def run_experiment_decaying_epsilon(m1, m2, m3, N):
bandits = [Bandit(m1), Bandit(m2), Bandit(m3)]
data = np.empty(N)
for i in range(N):
p = np.random.random()
if p < 1.0 / (i + 1):
j = np.random.choice(3)
else:
j = np.argmax([b.mean for b in bandits])
x = bandits[j].pull()
bandits[j].update(x)
data[i] = x
cumulative_average = np.cumsum(data) / (np.arange(N) + 1)
#plot moving average
plt.plot(cumulative_average)
plt.plot(np.ones(N) * m1)
plt.plot(np.ones(N) * m2)
plt.plot(np.ones(N) * m3)
plt.xscale('log')
plt.show()
return cumulative_average
def run_experiment_decaying_eps(m1, m2, m3, N):
bandits = [Bandit(m1), Bandit(m2), Bandit(m3)]
data = np.empty(N)
for i in range(N):
# Epsilon greedy
p = np.random.random()
if p < 1. / (i + 1):
selection = np.random.choice(3)
else:
selection = np.argmax([bandit.mean for bandit in bandits])
# Pull from bandit and update sample mean
x = bandits[selection].pull()
bandits[selection].update(x)
data[i] = x
cumulative_avg = np.cumsum(data) / (np.arange(N) + 1)
plt.plot(cumulative_avg)
plt.plot(np.ones(N) * m1)
plt.plot(np.ones(N) * m2)
plt.plot(np.ones(N) * m3)
plt.xscale('log')
plt.show()
# Print our esimate of each bandits mean and their actual mean
print('Estimate of mean Actual mean')
for bandit in bandits:
print('{:<20}{}'.format(bandit.mean, bandit.m))
return cumulative_avg
def run_experiment_decaying_epsilon(m1, m2, m3, N): bandits = [Bandit(m1), Bandit(m2), Bandit(m3)] data = np.empty(N)