def test_rank(fitness_dict, expected_results):
factory = ChromosomeFactory.ChromosomeRangeFactory(noOfGenes=4,
minValue=1,
maxValue=4)
ga = GAEngine.GAEngine(factory, 2, fitness_type=('equal', 4), mut_prob=0.3)
ga.population = Population.Population(factory, 2)
assert type((Utils.SelectionHandlers.rank(fitness_dict, ga))) == list
def test_false_population_control():
factory = ChromosomeFactory.ChromosomeRangeFactory(noOfGenes=8,minValue=1,maxValue=100)
ga = GAEngine.GAEngine(factory,population_size=100,fitness_type='max',mut_prob = 0.3,population_control = False)
ga.addCrossoverHandler(Utils.CrossoverHandlers.distinct, 4)
ga.addMutationHandler(Utils.MutationHandlers.swap)
ga.setSelectionHandler(select_all)
ga.setFitnessHandler(lambda x:sum(x))
ga.evolve(1)
assert len(ga.population.members) > 100
def evolve(self, noOfGenerations=100):
ga = GAEngine.GAEngine(self.factory,
population_size=self.population_size,
fitness_type='min')
ga.setFitnessHandler(ANNTopologyEvolve.fitness, self.X, self.Y,
self.hiddenLayers, self.input_dim, self.loss,
self.metrics, self.epochs, self.batch_size)
ga.setSelectionHandler(Utils.SelectionHandlers.best)
ga.addCrossoverHandler(Utils.CrossoverHandlers.onePoint, 1)
ga.addMutationHandler(ANNTopologyEvolve.mutation, 1,
self.neuronsPerLayer, self.activations,
self.optimizers)
ga.evolve(noOfGenerations)
from pygenetic import GAEngine, ChromosomeFactory, Utils
import matplotlib.pyplot as plt
matrix = [[0,172,145,607,329,72,312,120],[172,0,192,494,209,158,216,92],[145,192,0,490,237,75,205,100],[607,494,490,0,286,545,296,489],[329,209,237,286,0,421,49,208],[72,158,75,545,421,0,249,75],[312,216,205,296,49,249,9,194],[120,92,100,489,208,75,194,0]]
# best sequence i found: 0 5 2 7 1 6 4 3
factory = ChromosomeFactory.ChromosomeRangeFactory(noOfGenes=8,minValue=0,maxValue=7)
ga = GAEngine.GAEngine(factory,10,fitness_type='min',mut_prob = 0.02)
ga.addCrossoverHandler(Utils.CrossoverHandlers.PMX, 9)
def range_of_generation(fitness_mappings,ga):
return abs(fitness_mappings[0][1] - fitness_mappings[-1][1])
ga.addStatistic('range',range_of_generation)
ga.addCrossoverHandler(Utils.CrossoverHandlers.distinct, 4)
ga.addCrossoverHandler(Utils.CrossoverHandlers.OX, 3)
ga.addMutationHandler(Utils.MutationHandlers.swap)
ga.setSelectionHandler(Utils.SelectionHandlers.random)
ga.setFitnessHandler(Utils.Fitness.TSP, matrix)
ga.evolve(20)
fig = ga.statistics.plot_statistics(['range'])
plt.show()
fig = ga.statistics.plot()
plt.show()
fitness = 0
for i in range(len(board)):
isSafe = True
for j in range(len(board)):
if i != j:
if (board[i] == board[j]) or (abs(board[i] - board[j])
== abs(i - j)):
isSafe = False
break
if (isSafe == True):
fitness += 1
return fitness
factory = ChromosomeFactory.ChromosomeRangeFactory(noOfGenes=8,
minValue=1,
maxValue=8)
ga = GAEngine.GAEngine(factory, 100, fitness_type=('equal', 8), mut_prob=0.3)
ga.addCrossoverHandler(Utils.CrossoverHandlers.distinct, 4)
ga.addMutationHandler(Utils.MutationHandlers.swap)
ga.setSelectionHandler(Utils.SelectionHandlers.best)
ga.setFitnessHandler(fitness)
ga.evolve(10)
fig = ga.statistics.plot_statistics(
['best-fitness', 'worst-fitness', 'avg-fitness'])
plt.show()
fig = ga.statistics.plot_statistics(['diversity', 'mutation_rate'])
plt.show()