def test_simulation():
w = World(4)
w.populate(
Individual(is_zombie = True),
2, 1
)
w.populate(
Individual(),
1, 2
)
w.simulate(moves = "DLUURR")
score, positions = w.get_stats()
# One victim
assert 1 == score
# Two zombies
assert 2 == len(positions)
# First at (3, 0)
assert 3 == positions[0]['x']
assert 0 == positions[0]['y']
# Second at (2, 1)
assert 2 == positions[1]['x']
assert 1 == positions[1]['y']
def populate(self,
individual: Individual,
x_coord: int = 0,
y_coord: int = 0) -> None:
""" Populate the world with an individual.
Args:
individual (Individual) : Individual of the world. Either a zombie or victim.
x_coord (int) : The position, x coordinate, of the individual
y_coord (int) : The position, y coordinate, of the individual
"""
# Save coordinates
individual.x_coord = x_coord
individual.y_coord = y_coord
# Add to population
self.population.append(individual)
# Get the population index of this individual
population_index = len(self.population) - 1
# Place the individual on the map
self.map[x_coord][y_coord] = population_index
# Add zombie to the queue to action
if (individual.is_zombie):
self.__enqueue(individual)
def test_moves():
individual = Individual(is_zombie=is_zombie,
x_coord=x,
y_coord=y,
moves=moves)
for m in moves:
assert m.lower() == individual.get_move()
def test_add_score():
individual = Individual(is_zombie=is_zombie,
x_coord=x,
y_coord=y,
moves=moves)
individual.add_score(10)
assert individual.score == 10
def ConstructPopulation(self):
for _ in range(self.m_PopulationCount):
individual = Individual(self.m_StopsCount, self.m_RoadLength,
self.m_MutationProbability,
self.m_MinStopsDistance,
self.m_MaxStopsDistance,
self.m_bannedAreas)
individual.CreateIndividual()
self.m_Population.append(individual)
def cabbage_growing(parent_1, parent_2):
"""
скрещивание 2х особей
:param parent_1: родитель1
:param parent_2: родитель2
:return: 2 потомка особей
"""
# создание экземляров потомков
son = Individual(GENE_LEN, gene_number)
daughter = Individual(GENE_LEN, gene_number)
if rnd.randint(0, 1): # псевдоравномерное скрещивание
daughter.ind_gene = "".join(parent_2)
son.ind_gene = "".join([
rnd.choices([parent_1, parent_2], [1, 1])[0][i]
for i in range(GENE_LEN)
])
else: # создание 1 - точечного кроссовера
border1 = rnd.randint(0, int(GENE_LEN // 2))
#
son.ind_gene = "".join(parent_1[:border1] + parent_2[border1:])
daughter.ind_gene = "".join(parent_2[:(GENE_LEN - border1)] +
parent_1[(GENE_LEN - border1):])
return son, daughter
def TournamentSelection(self):
usedStops = []
newPopulation = []
while len(usedStops) < self.m_PopulationCount:
firstChallanger = random.randint(0, self.m_PopulationCount - 1)
secondChallanger = random.randint(0, self.m_PopulationCount - 1)
if (firstChallanger not in usedStops
and secondChallanger not in usedStops):
usedStops.append(firstChallanger)
usedStops.append(secondChallanger)
if self.m_Population[
firstChallanger].m_Quality >= self.m_Population[
secondChallanger].m_Quality:
newPopulation.append(self.m_Population[firstChallanger])
else:
newPopulation.append(self.m_Population[secondChallanger])
self.m_Population = newPopulation
for i in range(int(self.m_PopulationCount / 2),
self.m_PopulationCount):
self.m_Population.append(
Individual(self.m_StopsCount, self.m_RoadLength,
self.m_MutationProbability, self.m_MinStopsDistance,
self.m_MaxStopsDistance, self.m_bannedAreas))
self.m_Population[i].CreateIndividual()
self.m_Population[i].m_Quality = self.GoalFunction(
self.m_Population[i])
x = lambda a: a.m_Quality
self.m_Population.sort(key=x)
if self.m_Population[0].m_Quality < self.m_AlfaMale.m_Quality:
self.m_AlfaMale.m_Chromosome = self.m_Population[0].m_Chromosome[:]
self.m_AlfaMale.m_Quality = self.m_Population[0].m_Quality
def __init__(self,
halfPopulationCount,
roadLength,
stopsCount,
popularPlaces,
mutationProbability=0.05,
mutationMethod='mr',
iterationCount=10000,
minStopsDistance=2,
maxStopsDistance=8,
bannedAreas=[]):
self.m_PopulationCount = 2 * halfPopulationCount
self.m_RoadLength = roadLength
self.m_StopsCount = stopsCount
self.m_PopularPlaces = popularPlaces
self.m_Population = []
self.m_IterationCount = iterationCount
self.m_MutationProbability = mutationProbability
self.m_MutationMethod = mutationMethod
self.m_MinStopsDistance = minStopsDistance
self.m_MaxStopsDistance = maxStopsDistance
self.m_bannedAreas = bannedAreas
self.m_AlfaMale = Individual(self.m_StopsCount, self.m_RoadLength,
self.m_MutationProbability,
self.m_MinStopsDistance,
self.m_MaxStopsDistance,
self.m_bannedAreas)
self.m_AlfaMale.m_Quality = math.inf
class Population:
pop_size = 100
idv = Individual()
def __init__(self):
pass
def creat_pop(self):
P = []
for i in range(self.pop_size):
P.append(self.idv.creat_one())
return P
def next_Pop(self, Pop):
n_P = []
idv = self.idv
P = Pop
p_size = len(P)
for p in P:
p = idv.reset_one(p)
cp = idv.creat_one()
cp.X[:] = p.X[:]
cp.F_value[:] = p.F_value[:]
n_P.append(cp)
# 产生下一代
select(n_P)
for i in range(p_size):
j = np.random.randint(0, p_size, size=1)[0]
cross(n_P[i], n_P[j])
n_P[j] = mutate(n_P[j])
return n_P
def __move(self, individual: Individual, direction: str) -> None:
""" Move the individual around the map.
This moves the individual according to the direction given.
Zombie can move from one edge through the edge.
Args:
individual (Individual) : Individual of the world
direction (str) : Direction for individual to move
"""
# Make sure to lowercase
direction = direction.lower()
# Going Up
if (direction == 'u'):
if (individual.y_coord == 0):
if (individual.is_zombie):
individual.y_coord = self.N - 1
else:
individual.y_coord -= 1
# Going Down
if (direction == 'd'):
if (individual.y_coord == self.N - 1):
if (individual.is_zombie):
individual.y_coord = 0
else:
individual.y_coord += 1
# Going Left
if (direction == 'l'):
if (individual.x_coord == 0):
if (individual.is_zombie):
individual.x_coord = self.N - 1
else:
individual.x_coord -= 1
# Going Right
if (direction == 'r'):
if (individual.x_coord == self.N - 1):
if (individual.is_zombie):
individual.x_coord = 0
else:
individual.x_coord += 1
def test_populate():
w = World(4)
w.populate(
Individual(),
1, 3
)
assert 1 == len(w.get_population())
def test_queeu_when_zombie_is_populated():
w = World(4)
w.populate(
Individual(is_zombie = True),
1, 3
)
assert 1 == len(w.get_queue())
def play(self) -> None:
""" Run the world simulation
"""
i = 0
for inp in self.inputs:
i += 1
print("Simulation: ", i)
print("World: " + str(inp['n']) + " x " + str(inp['n']))
print()
# Create world
world = World(inp['n'])
# Place zombie
world.populate(Individual(is_zombie=True), inp['zombie']['x'],
inp['zombie']['y'])
# Place creatures
for creature in inp['creatures']:
world.populate(Individual(), creature['x'], creature['y'])
# Simulate
world.simulate(inp['moves'])
# Get stats
score, positions = world.get_stats()
print("zombies' score: ", score)
print("zombies' positions:")
for position in positions:
print("(" + str(position['x']) + ", " + str(position['y']) +
")",
end=" ")
print()
print("---")
print()
def test_has_no_move():
individual = Individual(is_zombie=is_zombie,
x_coord=x,
y_coord=y,
moves=moves)
while individual.has_move():
individual.get_move()
assert individual.has_move() == False
def RankSelection(self):
x = lambda a: a.m_Quality
self.m_Population.sort(key=x)
for i in range(int(self.m_PopulationCount / 2),
self.m_PopulationCount):
self.m_Population[i] = Individual(self.m_StopsCount,
self.m_RoadLength,
self.m_MutationProbability,
self.m_MinStopsDistance,
self.m_MaxStopsDistance,
self.m_bannedAreas)
self.m_Population[i].CreateIndividual()
self.m_Population[i].m_Quality = self.GoalFunction(
self.m_Population[i])
self.m_Population.sort(key=x)
if self.m_Population[0].m_Quality < self.m_AlfaMale.m_Quality:
self.m_AlfaMale.m_Chromosome = self.m_Population[0].m_Chromosome[:]
self.m_AlfaMale.m_Quality = self.m_Population[0].m_Quality
def make_evaluation(population):
"""
проводит оценку особей популяции.
:param population: оцениваемая популяция
:return: лучшая особь
"""
master_individ = Individual(
GENE_LEN, gene_number) # создание экземпляра доминантного индивида
for i in range(POP_LEN):
ind_list = list(population[i].ind_gene)
degree = 1
for j in range(GENE_LEN): # оценка схожести с идеалом
degree += 1 if GENE_LIST[j] == ind_list[j] else 0
population[i].ind_degree = degree # присвоение оценки
# перезапись доминанта при совпадении условий
master_individ = population[
i] if degree > master_individ.ind_degree else master_individ
return master_individ
def RouletteSelection(self):
goalFunctionSum = 0
for i in range(0, self.m_PopulationCount):
goalFunctionSum += self.m_Population[i].m_Quality
for i in range(0, self.m_PopulationCount):
if random.random(
) < self.m_Population[i].m_Quality / goalFunctionSum:
self.m_Population[i] = Individual(self.m_StopsCount,
self.m_RoadLength,
self.m_MutationProbability,
self.m_MinStopsDistance,
self.m_MaxStopsDistance,
self.m_bannedAreas)
self.m_Population[i].CreateIndividual()
self.m_Population[i].m_Quality = self.GoalFunction(
self.m_Population[i])
x = lambda a: a.m_Quality
self.m_Population.sort(key=x)
if self.m_Population[0].m_Quality < self.m_AlfaMale.m_Quality:
self.m_AlfaMale.m_Chromosome = self.m_Population[0].m_Chromosome[:]
self.m_AlfaMale.m_Quality = self.m_Population[0].m_Quality
def cabbage_growing(parent_1, parent_2):
"""
скрещивание 2х особей
:param parent_1: родитель1
:param parent_2: родитель2
:return: 2 потомка особей
"""
# создание экземляров потомков
son = Individual(GENE_LEN, gene_number)
daughter = Individual(GENE_LEN, gene_number)
parents = [parent_1, parent_2]
# псевдоравномерное скрещивание
son.ind_gene = "".join([
rnd.choices(parents, [1, rnd.uniform(1.1, 2.5)])[0][i]
for i in range(GENE_LEN)
])
daughter.ind_gene = "".join([
rnd.choices(parents, [1, rnd.uniform(1.1, 2.5)])[0][i]
for i in range(GENE_LEN)
])
# создание 1 - точечного кроссовера
# recessive_gene = parents[0]
# dominant_gene = parents[1]
# border1 = rnd.randint(0, int(GENE_LEN//2))
#
# son.ind_gene = "".join(recessive_gene[:border1] +
# dominant_gene[border1:])
# daughter.ind_gene = "".join(dominant_gene[:(GENE_LEN-border1)] +
# recessive_gene[(GENE_LEN-border1):])
return son, daughter
def __init__(self):
self.populations = Population()
ga.idv = Individual()
def test_coord():
individual = Individual(is_zombie=is_zombie,
x_coord=x,
y_coord=y,
moves=moves)
assert x == individual.x_coord and y == individual.y_coord
def make_genetic_algorithm(gene, gene_number):
"""
генетический алгоритм оптимизации
:param gene: идеальный геном
:param gene_number: возможная выборка генов
:return:
"""
def make_evaluation(population):
"""
проводит оценку особей популяции.
:param population: оцениваемая популяция
:return: лучшая особь
"""
master_individ = Individual(
GENE_LEN, gene_number) # создание экземпляра доминантного индивида
for i in range(POP_LEN):
ind_list = list(population[i].ind_gene)
degree = 1
for j in range(GENE_LEN): # оценка схожести с идеалом
degree += 1 if GENE_LIST[j] == ind_list[j] else 0
population[i].ind_degree = degree # присвоение оценки
# перезапись доминанта при совпадении условий
master_individ = population[
i] if degree > master_individ.ind_degree else master_individ
return master_individ
def get_selection(population):
"""
проведение селекции лучших особей методом рулетки
:return:
"""
# делим популяцию поровну на готовых к селекции и нет
best_individs = []
if rnd.randint(0, 1): # селекция рулеткой
while len(best_individs) < 0.3 * POP_LEN:
act_pop_len = len(population)
# псевдослучайный выбор особи на скрещивание
# winner = rnd.choices([i for i in range(act_pop_len)],
# [population[i].ind_degree/act_pop_len for i in
# range(act_pop_len)])[0]
# winner = rnd.choices([i for i in range(act_pop_len)],
# [population[i].ind_degree/math.log(math.e + (population[i].ind_degree)) for i in range(act_pop_len)])[0]
#
# турнир
pretendents = [
rnd.randint(0, act_pop_len - 1) for i in range(1, 5)
]
winner = rnd.choices(
pretendents,
[population[i].ind_degree for i in pretendents])[0]
best_individs.append(
population.pop(winner)) # добавление в список селекции
else: # селекция усечением
# сортировка популяции по оценкам и срез
population = sorted(population, key=(lambda ind: ind.ind_degree))
best_individs = population[int(len(population) * 0.7):]
population = population[:int(len(population) * 0.7)]
return best_individs, population
def cabbage_growing(parent_1, parent_2):
"""
скрещивание 2х особей
:param parent_1: родитель1
:param parent_2: родитель2
:return: 2 потомка особей
"""
# создание экземляров потомков
son = Individual(GENE_LEN, gene_number)
daughter = Individual(GENE_LEN, gene_number)
if rnd.randint(0, 1): # псевдоравномерное скрещивание
daughter.ind_gene = "".join(parent_2)
son.ind_gene = "".join([
rnd.choices([parent_1, parent_2], [1, 1])[0][i]
for i in range(GENE_LEN)
])
else: # создание 1 - точечного кроссовера
border1 = rnd.randint(0, int(GENE_LEN // 2))
#
son.ind_gene = "".join(parent_1[:border1] + parent_2[border1:])
daughter.ind_gene = "".join(parent_2[:(GENE_LEN - border1)] +
parent_1[(GENE_LEN - border1):])
return son, daughter
def crossbreed_individs(best_individs):
"""
скрещивание особей
:param best_individs: выборка особей на скрещивание
:return: скрещенная популяция
"""
new_individs = []
ind_len = len(best_individs)
while len(new_individs) < ind_len:
# случайный выбор партнеров на скрешивание
par_ind = [
rnd.randint(0,
len(best_individs) - 1),
rnd.randint(0,
len(best_individs) - 1)
]
# составление рабочих списков из выбранных особей
degrees_list = sorted(
[best_individs[par_ind[0]], best_individs[par_ind[1]]],
key=lambda ind: ind.ind_degree)
parents_list = [
list(degrees_list[0].ind_gene),
list(degrees_list[1].ind_gene)
]
coincidence = 0 # совпадение по генам
for i in range(GENE_LEN):
coincidence += 1 if parents_list[0][i] == parents_list[1][
i] == gene[i] else 0
" если худшая особь имеет некоторое количество оригинальных хороших" \
"генов, то пара скрещивается"
if math.fabs(degrees_list[0].ind_degree - coincidence) >= 1:
# скрещивание и добавление потомков в новую популяцию
new_individs.extend(
cabbage_growing(parents_list[0], parents_list[1]))
# сокращение списка потенциальных партнеров
best_individs = [
best_individs[i] for i in range(len(best_individs))
if i not in par_ind
]
return new_individs
def mutate_individs(feeble_individs):
"""
процесс мутации
:param feeble_individs:
:return:
"""
mutation_prob = 1 / GENE_LEN
for i in range(len(feeble_individs)):
individ_gene = list(feeble_individs[i].ind_gene)
for j in range(GENE_LEN):
if rnd.choices([0, 1],
[1 - mutation_prob,
math.fabs(mutation_prob)])[0]:
individ_gene[j] = rnd.choice(gene_number)
feeble_individs[i].ind_gene = "".join(individ_gene)
return
def output_inform(master):
master_list = list(master.ind_gene)
print(
"".join([
"\033[34m{}".format(master_list[i]) if master_list[i]
== GENE_LIST[i] else "\033[31m{}".format(master_list[i])
for i in range(GENE_LEN)
]), Style.RESET_ALL, master.ind_degree)
def handle_genotype(pop):
"""
обработка популяции
:param pop: популяция
:return: новая популяция
"""
best_individs, feeble_individs = get_selection(pop) # селекция
# print([best_individs[i].ind_degree for i in range(len(best_individs))], len(best_individs))
new_individs = crossbreed_individs(best_individs) # скрещивание
mutate_individs(feeble_individs) # мутация
return feeble_individs + new_individs # новая популяция
GENE_LIST = list(gene) # разбивка строки на символьный список
GENE_LEN = len(GENE_LIST) # длина генома
POP_LEN = 20
# размер популяции
STEP = 100 # шаг смены эры
assert POP_LEN % 4 == 0, "Размер популяции приведет к зацикливанию"
pop1 = [Individual(GENE_LEN, gene_number)
for i in range(POP_LEN)] # создание новой популяции
# pop2 = [Individual(GENE_LEN, gene_number) for j in range(POP_LEN)] # создание новой популяции
# pop3 = [Individual(GENE_LEN, gene_number) for i in range(POP_LEN)] # создание новой популяции
generation = 1
while True:
print("\nERA:", generation // STEP, "GENERATION: ", generation % STEP)
# оценка популяций
master1 = make_evaluation(pop1)
# master2 = make_evaluation(pop2)
# master3 = make_evaluation(pop3)
# вывод схожестей генов
output_inform(master1)
# output_inform(master2)
# output_inform(master3)
# условие выхода из алгоритма
# if (master1.ind_degree > GENE_LEN) or (master2.ind_degree > GENE_LEN)\
# or (master3.ind_degree > GENE_LEN): break
# if (master1.ind_degree > GENE_LEN) or (master2.ind_degree > GENE_LEN): break
if (master1.ind_degree > GENE_LEN): break
# проведение селекции, скрещивания и мутации
pop1 = handle_genotype(pop1)
# pop2 = handle_genotype(pop2)
# pop3 = handle_genotype(pop3, master3.ind_degree)
generation += 1
def test_is_zombie():
individual = Individual(is_zombie=is_zombie,
x_coord=x,
y_coord=y,
moves=moves)
assert is_zombie == individual.is_zombie
def test_has_move():
individual = Individual(is_zombie=is_zombie,
x_coord=x,
y_coord=y,
moves=moves)
assert individual.has_move() == True
