_sda.py 文件源码

def __init__(self, fun, x0, bounds, seed=None, minimizer_kwargs=None,
                 temperature_start=5230, qv=2.62, qa=-5.0,
                 maxfun=1e7, maxsteps=500, pure_sa=False):
        if x0 is not None and not len(x0) == len(bounds):
            raise ValueError('Bounds size does not match x0')
        lu = list(zip(*bounds))
        lower = np.array(lu[0])
        upper = np.array(lu[1])
        # Checking that bounds are consistent
        if not np.all(lower < upper):
            raise ValueError('Bounds are note consistent min < max')
        # Wrapper for the objective function and minimizer
        if minimizer_kwargs is None:
            minimizer_kwargs = dict()
        self.owf = ObjectiveFunWrapper(bounds, fun, **minimizer_kwargs)
        # Initialization of RandomState for reproducible runs if seed provided
        self.rs = check_random_state(seed)
        # Initialization of the energy state
        self.es = EnergyState(lower, upper)
        self.es.reset(self.owf, self.rs, x0)
        # Maximum number of function call that can be used a stopping criterion
        self.maxfun = maxfun
        # Maximum number of step (main iteration)  that can be used as
        # stopping criterion
        self.maxsteps = maxsteps
        # Minimum value of annealing temperature reached to perform
        # re-annealing
        self.temperature_start = temperature_start
        self.temperature_restart = 0.1
        # VisitingDistribution instance
        vd = VisitingDistribution(lower, upper, qv, self.rs)
        # Markov chain instance
        self.mc = MarkovChain(qa, vd, self.owf, self.rs, self.es)
        self.qv = qv
        self.pure_sa = pure_sa