def __init__(self, algorithm='qlearning', policy="eps_greedy", nA=10, nS=4**4, lvfa=False, feature_size=13, alpha=0.01, gamma=0.99, eligibility=True, episodes=1000, load_model=None, **kwargs): self.algorithm = Algorithm(algorithm=algorithm, nS=nS, nA=nA, lvfa=lvfa, feature_size=feature_size, eligibility=eligibility, alpha=alpha, gamma=gamma, **kwargs) self.policy = Policy(policy=policy, episodes=episodes, nA=nA, lvfa=lvfa, **kwargs)
from models import Elit from utils.converter import Converter from utils.operators.basic import Operator as op from utils.operators.modified import Operator as op_m from utils.policies import Policy from utils.test_functions import levi import numpy as np converter = Converter(type_='grey') crossover = op.crossover() mutator = op_m.Mutate.multi(3) policies = { 'include': Policy.elitarium(), 'exclude': Policy.elitarium(), 'parents': Policy.random() } n = 100 eps = 1 model = Elit(converter=converter, policies=policies, mutator=mutator, crossover=crossover) model.run(*(levi()), epochs=101, n=20, eps=eps, optimize='min', t=10,