def run(num_chromosomes, generations):
random.seed()
type1 = ['q12', 'q23', 'q34', 'q41']
type2 = ['q1', 'q2', 'q3', 'q4']
sample = SimpleChromosome([
SimpleGene('q12', type1, 'unit-1'),
SimpleGene('q12', type1, 'unit-2'),
SimpleGene('q1', type2, 'unit-3'),
SimpleGene('q1', type2, 'unit-4'),
SimpleGene('q1', type2, 'unit-5'),
SimpleGene('q1', type2, 'unit-6'),
SimpleGene('q1', type2, 'unit-7')
])
self_path = os.path.dirname(os.path.abspath(__file__))
f = open(os.path.join(self_path, 'power.cl'), 'r')
fstr = ''.join(f.readlines())
f.close()
import threading
evt = threading.Event()
evt.clear()
def state_changed(state):
if 'stopped' == state:
evt.set()
ga_cl = OpenCLGA(
{
'sample_chromosome': sample,
'termination': {
'type': 'count',
'count': generations
},
'population': num_chromosomes,
'fitness_kernel_str': fstr,
'fitness_func': 'power_station_fitness',
'opt_for_max': 'max',
'debug': True,
'generation_callback': show_generation_info
},
action_callbacks={'state': state_changed})
ga_cl.prepare()
prob_mutate = 0.05
prob_cross = 0.8
ga_cl.run(prob_mutate, prob_cross)
evt.wait()
utils.plot_ga_result(ga_cl.get_statistics())
print('run took', ga_cl.elapsed_time, 'seconds')
best_chromosome, best_fitness, best_info = ga_cl.get_the_best()
print('Best Fitness: %f' % (best_fitness))
print('1 ~ 7 units are maintained at: ' +
', '.join(str(g.dna) for g in best_info.genes))
def run(num_chromosomes, generations):
value_ranges = [10, 20, 50, 150, 250, 300, 250, 150, 50, 20, 10]
sample = SimpleChromosome(
[SimpleGene(0, list(range(v))) for v in value_ranges])
algebra_path = os.path.dirname(os.path.abspath(__file__))
ocl_kernels = os.path.realpath(
os.path.join(algebra_path, '..', '..', 'kernel'))
algebra_kernels = os.path.join(algebra_path, 'kernel')
f = open(os.path.join(algebra_kernels, 'expansion.cl'), 'r')
fstr = ''.join(f.readlines())
f.close()
import threading
evt = threading.Event()
evt.clear()
def state_changed(state):
if 'stopped' == state:
evt.set()
ga_cl = OpenCLGA(
{
'sample_chromosome': sample,
'termination': {
'type': 'count',
'count': generations
},
'population': num_chromosomes,
'fitness_kernel_str': fstr,
'fitness_func': 'expansion_fitness',
'extra_include_path': [ocl_kernels],
'opt_for_max': 'min',
'generation_callback': show_generation_info
},
action_callbacks={'state': state_changed})
ga_cl.prepare()
prob_mutate = 0.1
prob_cross = 0.8
ga_cl.run(prob_mutate, prob_cross)
evt.wait()
utils.plot_ga_result(ga_cl.get_statistics())
print('run took', ga_cl.elapsed_time, 'seconds')
best_chromosome, best_fitness, best_info = ga_cl.get_the_best()
print('Best Fitness: %f' % (best_fitness))
print('Expansion of (x + y)^10 is: ' +
' '.join(str(g) for g in best_chromosome))
def get_taiwan_travel_info():
'''
NOTE : Config TaiwanTravel GA parameters in dictionary 'dict_info'.
'''
cities, city_info, city_infoX, city_infoY = read_all_cities('TW319_368Addresses-no-far-islands.json')
city_ids = list(range(len(cities)))
random.seed()
tsp_path = os.path.dirname(os.path.abspath(__file__))
tsp_kernels = os.path.join(tsp_path, 'kernel')
sample = ShufflerChromosome([SimpleGene(v, cities) for v in city_ids])
f = open(os.path.join(tsp_kernels, 'taiwan_fitness.cl'), 'r')
fstr = ''.join(f.readlines())
f.close()
# It seems we don't need to use this helper if we enlarge the population size. Please
# re-evaluate and remove or uncomment the following line:
# sample.use_improving_only_mutation('improving_only_mutation_helper')
dict_info = { 'sample_chromosome' : sample,
'termination' : { 'type' : 'count',
'count' : 100 },
'population' : 10240,
'fitness_kernel_str' : fstr,
'fitness_func' : 'taiwan_fitness',
'fitness_args' : [{ 't' : 'float', 'v' : city_infoX, 'n' : 'x' },
{ 't' : 'float', 'v' : city_infoY, 'n' : 'y' }],
'opt_for_max' : 'min',
'saved_filename' : 'test%d%d.pickle',
'prob_mutation' : 0.1,
'prob_crossover' : 0.8,
'extinction' : { 'type' : 'best_avg',
'diff' : 1,
'ratio' : 0.9 },
'elitism_mode' : { 'every' : 2, 'interval' : 10, 'compress' : False },
'serializer': serializer}
return dict_info
def run(num_chromosomes, generations):
random.seed()
# Operations and Terminals
terminals = [('bid', 0), ('ask', 0), ('0.5', 0), ('bidv', 0), ('askv', 0)]
functions1 = [('^2', 1), ('neg', 1)]
functions2 = [('+', 2), ('-', 2), ('*', 2), ('/', 2)]
# Depth of the tree
min_depth = 4
max_depth = 8
depth = random.randint(min_depth, max_depth)
# Tree Initialization (should not require modification if the data structure is not changed)
i = 0
j = 0
n = 1
formula = []
setList = []
temp = []
tempSetList = []
numNodes = n
for i in range(depth):
while 1:
tmp_choice = random.randint(0, 2)
if tmp_choice == 0:
k = random.randint(0, len(terminals) - 1)
temp.append((terminals[k], terminals, terminals[k][0]))
tempSetList.append(tmp_choice)
j += 0
if tmp_choice == 1:
k = random.randint(0, len(functions1) - 1)
temp.append((functions1[k], functions1, functions1[k][0]))
tempSetList.append(tmp_choice)
j += 1
if tmp_choice == 2:
k = random.randint(0, len(functions2) - 1)
temp.append((functions2[k], functions2, functions2[k][0]))
tempSetList.append(tmp_choice)
j += 2
if len(temp) > n:
temp.clear()
tempSetList.clear()
j = 0
if len(temp) == n:
if (j > 0 & i + 1 == depth):
temp.clear()
tempSetList.clear()
j = 0
if (j == 0 & i + 1 < depth):
temp.clear()
tempSetList.clear()
j = 0
else:
break
formula.extend(temp)
setList.extend(tempSetList)
temp = []
tempSetList = []
n = j
j = 0
numNodes = numNodes + n
i += 1
for i in range(0, n):
k = random.randint(0, len(terminals) - 1)
formula.append((terminals[k], terminals, terminals[k][0]))
setList.append(0)
#tree_size = numNodes;
# GP run
# sample chromosome - which defines the format of following generations
# first arg - gene, second arg - mutation set
sample = SimpleChromosome(
[SimpleGene(node[0], node[1], node[2]) for node in formula])
# path of the cl file
self_path = os.path.dirname(os.path.abspath(__file__))
f = open(os.path.join(self_path, 'theo.cl'), 'r')
fstr = ''.join(f.readlines())
f.close()
import threading
evt = threading.Event()
evt.clear()
def state_changed(state):
if 'stopped' == state:
evt.set()
ga_cl = OpenCLGA(
{
'sample_chromosome':
sample,
'termination': {
'type': 'count',
'count': generations
},
'population':
num_chromosomes,
'fitness_kernel_str':
fstr,
'fitness_func':
'theo_fitness', # match this with the fitness func name in cl file
'fitness_args': [
{
't': 'float',
'v': theo,
'n': 'theo'
}, #change here when have more inputs
{
't': 'float',
'v': bid,
'n': 'bid'
},
{
't': 'float',
'v': ask,
'n': 'ask'
},
{
't': 'float',
'v': bidv,
'n': 'bidv'
},
{
't': 'float',
'v': askv,
'n': 'askv'
},
{
't': 'int',
'v': data_size,
'n': 'data_size'
},
{
't': 'int',
'v': numNodes,
'n': 'tree_size'
},
{
't': 'int',
'v': setList,
'n': 'setList'
}
],
'opt_for_max':
'min',
#'elitism_mode': {'top': 10,
# 'every': 10},
'debug':
False,
'generation_callback':
show_generation_info
},
action_callbacks={'state': state_changed})
ga_cl.prepare()
# set up the mutation/crossover probabilities
prob_mutate = 0.6
prob_cross = 0.3
ga_cl.run(prob_mutate, prob_cross)
evt.wait()
print('run took', ga_cl.elapsed_time, 'seconds')
best_chromosome, best_fitness, best_info = ga_cl.get_the_best()
print(' '.join(str(g.dna) for g in best_info.genes))
print('Fitness: %f' % (best_fitness))
# test data output chunk, can be removed
f = open('testdata.txt', 'a')
value = str(best_fitness)
time = str(ga_cl.elapsed_time)
f.write(value + ',' + time + '\n')
f.close()
def run(num_chromosomes, generations):
num_cities = 20
random.seed()
city_ids = list(range(0, num_cities))
city_info = {
city_id: (random.random() * 100, random.random() * 100)
for city_id in city_ids
}
sample = ShufflerChromosome([SimpleGene(v, city_ids) for v in city_ids])
tsp_path = os.path.dirname(os.path.abspath(__file__))
tsp_kernels = os.path.join(tsp_path, 'kernel')
f = open(os.path.join(tsp_kernels, 'simple_tsp.cl'), 'r')
fstr = ''.join(f.readlines())
f.close()
pointX = [str(city_info[v][0]) for v in city_info]
pointY = [str(city_info[v][1]) for v in city_info]
import threading
evt = threading.Event()
evt.clear()
def state_changed(state):
if 'stopped' == state:
evt.set()
tsp_ga_cl = OpenCLGA(
{
'sample_chromosome':
sample,
'termination': {
'type': 'count',
'count': generations
},
'population':
num_chromosomes,
'fitness_kernel_str':
fstr,
'fitness_func':
'simple_tsp_fitness',
'fitness_args': [{
't': 'float',
'v': pointX,
'n': 'x'
}, {
't': 'float',
'v': pointY,
'n': 'y'
}],
'opt_for_max':
'min',
'debug':
True,
'generation_callback':
show_generation_info
},
action_callbacks={'state': state_changed})
tsp_ga_cl.prepare()
prob_mutate = 0.1
prob_cross = 0.8
tsp_ga_cl.run(prob_mutate, prob_cross)
evt.wait()
utils.plot_ga_result(tsp_ga_cl.get_statistics())
print('run took', tsp_ga_cl.elapsed_time, 'seconds')
best_chromosome, best_fitness, best_info = tsp_ga_cl.get_the_best()
print('Best Fitness: %f' % (best_fitness))
print('Shortest Path: ' + ' => '.join(str(g) for g in best_chromosome))
utils.plot_tsp_result(city_info, best_chromosome)
def run(num_chromosomes, generations):
random.seed()
# The number of points randomly generated
num_points = 100
random.seed()
point_ids = list(range(0, num_points))
point_info = {
point_id: (random.random() * 100, random.random() * 100)
for point_id in point_ids
}
pointX = [str(point_info[v][0]) for v in point_info]
pointY = [str(point_info[v][1]) for v in point_info]
# The number of group you want to divide.
numOfGroups = 10
group_id_set = list(range(0, numOfGroups))
group_ids = [random.randint(0, numOfGroups - 1) for x in range(num_points)]
sample = SimpleChromosome(
[SimpleGene(groupd_id, group_id_set) for groupd_id in group_ids])
self_path = os.path.dirname(os.path.abspath(__file__))
f = open(os.path.join(self_path, 'grouping.cl'), 'r')
fstr = ''.join(f.readlines())
f.close()
import threading
evt = threading.Event()
evt.clear()
def state_changed(state):
if 'stopped' == state:
evt.set()
ga_cl = OpenCLGA(
{
'sample_chromosome':
sample,
'termination': {
'type': 'count',
'count': generations
},
'population':
num_chromosomes,
'fitness_kernel_str':
fstr,
'fitness_func':
'grouping_fitness',
'fitness_args': [{
't': 'int',
'v': numOfGroups,
'n': 'numOfGroups'
}, {
't': 'float',
'v': pointX,
'n': 'x'
}, {
't': 'float',
'v': pointY,
'n': 'y'
}],
'opt_for_max':
'min',
'debug':
True,
'generation_callback':
show_generation_info
},
action_callbacks={'state': state_changed})
ga_cl.prepare()
prob_mutate = 0.1
prob_cross = 0.8
ga_cl.run(prob_mutate, prob_cross)
evt.wait()
print('run took', ga_cl.elapsed_time, 'seconds')
best_chromosome, best_fitness, best_info = ga_cl.get_the_best()
print(best_chromosome)
utils.plot_grouping_result(group_id_set, best_chromosome, point_info)