Source code for iwopy.benchmarks.rosenbrock.rosenbrock

from iwopy import SimpleProblem, SimpleObjective



[docs]
class RosenbrockObjective(SimpleObjective):
    """
    The Rosenbrock function is defined as

    f(x,y) = (a-x)^2 + b(y-x^2)^2

    Recommended values for the parameters are:
    a = 1
    b = 100

    Domain:
    x = [-inf, inf]
    y = [-inf, inf]

    The unconstraint Rosenbrock function has a global minima at

    (x,y) = (1,1)

    with a function value of

    f(x,y) = 0

    :group: benchmarks.rosenbrock

    """


[docs]
    def __init__(
        self,
        problem,
        pars=(1.0, 100.0),
        ana_deriv=False,
        name="f",
    ):
        """
        Construtor

        Parameters
        ----------
        problem: iwopy.Problem
            The underlying optimization problem
        pars: tuple
            The a, b parameters
        ana_deriv: bool
            Switch for analytical derivatives
        name: str
            The function name

        """
        super().__init__(problem, name, n_components=1, has_ana_derivs=ana_deriv)

        # (a, b)
        self._pars = pars



[docs]
    def f(self, x, y):
        """
        The Rosenbrock function f(x, y)
        """
        a, b = self._pars
        return (a - x) ** 2 + b * (y - x**2) ** 2



[docs]
    def g(self, var, x, y, components=None):
        """
        The derivative of the Rosenbrock function
        """
        a, b = self._pars
        if var == 0:
            return -2 * (a - x) - 2 * b * (y - x**2) * 2 * x
        else:
            return 2 * b * (y - x**2)





[docs]
class RosenbrockProblem(SimpleProblem):
    """
    Problem definition of benchmark function Rosenbrock.

    Attributes
    ----------
    initial_values: list of float
        The initial values

    :group: benchmarks.rosenbrock

    """


[docs]
    def __init__(
        self,
        lower=[-5.0, -5.0],
        upper=[10.0, 10.0],
        initial=[0.0, 0.0],
        ana_deriv=False,
        name="rosenbrock",
    ):
        """
        Constructor

        Parameters
        ----------
        lower: list of float
            The minimal variable values
        upper: list of float
            The maximal variable values
        initial: list of float
            The initial values
        ana_deriv: bool
            Switch for analytical derivatives
        name: str
            The name of the problem

        """
        super().__init__(
            name,
            float_vars={"x": initial[0], "y": initial[1]},
            min_values_float={"x": lower[0], "y": lower[1]},
            max_values_float={"x": upper[0], "y": upper[1]},
        )

        self.add_objective(RosenbrockObjective(self, ana_deriv=ana_deriv))