def do_training():
print(genom_struct_path, MEMS_info, cuncurrency)
ga = GeneticAlgorithm(genom_struct_path)
start_time = time.time()
population = ga.run(init_population_size,
population_size,
mutation_rate,
num_iteratitions,
crossover_type,
calc_fitness,
fitness_goal,
cuncurrency=cuncurrency,
reverse_fitness_order=True,
population_np_path=population_np_path,
reload_np_population_rate=reload_np_population_rate)
save_models(population)
end_time = time.time()
print(population[:3].astype(float))
print(population[:, -1].astype(float))
print('Runtime :', end_time - start_time)
global logger
logging.config.dictConfig(yaml.load(open('logging.yaml')))
logger = logging.getLogger(GeneticAlgorithm.LOGGER_HANDLER_NAME)
path = 'genom_struct.csv'
init_population_size = 6000
population_size = 100
mutation_rate = 0.20
num_iteratitions = 100
crossover_type = GeneticAlgorithm.TWO_POINT_CROSSOVER
fitness_goal = 0.00001
load_dataset()
ga = GeneticAlgorithm(path)
start_time = time.time()
population = ga.run(init_population_size,
population_size,
mutation_rate,
num_iteratitions,
crossover_type,
calc_fitness,
fitness_goal,
cuncurrency=20,
reverse_fitness_order=False)
save_models(population)
end_time = time.time()
print(population[:3].astype(float))
print(population[:, -1].astype(float))
print('Runtime :', end_time - start_time)
# Genetic Algorithm with Configuration
ga = GeneticAlgorithm(
SalesTerritories(salesman, clients,
dist_matrix, priority_matrix, initial_territory,
limit=args.generations,
size=args.init_population,
prob_crossover=args.prob_crossover,
prob_mutation=args.prob_mutation,
local_search=args.local_search,
elitism=args.elitism,
params=params)
)
log("Init Otimization...\n\n")
population = ga.run(args.processes)
# Best Solution
best_territory = ga.best_solution
log("Best Solution: "+str(ga.genetics.fitness(best_territory)))
# History
create_history_df(ga.avg_fitness).to_csv("{}/avg_fitness.csv".format(args.output))
create_history_df(ga.best_fitness).to_csv("{}/best_fitness.csv".format(args.output))
#for k in ga.genetics.fn.keys():
# create_history_df(ga.genetics.fn[k]).to_csv("{}/{}_fitness.csv".format(args.output, k))
# Summary
chromo_summary = ChromoSummary(salesman, clients, dist_matrix, priority_matrix)