reinforcement2

import numpy as np

R = np.matrix([ [-1,-1,-1,-1,0,-1],
[-1,-1,-1,0,-1,1],
[-1,-1,-1,0,-1,-1],
[-1,0,0,-1,0,-1],
[0,-1,-1,0,-1,1],
[-1,0,-1,-1,0,1] ])

Q = np.matrix(np.zeros([6,6]))

gamma = 0.8

initial_state = 1

def available_actions(state):
current_state_row = R[state,]
av_act = np.where(current_state_row >= 0)[1]
return av_act

available_act = available_actions(initial_state)
print(available_act)

def sample_next_action(available_actions_range):
next_action = int(np.random.choice(available_actions_range,1))
return next_action

action = sample_next_action(available_act)
print(action)

def update(current_state, action, gamma):

max_index = np.where(Q[action,] == np.max(Q[action,]))[1]

if max_index.shape[0] > 1:
max_index = int(np.random.choice(max_index, size = 1))
else:
max_index = int(max_index)
max_value = Q[action, max_index]

Q[current_state, action] = R[current_state, action] + gamma * max_value

update(initial_state,action,gamma)

for i in range(10000):
current_state = np.random.randint(0, int(Q.shape[0]))
available_act = available_actions(current_state)
action = sample_next_action(available_act)
update(current_state,action,gamma)

print("Trained Q matrix:")
print(Q/np.max(Q))

print("Testing")
current_state = 1
steps = [current_state]
while current_state != 5:
next_step_index = np.where(Q[current_state,] == np.max(Q[current_state,]))[1]

if next_step_index.shape[0] > 1:
next_step_index = int(np.random.choice(next_step_index, size = 1))
else:
next_step_index = int(next_step_index)

steps.append(next_step_index)
current_state = next_step_index

print("Selected path:")
print(steps)