n\n'.format(alpha, mu, sigma))
alpha_est_daem, mu_est_daem, sigma_est_daem, errors_daem, steps, beta_step, likelihoods_daem, actual_likelihood_daem = solver.DAEM_GMM(
X=X, thresh=1e-6, K=K)
errorss_daem.append(errors_daem)
alphass_daem.append(alpha_est_daem)
muss_daem.append(mu_est_daem)
mu = np.zeros((K, 1, 1))
sigma = np.ones((K, 1, 1))
errorss_daem = []
alphass_daem = []
muss_daem = []
errorss_em = []
alphass_em = []
muss_em = []
bs = []
alpha = [1 / K for i in range(K)]
alphas = [alpha]
solver = Solver(alpha=np.array(alpha), mu=mu, sigma=sigma)
# Data generation
X = cdf_inv(np.random.rand(N)).reshape((1, N))
# print('Actual:')
# print('alpha: {}\nmu:\n{}\nsigma:\n{}\n\n'.format(alpha, mu, sigma))
#alpha_est_daem, mu_est_daem, sigma_est_daem, errors_daem, steps, beta_step, likelihoods_daem, actual_likelihood_daem = solver.DAEM_GMM(X=X, thresh=1e-6, K=K, max_steps=5000)
K_daem, res = solver.fit_data(X=X,
thresh=1e-4,
min_thresh=1e-6,
max_steps=5000,
algo='DAEM',
Kmin=4,
Kmax=8)
alpha_est_daem, mu_est_daem, sigma_est_daem, er, steps, beta_step, likelihoods_daem, al = res
def solve(self):
try:
# We need to transofrm our GUI entries to variables !
po00 = (int(self.payoffEntryL00.get()),
int(self.payoffEntryR00.get()))
po01 = (int(self.payoffEntryL01.get()),
int(self.payoffEntryR01.get()))
po10 = (int(self.payoffEntryL10.get()),
int(self.payoffEntryR10.get()))
po11 = (int(self.payoffEntryL11.get()),
int(self.payoffEntryR11.get()))
payoff = [[po00, po01], [po10, po11]]
M = np.array(payoff)
s = Solver(payoff=M)
result = s.solve(1, 1)
pure, mixed = result[0], result[1]
strategies = []
for r in pure:
if (r[0] == 0 and r[1] == 0):
strategies.append("(A,C)")
elif (r[0] == 0 and r[1] == 1):
strategies.append("(A,D)")
elif (r[0] == 1 and r[1] == 0):
strategies.append("(B,C)")
else:
strategies.append("(B,D)")
ne = " , ".join(strategies)
ne_msg = f'Nash Equiliberia Pure Strategies :\n{ne}\n'
problem = 0
if (mixed[0] is None or mixed[1] is None):
if (mixed[0] is not None):
msg = f'Nash Equiliberia Mixed Strategies : \np={Fraction(mixed[0]).limit_denominator()}\nq∈[0,1]'
elif (mixed[1] is not None):
msg = f'Nash Equiliberia Mixed Strategies : \np∈[0,1]\nq={Fraction(mixed[1]).limit_denominator()}'
else:
msg = f'Nash Equiliberia Mixed Strategies : \np∈[0,1]\nq∈[0,1]'
else:
if (mixed[0] > 1 or mixed[1] > 1):
msg = f'No Nash Equilebria found in Mixed Strategies'
problem = 1
elif (mixed[0] < 0):
if (mixed[1] < 0):
msg = f'Nash Equiliberia Mixed Strategies : \np∈[0,1]\nq∈[0,1]'
else:
msg = f'Nash Equiliberia Mixed Strategies : \np∈[0,1]\nq={Fraction(mixed[1]).limit_denominator()}'
elif (mixed[1] < 0):
if (mixed[0] < 0):
msg = f'Nash Equiliberia Mixed Strategies : \np∈[0,1]\nq∈[0,1]'
else:
msg = f'Nash Equiliberia Mixed Strategies : \np={Fraction(mixed[0]).limit_denominator()}\nq∈[0,1]'
else:
msg = f'Nash Equiliberia Mixed Strategies : \np = {mixed[0]} \nq = {mixed[1]}'
print(ne_msg + msg)
self.resultLabel.config(text=ne_msg + msg)
# showinfo(title="Mixed Strategies Nash Equilibria", message=msg)
po00 = (self.payoffFrame00, self.payoffLP00, self.payoffV00,
self.payoffRP00, self.payoffEntryL00, self.payoffEntryR00)
po01 = (self.payoffFrame01, self.payoffLP01, self.payoffV01,
self.payoffRP01, self.payoffEntryL01, self.payoffEntryR01)
po10 = (self.payoffFrame10, self.payoffLP10, self.payoffV10,
self.payoffRP10, self.payoffEntryL10, self.payoffEntryR10)
po11 = (self.payoffFrame11, self.payoffLP11, self.payoffV11,
self.payoffRP11, self.payoffEntryL11, self.payoffEntryR11)
show_ne = [po00, po01, po10, po11]
res2Show = []
for r in pure:
bin_num = str(r[0]) + str(r[1])
idx = int(bin_num, 2)
res2Show.append(idx)
for i in res2Show:
self.showNE(show_ne[i][0], show_ne[i][1], show_ne[i][2],
show_ne[i][3], show_ne[i][4], show_ne[i][5])
if (not problem):
self.PQgraph(mixed[0], mixed[1])
plt.show()
except ValueError as verr:
showerror(title=" Invalid PayOff",
message=" All entries need to be Integers!")
return
colors = ['red', 'pink', 'brown']
color_dis = [
'green', 'blue', 'orange', 'yellow', 'cyan', 'magenta', 'indigo', 'red',
'pink', 'brown', 'black', 'violet', 'gray', 'maroon', 'blue', 'gold'
]
errorss_daem = []
alphass_daem = []
muss_daem = []
errorss_em = []
alphass_em = []
muss_em = []
bs = []
for alpha in alphas:
solver = Solver(alpha=np.array(alpha), mu=mu, sigma=sigma)
# Xi is 1 dimensional. N data points
# need to generate data properly as required
X = np.empty((n, 0))
X_classes = [np.empty((n, 0)) for k in range(K)]
decomps = [np.linalg.eig(sigma[k]) for k in range(K)]
for j in range(N):
k_chosen = toss(alpha)
my_mu = mu[k_chosen]
wsig, vsig = decomps[k_chosen]
sample = my_mu + vsig.dot((wsig**0.5).reshape(
(n, 1)) * np.random.randn(n, 1))
X_classes[k_chosen] = np.append(X_classes[k_chosen], sample, axis=1)
X = np.append(X, sample, axis=1)