def BreedBoids(self, _mother, _father):
if _mother != _father:
strength = choice([_mother.m_strength, _father.m_strength])
boostA = choice([_mother.m_boost[0], _father.m_boost[0]])
boostB = choice([_mother.m_boost[1], _father.m_boost[1]])
tirednessRate = choice(
[_mother.m_tirednessRate, _father.m_tirednessRate])
recoveryRate = choice(
[_mother.m_recoveryRate, _father.m_recoveryRate])
awareness = choice([_mother.m_awareness, _father.m_awareness])
align = choice([_mother.m_alignment, _father.m_alignment])
sep = choice([_mother.m_separation, _father.m_separation])
cohes = choice([_mother.m_cohesion, _father.m_cohesion])
avoid = choice([_mother.m_avoidance, _father.m_avoidance])
rand = choice([_mother.m_random, _father.m_random])
food = choice([_mother.m_food, _father.m_food])
self.m_boids.append(
boid2.Boid2(_id=len(self.m_boids),
_str=strength,
_boost=np.array([boostA, boostB]),
_tired=tirednessRate,
_rec=recoveryRate,
_aware=awareness,
_align=align,
_cohes=cohes,
_sep=sep,
_avoid=avoid,
_random=rand,
_food=food,
_colour=np.array([1.0, 0.4, 6])))
def MutateBoids(self, _boid):
if bool(getrandbits(1)):
strength = _boid.m_strength + uniform(-0.2, 0.2)
else:
strength = _boid.m_strength
if bool(getrandbits(1)):
boost = _boid.m_boost + np.random.uniform(-0.2, 0.2, 2)
else:
boost = _boid.m_boost
if bool(getrandbits(1)):
tired = _boid.m_tirednessRate + uniform(-0.02, 0.02)
else:
tired = _boid.m_tirednessRate
if bool(getrandbits(1)):
recover = _boid.m_recoveryRate + uniform(-0.02, 0.02)
else:
recover = _boid.m_recoveryRate
if bool(getrandbits(1)):
align = _boid.m_alignment + uniform(-0.2, 0.2)
else:
align = _boid.m_alignment
if bool(getrandbits(1)):
sep = _boid.m_separation + uniform(-0.2, 0.2)
else:
sep = _boid.m_separation
if bool(getrandbits(1)):
cohes = _boid.m_cohesion + uniform(-0.2, 0.2)
else:
cohes = _boid.m_cohesion
if bool(getrandbits(1)):
avoid = _boid.m_avoidance + uniform(-0.2, 0.2)
else:
avoid = _boid.m_avoidance
if bool(getrandbits(1)):
rand = _boid.m_random + uniform(-0.2, 0.2)
else:
rand = _boid.m_random
if bool(getrandbits(1)):
food = _boid.m_food + uniform(-0.2, 0.2)
else:
food = _boid.m_food
if bool(getrandbits(1)):
aware = _boid.m_awareness + uniform(-0.02, 0.02)
else:
aware = _boid.m_awareness
self.m_boids.append(
boid2.Boid2(_id=len(self.m_boids),
_str=strength,
_boost=boost,
_tired=tired,
_rec=recover,
_aware=aware,
_align=align,
_cohes=cohes,
_sep=sep,
_avoid=avoid,
_random=rand,
_food=food,
_colour=np.array([0, 1, 0])))
def AddPredator(self, _i):
for x in range(0, _i):
self.m_predators.append(
boid2.Boid2(_id=x,
_predator=True,
_predAtt=self.predAtt,
_predSig=self.predSig,
_predSpeed=self.predSpeed))
def NextGeneration(self):
topTwenty = sorted(self.m_boids,
key=lambda x: x.m_fitness,
reverse=True)
topTwenty = topTwenty[:20]
print 'Top Performers of Generation #' + str(self.m_genCount) + ':'
for x in topTwenty:
print str(x.m_id) + ':' + str(x.m_fitness) + ': ' + x.m_name
# Update our generation counter
self.m_genCount += 1
# Reset ticks
self.ticks = 0
# If we're done, print out results
if self.m_genCount % 10 == 0:
print self.m_survivorCount
else:
# Clear population
self.m_boids = []
# Randomly breed 1/2 of the surviving population
for x in range(0, 10):
mother = choice(topTwenty)
father = choice(topTwenty)
self.BreedBoids(mother, father)
# Randomly mutate 1/4 of the surviving population
for x in range(0, 5):
self.MutateBoids(choice(topTwenty))
# Add survivors back into the population
for x in range(0, 20):
tBoid = topTwenty[x]
tBoid.m_colour = np.array([0.0, 0.0, 1.0])
tBoid.m_id = len(self.m_boids)
tBoid.m_fitness = 0
tBoid.m_ticksAlive = 0
self.m_boids.append(tBoid)
# If we're not at max population count, randomly generate more boids
if len(self.m_boids) < self.m_maxBoids:
diff = self.m_maxBoids - len(self.m_boids)
for i in range(self.m_maxBoids - diff, self.m_maxBoids):
self.m_boids.append((boid2.Boid2(_id=i)))
# Reset Predators
self.m_predators = []
# Re-Add Predators
self.AddPredator(self.m_maxPredators)
def AddBoid(self, _i):
for x in range(0, _i):
self.m_boids.append(
boid2.Boid2(_id=x,
_aware=random.uniform(1, self.startAware),
_str=random.uniform(0.05, self.startStr),
_rec=random.uniform(0, self.startRecov),
_tired=random.uniform(0.005, self.startTired),
_globAlign=self.boidAlignment,
_globCoh=self.boidCohesion,
_globSep=self.boidSeparation))
random.seed(self.m_boids[x].GetRandomSeed())
self.m_run.append(False)