def gen_agents(size,
d_input,
r_input,
c_input,
num_pre_dem=False,
peak_loc=False):
drones = []
if type(d_input) == int:
for d in range(d_input):
drones.append(
m.drone(
[random.randint(0, size[0]),
random.randint(0, size[1])], d))
else:
for idx, d in enumerate(d_input):
drones.append(m.drone([d[0], d[1]], idx))
restaurants = []
if type(r_input) == int:
for r in range(r_input):
restaurants.append(
m.restaurant(
[random.randint(0, size[0]),
random.randint(0, size[1])]))
else:
for r in r_input:
restaurants.append(m.restaurant(r))
chargers = []
if not c_input:
c_loc = [[size[0] / 4, size[1] / 2], [size[0] * 3 / 4, size[1] / 2]]
for c in c_loc:
chargers.append(m.charge([c[0], c[1]]))
elif type(c_input) == int:
for c in range(c_input):
chargers.append(
m.charge(
[random.randint(0, size[0]),
random.randint(0, size[1])]))
else:
for c in c_input:
chargers.append(
m.charge(
[random.randint(0, size[0]),
random.randint(0, size[1])]))
jobs_na = []
if num_pre_dem:
for i in range(num_pre_dem):
demand = dem.mapGenerator(size, False, peak_loc)
rest, cust_to_charge = m.pick_restaurant(demand, restaurants,
chargers)
job = m.job(demand, rest, 0, random.randint(10, 25),
cust_to_charge)
jobs_na.append(job)
return drones, restaurants, chargers, jobs_na
def main_loop(drones,
restaurants,
chargers,
jobs_a,
jobs_na,
jobs_c,
t,
spawn_prob,
size,
peak_loc,
learning_weights=False):
demand = dem.mapGenerator(size, spawn_prob, peak_loc)
slow = False
if demand:
rest, cust_to_charge = m.pick_restaurant(demand, restaurants, chargers)
job = m.job(demand, rest, t, random.randint(10, 25), cust_to_charge)
jobs_na.append(job)
for d in drones:
d.move_to_goal(t, chargers)
if not learning_weights:
no_job_drone = [
x for x in drones
if x.job == 0 or x.job.state == 4 and x.battery >= 90
]
else:
no_job_drone = [x for x in drones if x.job == 0 or x.job.state == 4]
if learning_weights:
s = time.time()
score_matrix = m.NNscores(no_job_drone, jobs_na, t, learning_weights,
len(drones))
if time.time() - s > 0.008:
slow = True
else:
score_matrix = m.simpleScores(no_job_drone, jobs_na, t)
idx_drones, idx_jobs_na = m.hungarianMethod(score_matrix)
for idx, drone in enumerate(idx_drones):
if int(score_matrix[drone, idx_jobs_na[idx]]) != 100000:
no_job_drone[drone].set_job(jobs_na[idx_jobs_na[idx]])
jobs_a.append(jobs_na[idx_jobs_na[idx]])
# added for KPI
for job in jobs_a:
if job.state == 3:
jobs_c.append(job)
jobs_na = [x for x in jobs_na if x.state == 0]
jobs_a = [x for x in jobs_a if x.state != 3]
return jobs_na, jobs_a, jobs_c, slow
-0.0005058854834270043, 0.98813616677188, 1.124796722093522,
-0.32571743636542916, -0.17580636537189864,
-0.010327358384964347, 0.9861115479780318, 0.5705700709535082,
-0.9390895310960459, -0.6529715741626314
]]]
weights = [np.array(x) for x in weights]
#################################################################################
#### Restaurant Locations
#################################################################################
np.random.seed(1)
rest_loc = []
for i in range(20):
rest_loc.append(
dem.mapGenerator(gridSize=(1000, 1000),
spawn_prob=False,
peakCoordinates=False))
np.random.seed(None)
# resttest = np.random.rand(3)
#################################################################################
#### Simulation Function
#################################################################################
def run_sim(runs, ticks, num_drones, num_restaurants, demandRate, rest_loc,
learning_weights):
size = (1000, 1000)
rest_loc = rest_loc[0:num_restaurants]