def update_goAround(obstacle):
if (self.goAround == 1):
self.x += sin(tan_inv(obstacle.y - self.y / obstacle.x - self.x))
self.y += cos(tan_inv(obstacle.y - self.y / obstacle.x - self.x))
if ((pow(abs(self.x - obstacle.x), 2) +
pow(abs(self.y - obstacle.y), 2)) >= 100):
self.goAround = 2
return
if (self.goAround == 2):
self.x += sin(
tan_inv(self.target_y - self.y / self.target_x - self.x))
self.y += cos(
tan_inv(self.target_y - self.y / self.target_x - self.x))
def update_target(self, alpha, explore, x_limit, y_limit, mean_x, mean_y):
temp_x = self.x * (1 - alpha) + alpha * explore * (
self.x + (random() * (x_limit / 2) -
(x_limit / 4))) + alpha * (1 - explore) * mean_x
temp_y = self.y * (1 - alpha) + alpha * explore * (
self.y + (random() * (y_limit / 2) -
(y_limit / 4))) + alpha * (1 - explore) * mean_y
# flag=0
# if(explore==1):
# if(self.x>x_limit-25 or self.x<-x_limit+25):
# temp_x=-self.x
# flag=1
# else:
# temp_x=self.x
# if(self.y>y_limit-25 or self.y<-y_limit+25):
# temp_y=-self.y
# flag=1
# else:
# temp_y=self.y
# if(flag==1):
# self.angle=tan_inv(temp_y,temp_x)
# return
# temp_x=self.x*(1-alpha)+alpha*explore*(self.x+(random()*(x_limit/2)-(x_limit/4)))+alpha*(1-explore)*mean_x
# temp_y=self.y*(1-alpha)+alpha*explore*(self.y+(random()*(y_limit/2)-(y_limit/4)))+alpha*(1-explore)*mean_y
if (temp_x > x_limit or temp_x < -x_limit):
while (not (temp_x > x_limit or temp_x < -x_limit)):
temp_x = self.x + (random() * (x_limit / 2) - (x_limit / 4))
if (temp_y > y_limit or temp_y < -y_limit):
while (not (temp_y > y_limit or temp_y < -y_limit)):
temp_y = self.y + (random() * (x_limit / 2) - (x_limit / 4))
self.target_x = temp_x
self.target_y = temp_y
self.angle = tan_inv((temp_y - self.y), (temp_x - self.x))
def update_target_obst(self, x_limit, y_limit, mean_x, mean_y):
temp_x = mean_x
temp_y = mean_y
# if(temp_x>x_limit or temp_x<-x_limit):
# temp_x=self.x+(random()*(x_limit/2)-(x_limit/4))
# if(temp_y>y_limit or temp_y<-y_limit):
# temp_y=self.y+(random()*(x_limit/2)-(x_limit/4))
# temp_x=0.7*temp_x + 0.3*(self.x+(random()*(x_limit/2)-(x_limit/4)))
# temp_y=0.7*temp_x + 0.3*(self.y+(random()*(y_limit/2)-(y_limit/4)))
self.target_x = temp_x
self.target_y = temp_y
self.angle = tan_inv((temp_y - self.y), (temp_x - self.x))
def update_target(self, alpha, explore, x_limit, y_limit, mean_x, mean_y):
if self.sleep:
self.sleep = False
return
temp_x = self.x * (1 - alpha) + alpha * explore * (
self.x + (random() * (x_limit / 2) -
(x_limit / 4))) + alpha * (1 - explore) * mean_x
temp_y = self.y * (1 - alpha) + alpha * explore * (
self.y + (random() * (y_limit / 2) -
(y_limit / 4))) + alpha * (1 - explore) * mean_y
if (temp_x > x_limit or temp_x < -x_limit):
while (not (temp_x > x_limit or temp_x < -x_limit)):
temp_x = self.x + (random() * (x_limit / 2) - (x_limit / 4))
if (temp_y > y_limit or temp_y < -y_limit):
while (not (temp_y > y_limit or temp_y < -y_limit)):
temp_y = self.y + (random() * (x_limit / 2) - (x_limit / 4))
self.target_x = temp_x
self.target_y = temp_y
self.angle = tan_inv((temp_y - self.y), (temp_x - self.x))
def main_obstacle_1(no_targets, no_observers, no_obstacles, targets, observers,
obstacles):
step = 0
data_until_update = []
total_count_obst = 0
while (step <= total_steps):
if (step % update_steps == 0):
[observer_target_dict, observer_obstacle_dict
] = update_for_observers(observers, targets, obstacles)
[target_observer_dict, target_obstacle_dict
] = update_for_targets(observers, targets, obstacles)
for i in observer_target_dict:
temp_arr_x = []
temp_arr_y = []
for j in observer_target_dict[i]:
temp_arr_x.append(targets[j].x)
temp_arr_y.append(targets[j].y)
if (len(temp_arr_x) and len(temp_arr_y)):
mean_x = mean(temp_arr_x)
mean_y = mean(temp_arr_y)
explore = pow(1 / (len(observer_target_dict[i]) + 1), 2)
rwrd = reward_obs_1(observers[i], targets,
observer_target_dict[i], obstacles,
observer_obstacle_dict[i], x_limit,
y_limit, explore, mean_x, mean_y)
E_min = LP_CTO(rwrd, 1.0,
template_probability_distribution)[0]
alpha = BRLP_CTO(rwrd, template_probability_distribution,
E_min)
observers[i].update_target(alpha, explore, x_limit,
y_limit, mean_x, mean_y)
else:
observers[i].update_target(1, 1, x_limit, y_limit, 0, 0)
for i in target_observer_dict:
temp_arr_x = []
temp_arr_y = []
for j in target_observer_dict[i]:
temp_arr_x.append(observers[j].x)
temp_arr_y.append(observers[j].y)
target_obstacle_dict[i].sort(reverse=True,
key=lambda i: obstacles[i][0])
if (len(temp_arr_x) and len(temp_arr_y)):
mean_x = mean(temp_arr_x)
mean_y = mean(temp_arr_y)
obst_idx = -1
tan_inv_ot = tan_inv((targets[i].y - mean_y),
(targets[i].x - mean_x))
for j in target_obstacle_dict[i]:
tan_inv_oo = tan_inv((obstacles[j][1][0].y - mean_y),
(obstacles[j][1][0].x - mean_x))
if (abs(tan_inv_oo - tan_inv_ot) <= math.pi / 2):
obst_idx = j
break
if (obst_idx == -1):
targets[i].update_target(x_limit, y_limit, mean_x,
mean_y)
else:
sx = 0
sy = 0
for j in obstacles[obst_idx][1]:
sx += j.x
sy += j.y
n = obstacles[obst_idx][0]
targets[i].update_target_obst(x_limit, y_limit, sx / n,
sy / n)
for i in observers:
i.update(x_limit, y_limit)
for i in targets:
i.update(x_limit, y_limit)
step += 1
ans_dict = update_for_observers1(observers, targets)
for i in ans_dict:
for j in ans_dict[i]:
total_count_obst += 1
return total_count_obst