def crossover_weights(ind1, ind2, force):
#ind1.check()
#ind2.check()
child1, child2 = tools.cxUniform(ind1.weights, ind2.weights, indpb=0.5)
ind1.weights[:] = fix_weights(child1, force)
ind2.weights[:] = fix_weights(child2, force)
#ind1.check()
#ind2.check()
return ind1, ind2
def crossover_weights(ind1, ind2, force):
#ind1.check()
#ind2.check()
child1, child2 = tools.cxUniform(ind1.weights.flatten(),
ind2.weights.flatten(),
indpb=0.5)
ind1.weights[:, :] = child1.reshape((ind1.NC, ind1.ND))
ind2.weights[:, :] = child2.reshape((ind2.NC, ind2.ND))
NC, ND = ind1.weights.shape
for i in range(NC):
ind1.weights[i, :] = fix_weights(ind1.weights[i, :], force)
ind2.weights[i, :] = fix_weights(ind2.weights[i, :], force)
#ind1.check()
#ind2.check()
return ind1, ind2
def mutate_weights(ind, force):
new_solution = ind.weights.copy()
ND = len(new_solution)
#select at random a dimension to modify
sel_dim = np.random.randint(ND)
new_solution[sel_dim] += np.random.normal(loc=0.0, scale=0.1)
new_solution[sel_dim] = np.clip(new_solution[sel_dim], 0.0000001, 1.0)
new_solution[:] = fix_weights(new_solution[:], force)
ind.weights[:] = new_solution
return ind,
def mutate_weights(ind, force):
new_solution = ind.weights.copy()
NC, ND = new_solution.shape
#select at random a cluster to modify
sel_cluster = np.random.randint(NC)
#select at random a dimension to modify
sel_dim = np.random.randint(ND)
new_solution[sel_cluster, sel_dim] += np.random.normal(loc=0.0, scale=0.1)
new_solution[sel_cluster,
sel_dim] = np.clip(new_solution[sel_cluster, sel_dim],
0.0000001, 1.0)
new_solution[sel_cluster, :] = fix_weights(new_solution[sel_cluster, :],
force)
ind.weights[:, :] = new_solution
return ind,
cl.distances.set_categorical(1, 3, distances_cat)
#cl.distances.set_targeted(23,targets,False)
#force = (22,0.15)
force = None
#initial centroids at random
indices = np.random.choice(N, size=NC, replace=False)
current_centroids = np.asfortranarray(np.empty((NC, ND)))
current_centroids[:, :] = data[indices, :]
#initial weights are uniform
weights = np.asfortranarray(np.ones((NC, ND), dtype=np.float32) / ND)
for c in range(NC):
weights[c, :] = fix_weights(weights[c, :], force=force)
for k in range(100):
best_centroids, best_u, best_energy_centroids, best_jm, current_temperature, evals = clustering_ga.optimize_centroids(
data,
current_centroids,
weights,
m,
lambda_value,
var_types, {},
ngen=ngen,
npop=npop,
cxpb=cxpb,
mutpb=mutpb,
stop_after=stop_after,
min_values=min_values,
def evolve_weights(toolbox,
centroids,
initial_weights,
values,
ngen,
npop,
stop_after,
cxpb,
mutpb,
force,
verbose=False):
NC, ND = centroids.shape
N = len(values)
pop = [WeightIndividual(NC, ND, None) for _ in range(npop)]
for i, ind in enumerate(pop):
if i == 0:
ind.weights[:, :] = initial_weights
else:
#random
for k in range(NC):
ind.weights[k, :] = np.random.random(ND)
ind.weights[k, :] /= np.sum(ind.weights[k, :])
ind.weights[k, :] = fix_weights(ind.weights[k, :], force=force)
ind.check()
hof = tools.HallOfFame(1, similar=similar_op_weights)
stats = tools.Statistics(key=lambda ind: ind.fitness.values)
stats.register("avg", np.mean)
stats.register("min", np.min)
stats.register("max", np.max)
logbook = tools.Logbook()
logbook.header = "gen", "min", "avg", "max", "best"
#logger.info("Evaluating initial population")
# Evaluate the individuals with an invalid fitness
invalid_ind = [ind for ind in pop if not ind.fitness.valid]
fitnesses = toolbox.map(toolbox.evaluate, invalid_ind)
for ind, fit in zip(invalid_ind, fitnesses):
ind.fitness.values = fit
hof.update(pop)
record = stats.compile(pop) if stats else {}
logbook.record(gen=0, best=hof[0].fitness.values[0], **record)
if verbose: print(logbook.stream)
evaluations = 0
no_improvements = 0
#logger.info("Starting evolution!")
for gen in range(1, ngen + 1):
prev_fitness = hof[0].fitness.values[0]
# Select the next generation individuals
offspring = toolbox.select(pop, len(pop))
# Vary the pool of individuals
offspring = algorithms.varAnd(offspring, toolbox, cxpb, mutpb)
# Evaluate the individuals with an invalid fitness
invalid_ind = [ind for ind in offspring if not ind.fitness.valid]
fitnesses = toolbox.map(toolbox.evaluate, invalid_ind)
for ind, fit in zip(invalid_ind, fitnesses):
ind.fitness.values = fit
evals_gen = len(invalid_ind)
# Update the hall of fame with the generated individuals
if hof is not None:
hof.update(offspring)
# Replace the current population by the offspring
pop[:] = offspring
# Append the current generation statistics to the logbook
record = stats.compile(pop) if stats else {}
logbook.record(gen=gen, best=hof[0].fitness.values[0], **record)
#logger.info(logbook.stream)
if verbose: print(logbook.stream)
evaluations += evals_gen
current_fitness = hof[0].fitness.values[0]
if current_fitness >= prev_fitness:
no_improvements += 1
else:
no_improvements = 0
if no_improvements > stop_after:
break
return pop, stats, hof, logbook, gen, evaluations