def get_reward(self, reset_reason):
pre_potential = helper.dipole_potential(self.pre_ball[X],
self.pre_ball[Y], 2.1, 7)
cur_potential = helper.dipole_potential(self.cur_ball[X],
self.cur_ball[Y], 2.1, 7)
reward = cur_potential - pre_potential
# Reset
if (reset_reason != NONE):
reward = 0
reward -= helper.distance2(self.cur_ball[X], self.cur_my_posture[0][X],
self.cur_ball[Y], self.cur_my_posture[0][Y])
if self.cur_my_posture[0][TOUCH]:
reward += 500
if (reset_reason == SCORE_MYTEAM):
self.score_sum += 1
reward += 1000 # minimum 25
self.printConsole("my team goal")
if (reset_reason == SCORE_OPPONENT):
self.score_sum -= 1
reward -= 1000 # maxmimum -25
self.printConsole("op team goal")
self.printConsole("reward: " + str(reward))
self.pre_ball = self.cur_ball
return reward
def get_reward(self, reset_reason, i):
pre_potential = helper.dipole_potential(self.pre_ball[X],
self.pre_ball[Y], 2.1, 7)
cur_potential = helper.dipole_potential(self.cur_ball[X],
self.cur_ball[Y], 2.1, 7)
reward = cur_potential - pre_potential
# Reset
if (reset_reason != NONE):
reward = 0
reward -= 0.1 * helper.distance(
self.cur_ball[X], self.cur_my_posture[i][X], self.cur_ball[Y],
self.cur_my_posture[i][Y])
for k in range(self.number_of_robots):
if self.cur_my_posture[k][TOUCH]:
reward += 30
if (reset_reason == SCORE_MYTEAM):
self.score_sum += 1
reward += 200 # minimum 25
self.printConsole("my team goal")
if (reset_reason == SCORE_OPPONENT):
self.score_sum -= 1
reward -= 200 # maxmimum -25
self.printConsole("op team goal")
# self.printConsole("reward: " + str(reward))
return reward
def get_reward(self, reset_reason, i):
pre_potential = helper.dipole_potential(self.pre_ball[X],
self.pre_ball[Y], 2.1, 30)
cur_potential = helper.dipole_potential(self.cur_ball[X],
self.cur_ball[Y], 2.1, 30)
potential_rew = cur_potential - pre_potential
# Reset
if (reset_reason != NONE):
potential_rew = 0
self.printConsole(' potential reward ' + str(i) + ': ' +
str(potential_rew))
dist_rew = -0.1 * helper.distance(
self.cur_ball[X], self.cur_my_posture[i][X], self.cur_ball[Y],
self.cur_my_posture[i][Y])
self.printConsole(' distance reward ' + str(i) + ': ' +
str(dist_rew))
touch_rew = 0
touch_case = 0
if (abs(self.cur_ball[X]) < 1.85) and (abs(self.cur_ball[Y]) < 1.3):
if self.cur_my_posture[i][TOUCH]:
touch_rew += 20
touch_case = 1
elif (self.cur_ball[X] > 0) and (abs(self.cur_ball[Y]) < 0.3):
if self.cur_my_posture[i][TOUCH]:
touch_rew += 20
touch_case = 2
elif (self.cur_ball[X] < 0) and (abs(self.cur_ball[Y]) < 0.3):
pass
else:
if self.cur_my_posture[i][TOUCH]:
touch_rew -= 0.01
touch_case = 3
self.printConsole(' touch reward ' + str(i) + ': ' +
str(touch_rew) + ' (case: ' + str(touch_case) + ')')
goal_rew = 0
if (reset_reason == SCORE_MYTEAM):
self.score_sum += 1
goal_rew += 500 # minimum 25
self.printConsole("my team goal")
if (reset_reason == SCORE_OPPONENT):
self.score_sum -= 1
goal_rew -= 500 # maxmimum -25
self.printConsole("op team goal")
rew = potential_rew + dist_rew + touch_rew + goal_rew
self.printConsole(' reward ' + str(i) + ': ' +
str(rew))
return rew
def get_reward(self, reset_reason):
pre_potential = helper.dipole_potential(self.pre_ball[X],
self.pre_ball[Y], 2.1, 3)
cur_potential = helper.dipole_potential(self.cur_ball[X],
self.cur_ball[Y], 2.1, 3)
reward = cur_potential - pre_potential
# Add dead lock penalty
if (reset_reason == SCORE_MYTEAM):
self.score_sum += 1
reward += 10
self.printConsole("my team goal")
if (reset_reason == SCORE_OPPONENT):
self.score_sum -= 1
reward -= 10
self.printConsole("op team goal")
self.printConsole("reward: " + str(reward))
self.pre_ball = self.cur_ball
return reward
def get_reward(self, reset_reason):
pre_potential = helper.dipole_potential(self.pre_ball[X],
self.pre_ball[Y], 2.1, 5)
cur_potential = helper.dipole_potential(self.cur_ball[X],
self.cur_ball[Y], 2.1, 5)
reward = cur_potential - pre_potential
# if my robot become deactive, recieve penalty
for i in range(self.number_of_robots):
if (self.cur_my_posture[i][ACTIVE]
== False) and (self.active_flag[MY_TEAM][i] is True):
reward -= 1
self.printConsole("my team deactive")
self.active_flag[MY_TEAM][i] = self.cur_my_posture[i][ACTIVE]
# if opponent robot become deactive, recieve penalty
for i in range(self.number_of_robots):
if (self.cur_op_posture[i][ACTIVE]
== False) and (self.active_flag[OP_TEAM][i] is True):
reward += 1
self.printConsole("opponent team deactive")
self.active_flag[OP_TEAM][i] = self.cur_op_posture[i][ACTIVE]
if (reset_reason == SCORE_MYTEAM):
self.score_sum += 1
reward += 50 # minimum 25
self.printConsole("my team goal")
if (reset_reason == SCORE_OPPONENT):
self.score_sum -= 1
reward -= 50 # maxmimum -25
self.printConsole("op team goal")
self.printConsole("reward: " + str(reward))
self.pre_ball = self.cur_ball
return reward
def get_reward(self, reset_reason):
# pre_potential = helper.dipole_potential(self.pre_ball[X], self.pre_ball[Y], 2.1, 3)
# cur_potential = helper.dipole_potential(self.cur_ball[X], self.cur_ball[Y], 2.1, 3)
reward = helper.dipole_potential(self.cur_my_posture[4][X], self.cur_my_posture[4][Y], 2.1, 0.3)
# Add dead lock penalty
# if(reset_reason == SCORE_MYTEAM):
# self.score_sum += 1
# reward += 24 # minimum 24
# self.printConsole("my team goal")
# if(reset_reason == SCORE_OPPONENT):
# self.score_sum -= 1
# reward -= 24 # maxmimum -24
# self.printConsole("op team goal")
self.printConsole("reward: " + str(reward))
self.pre_ball = self.cur_ball
return reward
