def create_random_mapping():
# # Uncomment for non-selective start
# mapping = []
# taskIDs = [i.id for i in Simulation.btasks]
# for id in taskIDs:
# randX = np.random.randint(0, meshX)
# randY = np.random.randint(0, meshY)
# mapping.append((id, randX, randY))
# Uncomment to have selective start
yn = True
while yn:
taskIDs = [i.id for i in Simulation.btasks]
mapping = []
for id in taskIDs:
randX = np.random.randint(0, meshX)
randY = np.random.randint(0, meshY)
mapping.append((id, randX, randY))
try:
sim = Simulation.Simulation(meshX, meshY, 1, 1, mapping)
yn = False
except Exception:
print(mapping)
yn = True
print("Found sol")
return mapping
def main():
factorfi = 1
factorfc = 1
fc_passing = True
fi_passing = True
while fc_passing:
try:
sim = Simulation.Simulation(4, 3, factorfc, 1,
[(1, 2, 2), (2, 3, 0), (3, 3, 1), (4, 0, 1), (5, 3, 1), (6, 0, 2), (7, 1, 2),
(8, 3, 0), (9, 3, 1), (10, 3, 2), (11, 3, 2), (12, 2, 2), (13, 1, 2),
(14, 2, 2), (15, 2, 0), (16, 2, 2), (17, 0, 1), (18, 3, 2), (19, 0, 0),
(20, 0, 2), (21, 3, 2), (22, 1, 0), (23, 3, 2), (24, 2, 1), (25, 1, 0),
(26, 2, 2), (27, 0, 0), (28, 2, 0), (29, 2, 2), (30, 1, 2), (31, 0, 2),
(32, 3, 0), (33, 0, 2), (34, 3, 0), (35, 0, 1), (36, 1, 0), (37, 1, 2),
(38, 2, 1), (39, 0, 0), (40, 1, 2), (41, 1, 1), (42, 1, 0), (43, 1, 1),
(44, 1, 2)])
factorfc -= 0.01
except Exception:
factorfc += 0.01
fc_passing = False
while fi_passing:
try:
sim = Simulation.Simulation(4, 3, 1, factorfi,
[(1, 2, 2), (2, 3, 0), (3, 3, 1), (4, 0, 1), (5, 3, 1), (6, 0, 2), (7, 1, 2),
(8, 3, 0), (9, 3, 1), (10, 3, 2), (11, 3, 2), (12, 2, 2), (13, 1, 2),
(14, 2, 2), (15, 2, 0), (16, 2, 2), (17, 0, 1), (18, 3, 2), (19, 0, 0),
(20, 0, 2), (21, 3, 2), (22, 1, 0), (23, 3, 2), (24, 2, 1), (25, 1, 0),
(26, 2, 2), (27, 0, 0), (28, 2, 0), (29, 2, 2), (30, 1, 2), (31, 0, 2),
(32, 3, 0), (33, 0, 2), (34, 3, 0), (35, 0, 1), (36, 1, 0), (37, 1, 2),
(38, 2, 1), (39, 0, 0), (40, 1, 2), (41, 1, 1), (42, 1, 0), (43, 1, 1),
(44, 1, 2)])
factorfi -= 0.01
except Exception:
factorfi += 0.01
fi_passing = False
print(factorfc, factorfi)
from scipy.stats import expon, norm, erlang
from time import time
from math import factorial as fact
λ = 10
µ = 15
c = 1 # number of servers
max_time = 10000 #2000 / µ
debug_interval = max_time / 200
Ntr = 3 # number of simulations to run
# RUN SIMULATIONS
simulations = []
for i in range(Ntr):
print(f"Running simulation {i}")
sim = Simulation(max_time, λ, µ, c, debug_interval)
sim.run()
simulations.append(sim)
print(f'Finished simulating')
# 9: show average number of packets in system over time
ρ = λ / (c * µ)
pi_0 = 1 / (sum([(c * ρ)**k / fact(k)
for k in range(0, c)]) + (c * ρ)**c / (fact(c) * (1 - ρ)))
pi_c_plus = (c * ρ)**c / (fact(c) * (1 - ρ)) * pi_0
theor_avg_load = c * ρ + ρ / (1 - ρ) * pi_c_plus
plt.plot([0, max_time], [theor_avg_load] * 2, 'k-', linewidth=1)
for sim in simulations:
debug_times = [ds.event_time for ds in sim.debugStats]
cum_avg_loads = [ds.cum_avg_load for ds in sim.debugStats]
import sys
from lib import Simulation
from lib import Options
if __name__ == '__main__':
options = Options()
opts = options.parse(sys.argv[1:])
sim = Simulation(opts.cuboid_file, opts.steps_file)
sim.engage()
def setUp(self):
# configure self.attribute
self.configure_simulator()
self.sim = Simulation(self.cuboid_file, self.steps_file)
def gen_main():
graphx = []
graphy = []
bestmapping = []
bestmapping_fit = 0
try:
# Consts
pop_size = 500
mutation_rate = 0.01
meshX = 20
meshY = 20
runs = 100
# Generate population
def create_random_mapping():
# # Uncomment for non-selective start
# mapping = []
# taskIDs = [i.id for i in Simulation.btasks]
# for id in taskIDs:
# randX = np.random.randint(0, meshX)
# randY = np.random.randint(0, meshY)
# mapping.append((id, randX, randY))
# Uncomment to have selective start
yn = True
while yn:
taskIDs = [i.id for i in Simulation.btasks]
mapping = []
for id in taskIDs:
randX = np.random.randint(0, meshX)
randY = np.random.randint(0, meshY)
mapping.append((id, randX, randY))
try:
sim = Simulation.Simulation(meshX, meshY, 1, 1, mapping)
yn = False
except Exception:
print(mapping)
yn = True
print("Found sol")
return mapping
population_mappings = [create_random_mapping() for i in range(pop_size)]
for k in range(runs):
print('\nIteration {}'.format(k))
# Sim population collect costs and inf if fail
costs = []
for mapping in population_mappings:
try:
sim = Simulation.Simulation(meshX, meshY, 1, 1, mapping)
costs.append(sim.get_fitness())
except Exception:
costs.append(1000)
# costs.append(np.random.randint(0, 1000))
# Invert costs need to find range
inv_costs = np.asarray(costs)
inv_costs = np.abs(inv_costs - 1000)
print('Fitness = Max: {} Min: {} Ave: {} Fails: {}'.format(inv_costs.max(), inv_costs.min(),
inv_costs.mean(),
np.count_nonzero(inv_costs == 0)))
print('Best mapping is:\n{}'.format(population_mappings[np.argmax(inv_costs)]))
# Drawing pretty lines and collecting results
graphx.append(k)
graphy.append((inv_costs.max(), inv_costs.min(), inv_costs.mean()))
if inv_costs.max() > bestmapping_fit:
bestmapping_fit = inv_costs.max()
bestmapping = population_mappings[np.argmax(inv_costs)]
# ~~~~
# Breed population
if sum(inv_costs) != 0:
distribution = inv_costs / sum(inv_costs)
else:
distribution = np.asarray([1 / pop_size] * pop_size)
choices1 = np.random.choice(np.arange(pop_size), pop_size, p=distribution)
choices2 = np.random.choice(np.arange(pop_size), pop_size, p=distribution)
newpop = []
for i in range(pop_size):
mapping1 = population_mappings[choices1[i]]
mapping2 = population_mappings[choices2[i]]
newmapping = []
for (tsk1, x1, y1,), (tsk2, x2, y2) in zip(mapping1, mapping2):
if np.random.random() > mutation_rate:
# No mutation
if np.random.randint(1, 11, 1) > 5:
newmapping.append((tsk1, x1, y1))
else:
newmapping.append((tsk2, x2, y2))
else:
# Mutation
newmapping.append((tsk1, np.random.randint(0, meshX), np.random.randint(0, meshY)))
newpop.append(newmapping)
population_mappings = newpop
except:
pass
print('\n\n\n\n----------------------')
print(bestmapping_fit)
print(bestmapping)
plt.plot(graphx, graphy)
plt.show()