def simulation_random_walk(params):
STPG = read_problem("datasets", "ORLibrary", params["dataset"])
crossover = CrossoverRandomWalkRST(STPG)
evaluation = EvaluateEdgeSet(STPG)
mutate = MutationReplaceByRandomEdge(STPG)
output_data_dir = os.path.join("data", "test", "edgeset", STPG.name)
tracker = DataTracker(params['runtrial'], target=output_data_dir)
population = (GPopulation(
chromosomes=[
gen_random_walk(STPG) for _ in range(params["population_size"])
],
eval_function=evaluation,
maximize=True).evaluate().normalize(
norm_function=normalize).callback(update_best))
evol = (Evolution().evaluate().normalize(
norm_function=normalize).callback(update_best).callback(
tracker.log_evaluation).select(selection_func=roullete).crossover(
combiner=crossover).mutate(
mutate_function=mutate,
probability=params['tx_mutation']).callback(
update_generation).callback(display, every=100))
with Stagnation(interval=params["stagnation_interval"]), \
BestKnownReached(global_optimum=params['global_optimum']):
result = population.evolve(evol, n=params["n_iterations"])
tracker.log_simulation(params, STPG, result)
tracker.report()
print(result.stoppedby)
def simulation(simulation_name, params):
datasets_folder = path.join("datasets", "ORLibrary")
filename = path.join(datasets_folder, params["dataset"])
STPG = ReaderORLibrary().parser(filename)
print("STPG information", '\n', 10 * '- ', '\n')
print("Trial: ", parameters['runtrial'])
print('Instance: ', STPG.name)
print('Best Known cost: ', params['global_optimum'])
print("Nro. Node:", STPG.nro_nodes)
print("Nro. Edges:", STPG.nro_edges)
print("Nro. Terminals:", STPG.nro_terminals)
# print("Terminals: \n", STPG.terminals)
output_folder = path.join("data", simulation_name, STPG.name)
tracker = DataTracker(params['runtrial'], target=output_folder)
generator = GenerateBasedPrimRST(STPG)
evaluator = EvaluateTreeGraph(STPG)
crossover = CrossoverPrimRST(STPG)
prunner = Prunning(STPG)
mut_prim = PrimBasedMutation(STPG)
replace_random = ReplaceByRandomEdge(STPG)
population = (GPopulation(
chromosomes=[generator() for _ in range(params["population_size"])],
eval_function=evaluator,
maximize=True).evaluate().normalize(
norm_function=normalize).callback(update_best))
evol = (Evolution().evaluate().normalize(norm_function=normalize).callback(
update_best).callback(tracker.log_evaluation).select(
selection_func=roullete).crossover(combiner=crossover).mutate(
mutate_function=replace_random, probability=0.2).mutate(
mutate_function=mut_prim, probability=0.2).mutate(
mutate_function=prunner,
probability=0.2).callback(update_generation).callback(
display, every=100))
with Stagnation(interval=params["stagnation_interval"]), \
BestKnownReached(global_optimum=params['global_optimum']):
result = population.evolve(evol, n=params["n_iterations"])
tracker.log_simulation(params, STPG, result)
best_overall = result.documented_best
test, response = is_steiner_tree(best_overall.chromosome, STPG)
tracker.log_bestIndividual(best_overall, test, response)
tracker.report()
def simulation(simulation_name, params: dict, get_evol: callable):
STPG = read_problem("datasets", "ORLibrary", params["dataset"])
lenght = STPG.nro_nodes - STPG.nro_terminals
tracker = DataTracker(params['runtrial'],
target=os.path.join("data", simulation_name,
STPG.name))
population = (GPopulation(
chromosomes=[
random_binary(lenght) for _ in range(params["population_size"])
],
eval_function=EvaluateKruskalBased(STPG,
penality_function=lambda k:
(k - 1) * 100),
maximize=True).evaluate().normalize(
norm_function=normalize).callback(update_best))
evol = get_evol(STPG, tracker, params)
# with Stagnation(interval=params["stagnation_interval"]), \
# with BestSteinerTreeReachead(global_optimum=params["global_optimum"],
# STPG=STPG,
# decoder=Coder(STPG).binary2treegraph,
# ):
with Stagnation(interval=params["stagnation_interval"]), \
BestKnownReached(global_optimum=params['global_optimum']):
result = population.evolve(evol, n=params["n_iterations"])
tracker.log_simulation(params, STPG, result)
tracker.report()
print(result.stoppedby)
return result