def test_qap(n = 5, i = 0):
total_ind = 10
total_generations = 100
max_pareto_size = 20
op = GeneticOperators()
instancias = parse_qap()
flux_mats = instancias[i][:-1]
dist_mat = instancias[i][-1]
num_loc = len(flux_mats[0]) #nro de localidades
objs = []
for cost_mat in flux_mats:
objs.append(QAPObjectiveFunction(dist_mat, cost_mat))
spea = SPEA(len(objs), op, max_pareto_size)
pareto_set = ParetoSet(None)
for i in xrange(n):
pop = []
for i in xrange(total_ind):
sol = range(num_loc)
random.shuffle(sol)
pop.append(GaSolution(sol, objs))
spea.run(pop, total_generations)
pareto_set.update(pop)
pareto_front = ParetoFront(pareto_set)
pareto_front.draw()
return pareto_set
def run(self, P, num_generations):
"""
Ejecuta el algoritmo SPEA
@param P: la poblacion inicial
@param num_generations: el numero maximo de generaciones
"""
ps = ParetoSet()
for i in xrange(num_generations):
ps.update(P)
for s in ps.solutions:
if s in P:
P.remove(s)
if len(ps.solutions) > self.max_pareto_points:
self.reduce_pareto_set(ps)
self.fitness_assignment(ps, P)
mating_pool = self.selection(P, ps)
P = self.next_generation(mating_pool, len(P))
def test_tsp(n = 5, i = 0):
total_ind = 10
total_generations = 100
max_pareto_size = 20
op = GeneticOperators()
instancias = parse_tsp()
cost_mats = instancias[i]
objs = []
for cost_mat in cost_mats:
objs.append(TSPObjectiveFunction(cost_mat))
num_cities = len(objs[0].mat)
spea = SPEA(len(objs), op, max_pareto_size)
pareto_set = ParetoSet(None)
for i in xrange(n):
pop = []
for i in xrange(total_ind):
sol = range(num_cities)
random.shuffle(sol)
pop.append(GaSolution(sol, objs))
spea.run(pop, 100)
pareto_set.update(pop)
pareto_front = ParetoFront(pareto_set)
pareto_front.draw()
return pareto_set
