def car_reward(F, player):
"""
seeker's reward function
:param F: Node with status of all players
:param player: player number
:return:
"""
# print 'Player'
# print player
player_nr = int(player[3:])
# print 'Player Number'
# print player_nr
# reward = 0
# F = None
position_player = F[player_nr]
reward = position_player[1] * position_player[1]
reward *= 4
# if position_player[0] == west:
# reward = -position_player[1]
# if position_player[1] > int(north - blocked_north):
# reward -= 150
f_not_player = None
for f_b in range(len(F)):
# print f_b
if not f_b == player_nr:
if f_not_player is None:
f_not_player = F[f_b]
else:
f_not_player = np.vstack((f_not_player, F[f_b]))
# print position_player
# print np.array([east, north])
if np.array_equal(position_player, np.array([east, north])) is False:
if check_collision(position_player, f_not_player):
# print 'Collision'
reward -= 150
# sys.stdout.write('\n')
# sys.stdout.write(' p: ' + str(player_nr) + ' pos : ' + str(position_player[0]) +
# ' - ' + str(position_player[1]) + ' r: ' + str(reward))
# sys.stdout.write('\n')
return reward
def car_reward(F, player):
"""
seeker's reward function
:param F: Node with status of all players
:param player: player number
:return:
"""
# print 'Player'
# print player
player_nr = int(player[3:])
# print 'Player Number'
# print player_nr
# reward = 0
# F = None
position_player = F[player_nr]
reward = position_player[1] * position_player[1]
if position_player[0] == 0:
reward -= position_player[1]
if position_player[1] > 2:
reward -= 150
f_not_player = None
for f_b in range(len(F)):
if not f_b == player_nr:
if f_not_player is None:
f_not_player = F[f_b]
else:
f_not_player = np.vstack((f_not_player, F[f_b]))
if check_collision(position_player, f_not_player):
# print ' - Collision '
reward -= 20
# sys.stdout.write('\r')
# sys.stdout.write(' p: ' + str(player_nr) + ' pos : ' + str(position_player[0]) +
# ' - ' + str(position_player[1]) + ' r: ' + str(reward))
return reward
state_space.append(state)
# print 'state space: \n ' + str(state_space)
# starting locations
# Ego-vehicle starts at [0, 0]
starting_poses = np.array([west, south])
pose_ind = 1
while pose_ind < nr_vehicles:
new_pose = np.array([0, 0])
new_pose[0] = np.random.randint(east + 1, size=1)
new_pose[1] = np.random.randint(south + 3, size=1)
print 'new_pose' + str(new_pose)
if not check_collision(new_pose, starting_poses):
starting_poses = np.vstack((starting_poses, new_pose)) # axis=0))
pose_ind += 1
# print 'Pose index : ' + str(pose_ind)
# starting locations
print 'starting_poses : '
print starting_poses
# observational noise CPT (left, right, stay)
# obsnoiseCPT = np.array([prob_left, prob_right, prob_remain, prob_brake])
def adjust_loc(location):
""""
function that adjusts for moves off the grid