def makeProblem(num):
global N
N = num
global fill
fill = [2] * N
global ranges
ranges = array('i', fill)
global ef
ef = CountOnesEvaluationFunction()
global odd
odd = DiscreteUniformDistribution(ranges)
global nf
nf = DiscreteChangeOneNeighbor(ranges)
global mf
mf = DiscreteChangeOneMutation(ranges)
global cf
cf = SingleCrossOver()
global df
df = DiscreteDependencyTree(.1, ranges)
global hcp
hcp = GenericHillClimbingProblem(ef, odd, nf)
global gap
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
global pop
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
global GA_pop
global myWriter, GA_iters, GA_mut, GA_keep
GA_keep = int(GA_pop * .75)
GA_mut = int(GA_pop * .15)
def mimic_discrete(name, ef, odd, ranges, iter_time, iters_total, iters_step, n_trials, params):
for i_trial in range(n_trials):
for samples, keep, m in itertools.product(*params):
df = DiscreteDependencyTree(m, ranges)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
mimic_instance = MIMIC(samples, keep, pop)
mimic_trainer = FixedIterationTrainer(mimic_instance, iters_step)
mimic_state = {'problem': mimic_instance,
'trainer': mimic_trainer}
# wrap into class etc...
wrapper_mimic = AlgoWrapper(mimic_state,
lambda state: state['trainer'].train(),
lambda state: ef.value(state['problem'].getOptimal()),
lambda state: ef.value(state['problem'].getOptimal())
)
# create name and invalidate if super empty
decorated_name = ""
if name is not None and name != "":
decorated_name = name + "_samples_" + str(samples) + "_keep_" + str(keep) + "_m_" + str(m)
timed_trainer = TimedTrainer(decorated_name,
wrapper_mimic,
iter_time,
iters_total,
iters_step,
_param_dict={'name':name,
'samples':samples,
'keep':keep,
'm':m}
)
timed_trainer.run()
def run_mimic(t, samples, keep, m):
fill = [N] * N
ranges = array('i', fill)
ef = TravelingSalesmanRouteEvaluationFunction(points)
odd = DiscreteUniformDistribution(ranges)
fname = outfile.format('MIMIC{}_{}_{}'.format(samples, keep, m),
str(t + 1))
base.write_header(fname)
df = DiscreteDependencyTree(m, ranges)
ef = TravelingSalesmanRouteEvaluationFunction(points)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
mimic = MIMIC(samples, keep, pop)
fit = FixedIterationTrainer(mimic, 10)
times = [0]
for i in range(0, maxIters, 10):
start = clock()
fit.train()
elapsed = time.clock() - start
times.append(times[-1] + elapsed)
fevals = ef.fevals
score = ef.value(mimic.getOptimal())
ef.fevals -= 1
st = '{},{},{},{}\n'.format(i, score, times[-1], fevals)
# print st
base.write_to_file(fname, st)
return
def run_mimic(t, samples, keep, m):
fname = outfile.format('MIMIC{}_{}_{}'.format(samples, keep, m),
str(t + 1))
ef = ContinuousPeaksEvaluationFunction(T)
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mf = DiscreteChangeOneMutation(ranges)
cf = SingleCrossOver()
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
df = DiscreteDependencyTree(m, ranges)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
mimic = MIMIC(samples, keep, pop)
fit = FixedIterationTrainer(mimic, 10)
times = [0]
for i in range(0, maxIters, 10):
start = clock()
fit.train()
elapsed = time.clock() - start
times.append(times[-1] + elapsed)
fevals = ef.fevals
score = ef.value(mimic.getOptimal())
ef.fevals -= 1
st = '{},{},{},{}\n'.format(i, score, times[-1], fevals)
# print st
base.write_to_file(fname, st)
return
def solveit(oaname, params):
N = 60
T = N / 10
fill = [2] * N
ranges = array('i', fill)
iterations = 10000
tryi = 1
ef = ContinuousPeaksEvaluationFunction(T)
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mf = DiscreteChangeOneMutation(ranges)
cf = SingleCrossOver()
df = DiscreteDependencyTree(.1, ranges)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
# fit = FixedIterationTrainer(rhc, 200000)
# fit.train()
if oaname == 'RHC':
iterations = int(params[0])
tryi = int(params[1])
oa = RandomizedHillClimbing(hcp)
if oaname == 'SA':
oa = SimulatedAnnealing(float(params[0]), float(params[1]), hcp)
if oaname == 'GA':
oa = StandardGeneticAlgorithm(int(params[0]), int(params[1]),
int(params[2]), gap)
if oaname == 'MMC':
oa = MIMIC(int(params[0]), int(params[1]), pop)
print "Running %s using %s for %d iterations, try %d" % (
oaname, ','.join(params), iterations, tryi)
print "=" * 20
starttime = timeit.default_timer()
output = []
for i in range(iterations):
oa.train()
if i % 10 == 0:
optimal = oa.getOptimal()
score = ef.value(optimal)
elapsed = float(timeit.default_timer() - starttime)
output.append([str(i), str(score), str(elapsed)])
print 'score: %.3f' % score
print 'train time: %.3f secs' % (int(timeit.default_timer() - starttime))
scsv = 'cp-%s-%s.csv' % (oaname, '-'.join(params))
print "Saving to %s" % (scsv),
with open(scsv, 'w') as csvf:
writer = csv.writer(csvf)
for row in output:
writer.writerow(row)
print "saved."
print "=" * 20
def mimic():
for samples, keep, m in product([80], [40], [0.15, 0.35, 0.55, 0.75, 0.95]):
fname = outfile.replace('XXX','MIMIC{}_{}_{}'.format(samples,keep,m))
df = DiscreteDependencyTree(m, ranges)
with open(fname,'w') as f:
f.write('iterations,fitness,time,fevals\n')
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
mimic = MIMIC(samples, keep, pop)
perform(mimic, fname)
def run_four_peaks_exploringSA():
N=200
T=N/5
fill = [2] * N
ranges = array('i', fill)
ef = FourPeaksEvaluationFunction(T)
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mf = DiscreteChangeOneMutation(ranges)
cf = SingleCrossOver()
df = DiscreteDependencyTree(.1, ranges)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
iters = [50, 100, 250, 500, 1000, 2500, 5000, 10000, 25000, 30000, 35000, 40000, 45000, 50000]
num_repeats = 5
all_sa_results = []
all_sa_times = []
coolings = [0.15, 0.35, 0.55, 0.75, 0.95]
for cooling in coolings:
sa_results = []
sa_times = []
for i in iters:
print(i)
for j in range(num_repeats):
start = time.time()
sa = SimulatedAnnealing(1E11, cooling, hcp)
fit = FixedIterationTrainer(sa, i)
fit.train()
end = time.time()
sa_results.append(ef.value(sa.getOptimal()))
sa_times.append(end - start)
print "SA cooling " + str(cooling) + ": " + str(ef.value(sa.getOptimal()))
all_sa_results.append(sa_results)
all_sa_results.append(sa_times)
with open('four_peaks_exploringSA.csv', 'w') as csvfile:
writer = csv.writer(csvfile)
for sa_results in all_sa_results:
writer.writerow(sa_results)
for sa_times in all_sa_times:
writer.writerow(sa_times)
return all_sa_results, all_sa_times
def run_experiment(self, opName):
"""Run a MIMIC optimization experiment for a given optimization
problem.
Args:
ef (AbstractEvaluationFunction): Evaluation function.
ranges (array): Search space ranges.
op (str): Name of optimization problem.
"""
outdir = 'results/OPT/{}'.format(opName) # get results directory
outfile = 'MIMIC_{}_{}_{}_results.csv'.format(self.s, self.k, self.m)
fname = get_abspath(outfile, outdir) # get output filename
# delete existing results file, if it already exists
try:
os.remove(fname)
except Exception as e:
print e
pass
with open(fname, 'w') as f:
f.write('iterations,fitness,time,fevals,trial\n')
# start experiment
for t in range(self.numTrials):
# initialize optimization problem and training functions
ranges, ef = self.op.get_ef()
mimic = None
df = DiscreteDependencyTree(self.m, ranges)
odd = DiscreteUniformDistribution(ranges)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
mimic = MIMIC(self.s, self.k, pop)
fit = FixedIterationTrainer(mimic, 10)
# run experiment and train evaluation function
start = time.clock()
for i in range(0, self.maxIters, 10):
fit.train()
elapsed = time.clock() - start
fe = ef.valueCallCount
score = ef.value(mimic.getOptimal())
ef.valueCallCount -= 1
# write results to output file
s = '{},{},{},{},{}\n'.format(i + 10, score, elapsed, fe, t)
with open(fname, 'a+') as f:
f.write(s)
def mimicGATest():
popBegin = 1
popEnd = 101
keepBegin = 1
keepEnd = 90
mutBegin = 1
mutEnd = 90
itersBegin = 1
itersEnd = 200
samples = 10
keep = 2
problemSize = N
mimicRange = (problemSize)
iters = 1
paramRanges = Vector(8)
paramRanges.addElement(popBegin)
paramRanges.addElement(popEnd)
paramRanges.addElement(keepBegin)
paramRanges.addElement(keepEnd)
paramRanges.addElement(mutBegin)
paramRanges.addElement(mutEnd)
paramRanges.addElement(itersBegin)
paramRanges.addElement(itersEnd)
totalParamSize1 = (popEnd - popBegin + 1) + (keepEnd - keepBegin + 1) + (
mutEnd - mutBegin + 1) + (itersEnd - itersBegin + 1)
allParamValues = range(popBegin, popEnd + 1) + range(
keepBegin, keepEnd + 1) + range(mutBegin, mutEnd + 1) + range(
itersBegin, itersEnd + 1)
totalParamSize = len(allParamValues)
metaFun = RamysEvalMetafunc(ranges)
discreteDist = RamysMimicDistribution(
paramRanges) #DiscreteUniformDistribution(problemSize)
distFunc = DiscreteDependencyTree(.1, allParamValues)
findGA = GenericProbabilisticOptimizationProblem(metaFun, discreteDist,
distFunc)
mimic = MIMIC(samples, keep, findGA)
fit = FixedIterationTrainer(mimic, iters)
fit.train()
print str(N) + ": MIMIC finds GA : " + str(ef.value(mimic.getOptimal()))
def main():
"""Run this experiment"""
training_ints = initialize_instances(PATH + "X_train.csv")
testing_ints = initialize_instances(PATH + "X_test.csv")
validation_ints = initialize_instances(PATH + "y_train.csv")
factory = BackPropagationNetworkFactory()
measure = SumOfSquaresError()
logistic_sigmoid = LogisticSigmoid()
data_set = DataSet(training_ints)
data_set_size = data_set.size()
print(data_set_size)
print(type(data_set_size))
odd = DiscreteUniformDistribution([data_set_size])
df = DiscreteDependencyTree(.1, [data_set_size])
classification_network = factory.createClassificationNetwork(
[INPUT_LAYER, HIDDEN_LAYER1, OUTPUT_LAYER], logistic_sigmoid)
evaluation = NeuralNetworkEvaluationFunction(classification_network,
data_set, measure)
pop = GenericProbabilisticOptimizationProblem(evaluation, odd, df)
oa = MIMIC(data_set_size, int(0.1 * data_set_size), pop)
train(oa, classification_network, 'GA', training_ints, validation_ints,
testing_ints, measure)
# create range
fill = [COPIES_EACH + 1] * NUM_ITEMS
ranges = array('i', fill)
ef = KnapsackEvaluationFunction(weights, volumes, KNAPSACK_VOLUME, copies)
initial_distribution = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mutation_function = DiscreteChangeOneMutation(ranges)
cf = SingleCrossOver()
df = DiscreteDependencyTree(.1, ranges)
hill_climbing_problem = GenericHillClimbingProblem(ef, initial_distribution, nf)
genetic_problem = GenericGeneticAlgorithmProblem(ef, initial_distribution, mutation_function, cf)
probablistic_optimization = GenericProbabilisticOptimizationProblem(ef, initial_distribution, df)
from time import time
f = open("experiments/results/knapsack_optimal2.txt", "w")
f.write("starting RHC\n")
rhc = RandomizedHillClimbing(hill_climbing_problem)
score = 0
iters = 0
t0 = time()
while iters < 80000:
score = rhc.train()
f.write(str(iters) + "," + str(score) +"\n")
iters += 1
def run_all():
problem = 'knapsack'
# Random number generator */
random = Random()
# The number of items
NUM_ITEMS = 40
# The number of copies each
COPIES_EACH = 4
# The maximum weight for a single element
MAX_WEIGHT = 50
# The maximum volume for a single element
MAX_VOLUME = 50
# The volume of the knapsack
KNAPSACK_VOLUME = MAX_VOLUME * NUM_ITEMS * COPIES_EACH * .4
# create copies
fill = [COPIES_EACH] * NUM_ITEMS
copies = array('i', fill)
# create weights and volumes
fill = [0] * NUM_ITEMS
weights = array('d', fill)
volumes = array('d', fill)
for i in range(0, NUM_ITEMS):
weights[i] = random.nextDouble() * MAX_WEIGHT
volumes[i] = random.nextDouble() * MAX_VOLUME
# create range
fill = [COPIES_EACH + 1] * NUM_ITEMS
ranges = array('i', fill)
ef = KnapsackEvaluationFunction(weights, volumes, KNAPSACK_VOLUME, copies)
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mf = DiscreteChangeOneMutation(ranges)
cf = UniformCrossOver()
df = DiscreteDependencyTree(.1, ranges)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
maxEpochs = 400
columns = ['problem','label','score','epoch','time','avgTrainTime','iterations']
outFile = open(filename+'_all.csv','wb')
fout = csv.writer(outFile,delimiter=',')
fout.writerow(columns)
def run_algo(alg,fit,label,iters):
print(alg)
trainTimes = [0.]
trainTime = []
scores = [0]
deltaScores = []
for epoch in range(0,maxEpochs,1):
st = time.clock()
fit.train()
et = time.clock()
trainTimes.append(trainTimes[-1]+(et-st))
trainTime.append((et-st))
rollingMean = 10
avgTime = (math.fsum(trainTime[-rollingMean:]) / float(rollingMean))
score = ef.value(alg.getOptimal())
scores.append(score)
deltaScores.append(math.fabs(scores[-2] - scores[-1]))
# trialString = '{}-{}-{}-{}'.format(label,score,epoch,trainTimes[-1])
trialData = [problem,label,score,epoch,trainTimes[-1],avgTime,iters]
print(trialData)
fout.writerow(trialData)
iters = 10
rhc = RandomizedHillClimbing(hcp)
fit = FixedIterationTrainer(rhc, iters)
run_algo(rhc,fit,'RHC',10)
startTemp = 1E11
coolingFactor = .95
sa = SimulatedAnnealing(startTemp, coolingFactor, hcp)
fit = FixedIterationTrainer(sa, iters)
run_algo(sa,fit,'HCP',10)
population = 300
mates = 100
mutations = 50
ga = StandardGeneticAlgorithm(population, mates, mutations, gap)
fit = FixedIterationTrainer(ga, iters)
run_algo(ga,fit,'GA',10)
samples = 200
keep = 20
mimic = MIMIC(samples, keep, pop)
fit = FixedIterationTrainer(mimic, iters)
run_algo(mimic,fit,'MIMIC',10)
outFile.close()
def run_knapsack_experiments():
OUTPUT_DIRECTORY = './output'
# Random number generator */
random = Random()
# The number of items
NUM_ITEMS = 40
# The number of copies each
COPIES_EACH = 4
# The maximum weight for a single element
MAX_WEIGHT = 50
# The maximum volume for a single element
MAX_VOLUME = 50
# The volume of the knapsack
KNAPSACK_VOLUME = MAX_VOLUME * NUM_ITEMS * COPIES_EACH * .4
# create copies
fill = [COPIES_EACH] * NUM_ITEMS
copies = array('i', fill)
# create weights and volumes
fill = [0] * NUM_ITEMS
weights = array('d', fill)
volumes = array('d', fill)
for i in range(0, NUM_ITEMS):
weights[i] = random.nextDouble() * MAX_WEIGHT
volumes[i] = random.nextDouble() * MAX_VOLUME
# create range
fill = [COPIES_EACH + 1] * NUM_ITEMS
ranges = array('i', fill)
ef = KnapsackEvaluationFunction(weights, volumes, KNAPSACK_VOLUME, copies)
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mf = DiscreteChangeOneMutation(ranges)
cf = UniformCrossOver()
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
max_iter = 5000
outfile = OUTPUT_DIRECTORY + '/knapsack_{}_log.csv'
# Randomized Hill Climber
filename = outfile.format('rhc')
with open(filename, 'w') as f:
f.write('iterations,fitness,time\n')
for it in range(0, max_iter, 10):
rhc = RandomizedHillClimbing(hcp)
fit = FixedIterationTrainer(rhc, it)
start_time = time.clock()
fit.train()
elapsed_time = time.clock() - start_time
# fevals = ef.fevals
score = ef.value(rhc.getOptimal())
data = '{},{},{}\n'.format(it, score, elapsed_time)
print(data)
with open(filename, 'a') as f:
f.write(data)
# Simulated Annealing
filename = outfile.format('sa')
with open(filename, 'w') as f:
f.write('iteration,cooling_value,fitness,time\n')
for cooling_value in (.19, .38, .76, .95):
for it in range(0, max_iter, 10):
sa = SimulatedAnnealing(200, cooling_value, hcp)
fit = FixedIterationTrainer(sa, it)
start_time = time.clock()
fit.train()
elapsed_time = time.clock() - start_time
# fevals = ef.fevals
score = ef.value(sa.getOptimal())
data = '{},{},{},{}\n'.format(it, cooling_value, score,
elapsed_time)
print(data)
with open(filename, 'a') as f:
f.write(data)
# Genetic Algorithm
filename = outfile.format('ga')
with open(filename, 'w') as f:
f.write('iteration,population_size,to_mate,to_mutate,fitness,time\n')
for population_size, to_mate, to_mutate in itertools.product(
[200], [110, 120, 130, 140, 150], [2, 4, 6, 8]):
for it in range(0, max_iter, 10):
ga = StandardGeneticAlgorithm(population_size, to_mate, to_mutate,
gap)
fit = FixedIterationTrainer(ga, it)
start_time = time.clock()
fit.train()
elapsed_time = time.clock() - start_time
# fevals = ef.fevals
score = ef.value(ga.getOptimal())
data = '{},{},{},{},{},{}\n'.format(it, population_size, to_mate,
to_mutate, score, elapsed_time)
print(data)
with open(filename, 'a') as f:
f.write(data)
# MIMIC
filename = outfile.format('mm')
with open(filename, 'w') as f:
f.write('iterations,samples,to_keep,m,fitness,time\n')
for samples, to_keep, m in itertools.product([200], [100],
[0.1, 0.3, 0.5, 0.7, 0.9]):
for it in range(0, 500, 10):
df = DiscreteDependencyTree(m, ranges)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
mm = MIMIC(samples, 20, pop)
fit = FixedIterationTrainer(mm, it)
start_time = time.clock()
fit.train()
elapsed_time = time.clock() - start_time
# fevals = ef.fevals
score = ef.value(mm.getOptimal())
data = '{},{},{},{},{},{}\n'.format(it, samples, to_keep, m, score,
elapsed_time)
print(data)
with open(filename, 'a') as f:
f.write(data)
def solveit(oaname, params):
# set N value. This is the number of points
N = 50
iterations = 1000
tryi = 1
random = Random()
points = [[0 for x in xrange(2)] for x in xrange(N)]
for i in range(0, len(points)):
points[i][0] = random.nextDouble()
points[i][1] = random.nextDouble()
ef = TravelingSalesmanRouteEvaluationFunction(points)
odd = DiscretePermutationDistribution(N)
nf = SwapNeighbor()
mf = SwapMutation()
cf = TravelingSalesmanCrossOver(ef)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
if oaname == "RHC":
iterations = int(params[0])
tryi = int(params[1])
oa = RandomizedHillClimbing(hcp)
if oaname == "SA":
oa = SimulatedAnnealing(float(params[0]), float(params[1]), hcp)
if oaname == "GA":
iterations=1000
oa = StandardGeneticAlgorithm(int(params[0]), int(params[1]), int(params[2]), gap)
if oaname == "MMC":
iterations=1000
# for mimic we use a sort encoding
ef = TravelingSalesmanSortEvaluationFunction(points)
fill = [N] * N
ranges = array('i', fill)
odd = DiscreteUniformDistribution(ranges)
df = DiscreteDependencyTree(.1, ranges)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
oa = MIMIC(int(params[0]), int(params[1]), pop)
print "Running %s using %s for %d iterations, try %d" % (oaname, ','.join(params), iterations, tryi)
print "="*20
starttime = timeit.default_timer()
output = []
for i in range(iterations):
oa.train()
if i%10 == 0:
optimal = oa.getOptimal()
score = ef.value(optimal)
elapsed = int(timeit.default_timer()-starttime)
output.append([str(i), str(score), str(elapsed)])
print 'Inverse of Distance [score]: %.3f' % score
print 'train time: %d secs' % (int(timeit.default_timer()-starttime))
scsv = 'tsp-%s-%s.csv' % (oaname, '-'.join(params))
print "Saving to %s" % (scsv),
with open(scsv, 'w') as csvf:
writer = csv.writer(csvf)
for row in output:
writer.writerow(row)
print "saved."
print "="*20
print "Route:"
if oaname == 'MMC':
optimal = oa.getOptimal()
fill = [0] * optimal.size()
ddata = array('d', fill)
for i in range(0,len(ddata)):
ddata[i] = optimal.getContinuous(i)
order = ABAGAILArrays.indices(optimal.size())
ABAGAILArrays.quicksort(ddata, order)
print order
else:
path = []
for x in range(0,N):
path.append(oa.getOptimal().getDiscrete(x))
print path
def run_traveling_salesman():
# set N value. This is the number of points
N = 50
random = Random()
points = [[0 for x in xrange(2)] for x in xrange(N)]
for i in range(0, len(points)):
points[i][0] = random.nextDouble()
points[i][1] = random.nextDouble()
ef = TravelingSalesmanRouteEvaluationFunction(points)
odd = DiscretePermutationDistribution(N)
nf = SwapNeighbor()
mf = SwapMutation()
cf = TravelingSalesmanCrossOver(ef)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
iters = [50, 100, 250, 500, 1000, 2500, 5000, 10000, 25000, 50000, 100000]
num_repeats = 5
rhc_results = []
rhc_times = []
for i in iters:
print(i)
for j in range(num_repeats):
start = time.time()
rhc = RandomizedHillClimbing(hcp)
fit = FixedIterationTrainer(rhc, i)
fit.train()
end = time.time()
rhc_results.append(ef.value(rhc.getOptimal()))
rhc_times.append(end - start)
print "RHC Inverse of Distance: " + str(ef.value(rhc.getOptimal()))
# print "Route:"
# path = []
# for x in range(0,N):
# path.append(rhc.getOptimal().getDiscrete(x))
# print path
sa_results = []
sa_times = []
for i in iters:
print(i)
for j in range(num_repeats):
start = time.time()
sa = SimulatedAnnealing(1E12, .999, hcp)
fit = FixedIterationTrainer(sa, i)
fit.train()
sa_results.append(ef.value(sa.getOptimal()))
sa_times.append(end - start)
print "SA Inverse of Distance: " + str(ef.value(sa.getOptimal()))
# print "Route:"
# path = []
# for x in range(0,N):
# path.append(sa.getOptimal().getDiscrete(x))
# print path
ga_results = []
ga_times = []
for i in iters:
print(i)
for j in range(num_repeats):
start = time.time()
ga = StandardGeneticAlgorithm(2000, 1500, 250, gap)
fit = FixedIterationTrainer(ga, i)
fit.train()
end = time.time()
ga_results.append(ef.value(ga.getOptimal()))
print "GA Inverse of Distance: " + str(ef.value(ga.getOptimal()))
ga_times.append(end - start)
# print "Route:"
# path = []
# for x in range(0,N):
# path.append(ga.getOptimal().getDiscrete(x))
# print path
# for mimic we use a sort encoding
ef = TravelingSalesmanSortEvaluationFunction(points)
fill = [N] * N
ranges = array('i', fill)
odd = DiscreteUniformDistribution(ranges)
df = DiscreteDependencyTree(.1, ranges)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
mimic_results = []
mimic_times = []
for i in iters[0:6]:
print(i)
for j in range(num_repeats):
start = time.time()
mimic = MIMIC(500, 100, pop)
fit = FixedIterationTrainer(mimic, i)
fit.train()
end = time.time()
mimic_results.append(ef.value(mimic.getOptimal()))
print "MIMIC Inverse of Distance: " + str(
ef.value(mimic.getOptimal()))
# print "Route:"
# path = []
# optimal = mimic.getOptimal()
# fill = [0] * optimal.size()
# ddata = array('d', fill)
# for i in range(0,len(ddata)):
# ddata[i] = optimal.getContinuous(i)
# order = ABAGAILArrays.indices(optimal.size())
# ABAGAILArrays.quicksort(ddata, order)
# print order
mimic_times.append(end - start)
with open('travelingsalesman.csv', 'w') as csvfile:
writer = csv.writer(csvfile)
writer.writerow(rhc_results)
writer.writerow(rhc_times)
writer.writerow(sa_results)
writer.writerow(sa_times)
writer.writerow(ga_results)
writer.writerow(ga_times)
writer.writerow(mimic_results)
writer.writerow(mimic_times)
return rhc_results, rhc_times, sa_results, sa_times, ga_results, ga_times, mimic_results, mimic_times
def fourpeaksfunc(N, iterations):
rhcMult = 200
saMult = 200
gaMult = 2
mimicMult = 1
optimalOut = []
timeOut = []
evalsOut = []
T = N / 5
fill = [2] * N
ranges = array('i', fill)
ef = FourPeaksEvaluationFunction(T)
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mf = DiscreteChangeOneMutation(ranges)
cf = SingleCrossOver()
df = DiscreteDependencyTree(.1, ranges)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
for niter in iterations:
iterOptimalOut = [N, niter]
iterTimeOut = [N, niter]
iterEvals = [N, niter]
start = time.time()
rhc = RandomizedHillClimbing(hcp)
fit = FixedIterationTrainer(rhc, niter * rhcMult)
fit.train()
end = time.time()
rhcOptimal = ef.value(rhc.getOptimal())
rhcTime = end - start
print "RHC optimum: " + str(rhcOptimal)
print "RHC time: " + str(rhcTime)
iterOptimalOut.append(rhcOptimal)
iterTimeOut.append(rhcTime)
functionEvals = ef.getNumEvals()
ef.zeroEvals()
iterEvals.append(functionEvals)
start = time.time()
sa = SimulatedAnnealing(1E20, .8, hcp)
fit = FixedIterationTrainer(sa, niter * saMult)
fit.train()
end = time.time()
saOptimal = ef.value(sa.getOptimal())
saTime = end - start
print "SA optimum: " + str(saOptimal)
print "SA time: " + str(saTime)
iterOptimalOut.append(saOptimal)
iterTimeOut.append(saTime)
functionEvals = ef.getNumEvals()
ef.zeroEvals()
iterEvals.append(functionEvals)
start = time.time()
ga = StandardGeneticAlgorithm(200, 100, 10, gap)
fit = FixedIterationTrainer(ga, niter * gaMult)
fit.train()
end = time.time()
gaOptimal = ef.value(ga.getOptimal())
gaTime = end - start
print "GA optimum: " + str(gaOptimal)
print "GA time: " + str(gaTime)
iterOptimalOut.append(gaOptimal)
iterTimeOut.append(gaTime)
functionEvals = ef.getNumEvals()
ef.zeroEvals()
iterEvals.append(functionEvals)
start = time.time()
mimic = MIMIC(200, 20, pop)
fit = FixedIterationTrainer(mimic, niter * mimicMult)
fit.train()
end = time.time()
mimicOptimal = ef.value(mimic.getOptimal())
mimicTime = end - start
print "MIMIC optimum: " + str(mimicOptimal)
print "MIMIC time: " + str(mimicTime)
iterOptimalOut.append(mimicOptimal)
iterTimeOut.append(mimicTime)
functionEvals = ef.getNumEvals()
ef.zeroEvals()
iterEvals.append(functionEvals)
optimalOut.append(iterOptimalOut)
timeOut.append(iterTimeOut)
evalsOut.append(iterEvals)
return [optimalOut, timeOut, evalsOut]
ef = TravelingSalesmanRouteEvaluationFunction(points)
odd = DiscretePermutationDistribution(N)
nf = SwapNeighbor()
mf = SwapMutation()
cf = TravelingSalesmanCrossOver(ef)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
# # for mimic we use a sort encoding
efm = TravelingSalesmanSortEvaluationFunction(points)
fill = [N] * N
ranges = array('i', fill)
oddm = DiscreteUniformDistribution(ranges)
dfm = DiscreteDependencyTree(.1, ranges);
pop = GenericProbabilisticOptimizationProblem(efm, oddm, dfm)
def merge_two_dicts(x, y):
z = x.copy() # start with x's keys and values
z.update(y) # modifies z with y's keys and values & returns None
return z
def rhc_fac(args = {}):
constant_params = {'hcp':hcp}
params = merge_two_dicts(args, constant_params)
print(params)
rhc = RandomizedHillClimbing(hcp)
return FixedIterationTrainer(rhc, num_iterations)
def sa_fac(args = {}):
constant_params = {'hcp':hcp}
def travelingsalesmanfunc(N, iterations):
rhcMult = 1500
saMult = 1500
gaMult = 1
mimicMult = 3
random = Random()
points = [[0 for x in xrange(2)] for x in xrange(N)]
for i in range(0, len(points)):
points[i][0] = random.nextDouble()
points[i][1] = random.nextDouble()
optimalOut = []
timeOut = []
evalsOut = []
for niter in iterations:
ef = TravelingSalesmanRouteEvaluationFunction(points)
odd = DiscretePermutationDistribution(N)
nf = SwapNeighbor()
mf = SwapMutation()
cf = TravelingSalesmanCrossOver(ef)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
iterOptimalOut = [N, niter]
iterTimeOut = [N, niter]
iterEvals = [N, niter]
start = time.time()
rhc = RandomizedHillClimbing(hcp)
fit = FixedIterationTrainer(rhc, niter * rhcMult)
fit.train()
end = time.time()
rhcOptimal = ef.value(rhc.getOptimal())
rhcTime = end - start
print "RHC Inverse of Distance: optimum: " + str(rhcOptimal)
print "RHC time: " + str(rhcTime)
#print "RHC Inverse of Distance: " + str(ef.value(rhc.getOptimal()))
print "Route:"
path = []
for x in range(0, N):
path.append(rhc.getOptimal().getDiscrete(x))
print path
iterOptimalOut.append(rhcOptimal)
iterTimeOut.append(rhcTime)
functionEvals = ef.getNumEvals()
ef.zeroEvals()
iterEvals.append(functionEvals)
start = time.time()
sa = SimulatedAnnealing(1E12, .999, hcp)
fit = FixedIterationTrainer(sa, niter * saMult)
fit.train()
end = time.time()
saOptimal = ef.value(sa.getOptimal())
saTime = end - start
print "SA Inverse of Distance optimum: " + str(saOptimal)
print "SA time: " + str(saTime)
#print "SA Inverse of Distance: " + str(ef.value(sa.getOptimal()))
print "Route:"
path = []
for x in range(0, N):
path.append(sa.getOptimal().getDiscrete(x))
print path
iterOptimalOut.append(saOptimal)
iterTimeOut.append(saTime)
functionEvals = ef.getNumEvals()
ef.zeroEvals()
iterEvals.append(functionEvals)
start = time.time()
ga = StandardGeneticAlgorithm(2000, 1500, 250, gap)
fit = FixedIterationTrainer(ga, niter * gaMult)
fit.train()
end = time.time()
gaOptimal = ef.value(ga.getOptimal())
gaTime = end - start
print "GA Inverse of Distance optimum: " + str(gaOptimal)
print "GA time: " + str(gaTime)
#print "GA Inverse of Distance: " + str(ef.value(ga.getOptimal()))
print "Route:"
path = []
for x in range(0, N):
path.append(ga.getOptimal().getDiscrete(x))
print path
iterOptimalOut.append(gaOptimal)
iterTimeOut.append(gaTime)
functionEvals = ef.getNumEvals()
ef.zeroEvals()
iterEvals.append(functionEvals)
start = time.time()
# for mimic we use a sort encoding
ef = TravelingSalesmanSortEvaluationFunction(points)
fill = [N] * N
ranges = array('i', fill)
odd = DiscreteUniformDistribution(ranges)
df = DiscreteDependencyTree(.1, ranges)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
start = time.time()
mimic = MIMIC(500, 100, pop)
fit = FixedIterationTrainer(mimic, niter * mimicMult)
fit.train()
end = time.time()
mimicOptimal = ef.value(mimic.getOptimal())
mimicTime = end - start
print "MIMIC Inverse of Distance optimum: " + str(mimicOptimal)
print "MIMIC time: " + str(mimicTime)
#print "MIMIC Inverse of Distance: " + str(ef.value(mimic.getOptimal()))
print "Route:"
path = []
optimal = mimic.getOptimal()
fill = [0] * optimal.size()
ddata = array('d', fill)
for i in range(0, len(ddata)):
ddata[i] = optimal.getContinuous(i)
order = ABAGAILArrays.indices(optimal.size())
ABAGAILArrays.quicksort(ddata, order)
print order
iterOptimalOut.append(mimicOptimal)
iterTimeOut.append(mimicTime)
functionEvals = ef.getNumEvals()
ef.zeroEvals()
iterEvals.append(functionEvals)
optimalOut.append(iterOptimalOut)
timeOut.append(iterTimeOut)
evalsOut.append(iterEvals)
return [optimalOut, timeOut, evalsOut]
def run_four_peaks():
N = 200
T = N / 5
fill = [2] * N
ranges = array('i', fill)
ef = FourPeaksEvaluationFunction(T)
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mf = DiscreteChangeOneMutation(ranges)
cf = SingleCrossOver()
df = DiscreteDependencyTree(.1, ranges)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
iters = [50, 100, 250, 500, 1000, 2500, 5000, 10000, 25000, 50000, 100000]
num_repeats = 5
rhc_results = []
rhc_times = []
for i in iters:
print(i)
for j in range(num_repeats):
start = time.time()
rhc = RandomizedHillClimbing(hcp)
fit = FixedIterationTrainer(rhc, i)
fit.train()
end = time.time()
rhc_results.append(ef.value(rhc.getOptimal()))
rhc_times.append(end - start)
print "RHC: " + str(ef.value(rhc.getOptimal()))
sa_results = []
sa_times = []
for i in iters:
print(i)
for j in range(num_repeats):
start = time.time()
sa = SimulatedAnnealing(1E11, .95, hcp)
fit = FixedIterationTrainer(sa, i)
fit.train()
end = time.time()
sa_results.append(ef.value(sa.getOptimal()))
sa_times.append(end - start)
print "SA: " + str(ef.value(sa.getOptimal()))
ga_results = []
ga_times = []
for i in iters:
print(i)
for j in range(num_repeats):
start = time.time()
ga = StandardGeneticAlgorithm(200, 100, 10, gap)
fit = FixedIterationTrainer(ga, i)
fit.train()
end = time.time()
ga_results.append(ef.value(ga.getOptimal()))
ga_times.append(end - start)
print "GA: " + str(ef.value(ga.getOptimal()))
mimic_results = []
mimic_times = []
for i in iters[0:6]:
print(i)
for j in range(num_repeats):
start = time.time()
mimic = MIMIC(200, 20, pop)
fit = FixedIterationTrainer(mimic, i)
fit.train()
end = time.time()
mimic_results.append(ef.value(mimic.getOptimal()))
mimic_times.append(end - start)
print "MIMIC: " + str(ef.value(mimic.getOptimal()))
with open('four_peaks.csv', 'w') as csvfile:
writer = csv.writer(csvfile)
writer.writerow(rhc_results)
writer.writerow(rhc_times)
writer.writerow(sa_results)
writer.writerow(sa_times)
writer.writerow(ga_results)
writer.writerow(ga_times)
writer.writerow(mimic_results)
writer.writerow(mimic_times)
return rhc_results, rhc_times, sa_results, sa_times, ga_results, ga_times, mimic_results, mimic_times
def run_knapsack():
# Random number generator */
random = Random()
# The number of items
NUM_ITEMS = 40
# The number of copies each
COPIES_EACH = 4
# The maximum weight for a single element
MAX_WEIGHT = 50
# The maximum volume for a single element
MAX_VOLUME = 50
# The volume of the knapsack
KNAPSACK_VOLUME = MAX_VOLUME * NUM_ITEMS * COPIES_EACH * .4
# create copies
fill = [COPIES_EACH] * NUM_ITEMS
copies = array('i', fill)
# create weights and volumes
fill = [0] * NUM_ITEMS
weights = array('d', fill)
volumes = array('d', fill)
for i in range(0, NUM_ITEMS):
weights[i] = random.nextDouble() * MAX_WEIGHT
volumes[i] = random.nextDouble() * MAX_VOLUME
# create range
fill = [COPIES_EACH + 1] * NUM_ITEMS
ranges = array('i', fill)
ef = KnapsackEvaluationFunction(weights, volumes, KNAPSACK_VOLUME, copies)
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mf = DiscreteChangeOneMutation(ranges)
cf = UniformCrossOver()
df = DiscreteDependencyTree(.1, ranges)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
iters = [50, 100, 250, 500, 1000, 2500, 5000, 10000, 25000, 50000, 100000]
num_repeats = 5
rhc_results = []
rhc_times = []
for i in iters:
print(i)
for j in range(num_repeats):
start = time.time()
rhc = RandomizedHillClimbing(hcp)
fit = FixedIterationTrainer(rhc, i)
fit.train()
end = time.time()
rhc_results.append(ef.value(rhc.getOptimal()))
rhc_times.append(end - start)
#print "RHC: " + str(ef.value(rhc.getOptimal()))
sa_results = []
sa_times = []
for i in iters:
print(i)
for j in range(num_repeats):
start = time.time()
sa = SimulatedAnnealing(100, .95, hcp)
fit = FixedIterationTrainer(sa, i)
fit.train()
end = time.time()
sa_results.append(ef.value(sa.getOptimal()))
sa_times.append(end - start)
#print "SA: " + str(ef.value(sa.getOptimal()))
ga_results = []
ga_times = []
for i in iters:
print(i)
for j in range(num_repeats):
start = time.time()
ga = StandardGeneticAlgorithm(200, 150, 25, gap)
fit = FixedIterationTrainer(ga, i)
fit.train()
end = time.time()
ga_results.append(ef.value(sa.getOptimal()))
ga_times.append(end - start)
#print "GA: " + str(ef.value(ga.getOptimal()))
mimic_results = []
mimic_times = []
for i in iters[0:6]:
print(i)
for j in range(num_repeats):
start = time.time()
mimic = MIMIC(200, 100, pop)
fit = FixedIterationTrainer(mimic, i)
fit.train()
end = time.time()
mimic_results.append(ef.value(mimic.getOptimal()))
mimic_times.append(end - start)
#print "MIMIC: " + str(ef.value(mimic.getOptimal()))
with open('knapsack.csv', 'w') as csvfile:
writer = csv.writer(csvfile)
writer.writerow(rhc_results)
writer.writerow(rhc_times)
writer.writerow(sa_results)
writer.writerow(sa_times)
writer.writerow(ga_results)
writer.writerow(ga_times)
writer.writerow(mimic_results)
writer.writerow(mimic_times)
return rhc_results, rhc_times, sa_results, sa_times, ga_results, ga_times, mimic_results, mimic_times
# Problem Definition
ef = TravelingSalesmanRouteEvaluationFunction(points)
odd = DiscretePermutationDistribution(N)
nf = SwapNeighbor()
mf = SwapMutation()
cf = TravelingSalesmanCrossOver(ef)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
ef2 = TravelingSalesmanSortEvaluationFunction(points)
fill = [N] * N
ranges = array('i', fill)
odd2 = DiscreteUniformDistribution(ranges)
df = DiscreteDependencyTree(.1, ranges)
pop = GenericProbabilisticOptimizationProblem(ef2, odd2, df)
# Algorithm declaration
rhc = RandomizedHillClimbing(hcp)
sa = SimulatedAnnealing(SA_TEMPERATURE, SA_COOLING_FACTOR, hcp)
ga = StandardGeneticAlgorithm(GA_POPULATION, GA_CROSSOVER, GA_MUTATION,
gap)
mimic = MIMIC(MIMIC_SAMPLES, MIMIC_TO_KEEP, pop)
# Trainer declaration
fit_rhc = FixedIterationTrainer(rhc, current_iteration_count)
fit_sa = FixedIterationTrainer(sa, current_iteration_count)
fit_ga = FixedIterationTrainer(ga, current_iteration_count)
fit_mimic = FixedIterationTrainer(mimic, current_iteration_count)
print("Computing for %d iterations" % current_iteration_count)
N = 1000
maxIters = 3001
numTrials = 5
fill = [2] * N
ranges = array('i', fill)
outfile = OUTPUT_DIRECTORY + '/FLIPFLOP/FLIPFLOP_{}_{}_LOG.csv'
ef = FlipFlopEvaluationFunction()
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mf = DiscreteChangeOneMutation(ranges)
cf = SingleCrossOver()
df = DiscreteDependencyTree(.1, ranges)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
# RHC
for t in range(numTrials):
fname = outfile.format('RHC', str(t + 1))
with open(fname, 'w') as f:
f.write('iterations,fitness,time,fevals\n')
ef = FlipFlopEvaluationFunction()
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
hcp = GenericHillClimbingProblem(ef, odd, nf)
rhc = RandomizedHillClimbing(hcp)
fit = FixedIterationTrainer(rhc, 10)
times = [0]
for i in range(0, maxIters, 10):
start = clock()
def knapsackfunc(NUM_ITEMS, iterations):
rhcMult = 600
saMult = 600
gaMult = 4
mimicMult = 3
# Random number generator */
random = Random()
# The number of items
#NUM_ITEMS = 40
# The number of copies each
COPIES_EACH = 4
# The maximum weight for a single element
MAX_WEIGHT = 50
# The maximum volume for a single element
MAX_VOLUME = 50
# The volume of the knapsack
KNAPSACK_VOLUME = MAX_VOLUME * NUM_ITEMS * COPIES_EACH * .4
# create copies
fill = [COPIES_EACH] * NUM_ITEMS
copies = array('i', fill)
# create weights and volumes
fill = [0] * NUM_ITEMS
weights = array('d', fill)
volumes = array('d', fill)
for i in range(0, NUM_ITEMS):
weights[i] = random.nextDouble() * MAX_WEIGHT
volumes[i] = random.nextDouble() * MAX_VOLUME
# create range
fill = [COPIES_EACH + 1] * NUM_ITEMS
ranges = array('i', fill)
ef = KnapsackEvaluationFunction(weights, volumes, KNAPSACK_VOLUME, copies)
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mf = DiscreteChangeOneMutation(ranges)
cf = UniformCrossOver()
df = DiscreteDependencyTree(.1, ranges)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
optimalOut = []
timeOut = []
evalsOut = []
for niter in iterations:
iterOptimalOut = [NUM_ITEMS, niter]
iterTimeOut = [NUM_ITEMS, niter]
iterEvals = [NUM_ITEMS, niter]
start = time.time()
rhc = RandomizedHillClimbing(hcp)
fit = FixedIterationTrainer(rhc, niter*rhcMult)
fit.train()
end = time.time()
rhcOptimal = ef.value(rhc.getOptimal())
rhcTime = end-start
print "RHC optimum: " + str(rhcOptimal)
print "RHC time: " + str(rhcTime)
iterOptimalOut.append(rhcOptimal)
iterTimeOut.append(rhcTime)
functionEvals = ef.getNumEvals()
ef.zeroEvals()
iterEvals.append(functionEvals)
start = time.time()
sa = SimulatedAnnealing(100, .95, hcp)
fit = FixedIterationTrainer(sa, niter*saMult)
fit.train()
end = time.time()
saOptimal = ef.value(sa.getOptimal())
saTime = end-start
print "SA optimum: " + str(saOptimal)
print "SA time: " + str(saTime)
iterOptimalOut.append(saOptimal)
iterTimeOut.append(saTime)
functionEvals = ef.getNumEvals()
ef.zeroEvals()
iterEvals.append(functionEvals)
start = time.time()
ga = StandardGeneticAlgorithm(200, 150, 25, gap)
fit = FixedIterationTrainer(ga, niter*gaMult)
fit.train()
end = time.time()
gaOptimal = ef.value(ga.getOptimal())
gaTime = end - start
print "GA optimum: " + str(gaOptimal)
print "GA time: " + str(gaTime)
iterOptimalOut.append(gaOptimal)
iterTimeOut.append(gaTime)
functionEvals = ef.getNumEvals()
ef.zeroEvals()
iterEvals.append(functionEvals)
start = time.time()
mimic = MIMIC(200, 100, pop)
fit = FixedIterationTrainer(mimic, niter*mimicMult)
fit.train()
end = time.time()
mimicOptimal = ef.value(mimic.getOptimal())
mimicTime = end - start
print "MIMIC optimum: " + str(mimicOptimal)
print "MIMIC time: " + str(mimicTime)
iterOptimalOut.append(mimicOptimal)
iterTimeOut.append(mimicTime)
functionEvals = ef.getNumEvals()
ef.zeroEvals()
iterEvals.append(functionEvals)
optimalOut.append(iterOptimalOut)
timeOut.append(iterTimeOut)
evalsOut.append(iterEvals)
return [optimalOut, timeOut, evalsOut]
def run_all_2(N=40,fout=None):
maxEpochs = 10**5
maxTime = 300 #5 minutes
problem = 'knapsack'
# Random number generator */
random = Random()
# The number of items
NUM_ITEMS = N
# The number of copies each
COPIES_EACH = 4
# The maximum weight for a single element
MAX_WEIGHT = 50
# The maximum volume for a single element
MAX_VOLUME = 50
# The volume of the knapsack
KNAPSACK_VOLUME = MAX_VOLUME * NUM_ITEMS * COPIES_EACH * .4
# create copies
fill = [COPIES_EACH] * NUM_ITEMS
copies = array('i', fill)
# create weights and volumes
fill = [0] * NUM_ITEMS
weights = array('d', fill)
volumes = array('d', fill)
for i in range(0, NUM_ITEMS):
weights[i] = random.nextDouble() * MAX_WEIGHT
volumes[i] = random.nextDouble() * MAX_VOLUME
# create range
fill = [COPIES_EACH + 1] * NUM_ITEMS
ranges = array('i', fill)
ef = KnapsackEvaluationFunction(weights, volumes, KNAPSACK_VOLUME, copies)
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mf = DiscreteChangeOneMutation(ranges)
cf = UniformCrossOver()
df = DiscreteDependencyTree(.1, ranges)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
def run_algo(alg,fit,label,difficulty,iters):
trainTimes = [0.]
trainTime = []
scoreChange = [0.]
stuckCount = 10**3
prev = 0.
for epoch in range(0,maxEpochs,1):
st = time.clock()
fit.train()
et = time.clock()
trainTimes.append(trainTimes[-1]+(et-st))
trainTime.append((et-st))
rollingMean = 10
avgTime = (math.fsum(trainTime[-rollingMean:]) / float(rollingMean))
score = ef.value(alg.getOptimal())
# trialString = '{}-{}-{}-{}'.format(label,score,epoch,trainTimes[-1])
trialData = [problem,difficulty,label,score,epoch,trainTimes[-1],avgTime,iters]
# print(trialData)
# fout.writerow(trialData)
# print(trialData)
print(trialData,max(scoreChange))
# print(max(scoreChange))
# optimum = (difficulty-1-T) + difficulty
# if score >= optimum: break
scoreChange.append(abs(score-prev))
prev = score
scoreChange = scoreChange[-stuckCount:]
# print(scoreChange)
if max(scoreChange) < 1.0: break
if trainTimes[-1] > maxTime: break
# print(trialData)
fout.writerow(trialData)
iters = 1000
rhc = RandomizedHillClimbing(hcp)
fit = FixedIterationTrainer(rhc, iters)
run_algo(rhc,fit,'RHC',N,iters)
iters = 1000
startTemp = 1E10
coolingFactor = .99
sa = SimulatedAnnealing(startTemp, coolingFactor, hcp)
fit = FixedIterationTrainer(sa, iters)
run_algo(sa,fit,'SA',N,iters)
iters = 10
population = 300
mates = 100
mutations = 50
ga = StandardGeneticAlgorithm(population, mates, mutations, gap)
fit = FixedIterationTrainer(ga, iters)
run_algo(ga,fit,'GA',N,iters)
iters = 10
samples = 200
keep = 20
mimic = MIMIC(samples, keep, pop)
fit = FixedIterationTrainer(mimic, iters)
run_algo(mimic,fit,'MIMIC',N,iters)
def run_count_ones_experiments():
OUTPUT_DIRECTORY = './output'
N = 80
fill = [2] * N
ranges = array('i', fill)
ef = CountOnesEvaluationFunction()
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mf = DiscreteChangeOneMutation(ranges)
cf = SingleCrossOver()
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
max_iter = 5000
outfile = OUTPUT_DIRECTORY + '/count_ones_{}_log.csv'
# Randomized Hill Climber
filename = outfile.format('rhc')
with open(filename, 'w') as f:
f.write('iteration,fitness,time\n')
for it in range(0, max_iter, 10):
rhc = RandomizedHillClimbing(hcp)
fit = FixedIterationTrainer(rhc, it)
start_time = time.clock()
fit.train()
elapsed_time = time.clock() - start_time
# fevals = ef.fevals
score = ef.value(rhc.getOptimal())
data = '{},{},{}\n'.format(it, score, elapsed_time)
print(data)
with open(filename, 'a') as f:
f.write(data)
# Simulated Annealing
filename = outfile.format('sa')
with open(filename, 'w') as f:
f.write('iteration,cooling_value,fitness,time\n')
for cooling_value in (.19, .38, .76, .95):
for it in range(0, max_iter, 10):
sa = SimulatedAnnealing(100, cooling_value, hcp)
fit = FixedIterationTrainer(sa, it)
start_time = time.clock()
fit.train()
elapsed_time = time.clock() - start_time
# fevals = ef.fevals
score = ef.value(sa.getOptimal())
data = '{},{},{},{}\n'.format(it, cooling_value, score, elapsed_time)
print(data)
with open(filename, 'a') as f:
f.write(data)
# Genetic Algorithm
filename = outfile.format('ga')
with open(filename, 'w') as f:
f.write('iteration,population_size,to_mate,to_mutate,fitness,time\n')
for population_size, to_mate, to_mutate in itertools.product([20], [4, 8, 16, 20], [0, 2, 4, 6]):
for it in range(0, max_iter, 10):
ga = StandardGeneticAlgorithm(population_size, to_mate, to_mutate, gap)
fit = FixedIterationTrainer(ga, it)
start_time = time.clock()
fit.train()
elapsed_time = time.clock() - start_time
# fevals = ef.fevals
score = ef.value(ga.getOptimal())
data = '{},{},{},{},{},{}\n'.format(it, population_size, to_mate, to_mutate, score, elapsed_time)
print(data)
with open(filename, 'a') as f:
f.write(data)
# MIMIC
filename = outfile.format('mm')
with open(filename, 'w') as f:
f.write('iterations,samples,to_keep,m,fitness,time\n')
for samples, to_keep, m in itertools.product([50], [10], [0.1, 0.3, 0.5, 0.7, 0.9]):
for it in range(0, 500, 10):
df = DiscreteDependencyTree(m, ranges)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
mm = MIMIC(samples, 20, pop)
fit = FixedIterationTrainer(mm, it)
start_time = time.clock()
fit.train()
elapsed_time = time.clock() - start_time
# fevals = ef.fevals
score = ef.value(mm.getOptimal())
data = '{},{},{},{},{},{}\n'.format(it, samples, to_keep, m, score, elapsed_time)
print(data)
with open(filename, 'a') as f:
f.write(data)
def run_all_2(N=200,T=40,fout=None):
problem = 'flipflop'
# N=200
# T=N/10
maxEpochs = 10**5
maxTime = 300 #5 minutes
fill = [2] * N
ranges = array('i', fill)
# ef = FourPeaksEvaluationFunction(T)
ef = FlipFlopEvaluationFunction()
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mf = DiscreteChangeOneMutation(ranges)
# mf = SwapMutation()
cf = SingleCrossOver()
df = DiscreteDependencyTree(.1, ranges)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
def run_algo(alg,fit,label,difficulty,iters):
trainTimes = [0.]
trainTime = []
scoreChange = [0.]
stuckCount = 10**3
prev = 0.
for epoch in range(0,maxEpochs,1):
st = time.clock()
fit.train()
et = time.clock()
trainTimes.append(trainTimes[-1]+(et-st))
trainTime.append((et-st))
rollingMean = 10
avgTime = (math.fsum(trainTime[-rollingMean:]) / float(rollingMean))
score = ef.value(alg.getOptimal())
# print(score,difficulty)
# trialString = '{}-{}-{}-{}'.format(label,score,epoch,trainTimes[-1])
trialData = [problem,difficulty,label,score,epoch,trainTimes[-1],avgTime,iters]
print(trialData)
# fout.writerow(trialData)
optimum = (difficulty-1)
if score >= optimum: break
scoreChange.append(abs(score-prev))
prev = score
scoreChange = scoreChange[-stuckCount:]
# print(scoreChange)
if max(scoreChange) == 0: break
if trainTimes[-1] > maxTime: break
# print(trialData)
fout.writerow(trialData)
iters = 10
rhc = RandomizedHillClimbing(hcp)
fit = FixedIterationTrainer(rhc, iters)
run_algo(rhc,fit,'RHC',N,iters)
iters = 10
startTemp = 1E10
coolingFactor = .99
sa = SimulatedAnnealing(startTemp, coolingFactor, hcp)
fit = FixedIterationTrainer(sa, iters)
run_algo(sa,fit,'SA',N,iters)
iters = 10
population = 300
mates = 40
mutations = 10
ga = StandardGeneticAlgorithm(population, mates, mutations, gap)
fit = FixedIterationTrainer(ga, iters)
run_algo(ga,fit,'GA',N,iters)
iters = 10
samples = 200
keep = 5
mimic = MIMIC(samples, keep, pop)
fit = FixedIterationTrainer(mimic, iters)
run_algo(mimic,fit,'MIMIC',N,iters)
def run_algorithm_test(T,
ranges,
algorithms,
output_file_name,
trial_number,
iterations=False):
with open(output_file_name, 'w') as f:
f.write('algorithm,optimal_result,iterations,time,trial\n')
ef = FourPeaksEvaluationFunction(T)
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mf = DiscreteChangeOneMutation(ranges)
cf = SingleCrossOver()
df = DiscreteDependencyTree(.1, ranges)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
for trial in range(trial_number):
if iterations is False:
for item in algorithms:
start_time = time.time()
if item in ['rhc']:
optimal_result, run_iters = run_rhc(hcp, ef)
elif item in ['sa']:
optimal_result, run_iters = run_sa(hcp, ef)
elif item in ['ga']:
optimal_result, run_iters = run_ga(gap, ef)
elif item in ['mimic']:
optimal_result, run_iters = run_mimic(pop, ef)
else:
print "The algorithm type {} is not supported.".format(
item)
end_time = time.time()
time_elapsed = end_time - start_time
run_output = '{},{},{},{},{}\n'.format(item, optimal_result,
run_iters, time_elapsed,
trial)
with open(output_file_name, 'a') as f:
f.write(run_output)
else:
for iter in iterations:
for item in algorithms:
start_time = time.time()
if item in ['rhc']:
optimal_result, run_iters = run_rhc(hcp, ef, iter)
elif item in ['sa']:
optimal_result, run_iters = run_sa(hcp, ef, iter)
elif item in ['ga']:
optimal_result, run_iters = run_ga(gap, ef, iter)
elif item in ['mimic']:
optimal_result, run_iters = run_mimic(pop, ef, iter)
else:
print "The algorithm type {} is not supported.".format(
item)
end_time = time.time()
time_elapsed = end_time - start_time
run_output = '{},{},{},{},{}\n'.format(
item, optimal_result, run_iters, time_elapsed, trial)
with open(output_file_name, 'a') as f:
f.write(run_output)
print "time elapsed is {}".format(time_elapsed)
return
def solveit(oaname, params):
iterations = 10000
tryi = 1
# Random number generator */
random = Random()
# The number of items
NUM_ITEMS = 40
# The number of copies each
COPIES_EACH = 4
# The maximum weight for a single element
MAX_WEIGHT = 50
# The maximum volume for a single element
MAX_VOLUME = 50
# The volume of the knapsack
KNAPSACK_VOLUME = MAX_VOLUME * NUM_ITEMS * COPIES_EACH * .4
# create copies
fill = [COPIES_EACH] * NUM_ITEMS
copies = array('i', fill)
# create weights and volumes
fill = [0] * NUM_ITEMS
weights = array('d', fill)
volumes = array('d', fill)
for i in range(0, NUM_ITEMS):
weights[i] = random.nextDouble() * MAX_WEIGHT
volumes[i] = random.nextDouble() * MAX_VOLUME
# create range
fill = [COPIES_EACH + 1] * NUM_ITEMS
ranges = array('i', fill)
ef = KnapsackEvaluationFunction(weights, volumes, KNAPSACK_VOLUME, copies)
odd = DiscreteUniformDistribution(ranges)
nf = DiscreteChangeOneNeighbor(ranges)
mf = DiscreteChangeOneMutation(ranges)
cf = UniformCrossOver()
df = DiscreteDependencyTree(.1, ranges)
hcp = GenericHillClimbingProblem(ef, odd, nf)
gap = GenericGeneticAlgorithmProblem(ef, odd, mf, cf)
pop = GenericProbabilisticOptimizationProblem(ef, odd, df)
if oaname == 'RHC':
iterations = int(params[0])
tryi = int(params[1])
oa = RandomizedHillClimbing(hcp)
if oaname == 'SA':
oa = SimulatedAnnealing(float(params[0]), float(params[1]), hcp)
if oaname == 'GA':
iterations = 1000
oa = StandardGeneticAlgorithm(int(params[0]), int(params[1]),
int(params[2]), gap)
if oaname == 'MMC':
iterations = 1000
oa = MIMIC(int(params[0]), int(params[1]), pop)
print "Running %s using %s for %d iterations, try %d" % (
oaname, ','.join(params), iterations, tryi)
print "=" * 20
starttime = timeit.default_timer()
output = []
for i in range(iterations):
oa.train()
if i % 10 == 0:
optimal = oa.getOptimal()
score = ef.value(optimal)
elapsed = int(timeit.default_timer() - starttime)
output.append([str(i), str(score), str(elapsed)])
print 'score: %.3f' % score
print 'train time: %d secs' % (int(timeit.default_timer() - starttime))
scsv = 'kn-%s-%s.csv' % (oaname, '-'.join(params))
print "Saving to %s" % (scsv),
with open(scsv, 'w') as csvf:
writer = csv.writer(csvf)
for row in output:
writer.writerow(row)
print "saved."
print "=" * 20
