import numpy as np
from iwopy import SimpleProblem, SimpleObjective
[docs]class BraninObjective(SimpleObjective):
"""
The objective function for the Branin problem.
The Branin (or Branin-Hoo) function is defined as
f(x,y) = a(y-bx^2+cx-r)^2 + s(1-t)cos(x)+s
Recommended values for the parameters are:
a = 1
b = 5.1/(4*pi^2)
c = 5/pi
r = 6
s = 10
t = 1/(8*pi)
Domain:
x = [-5, 10]
y = [0, 15]
The Branin function has three global minima at
(x,y) = (-pi, 12.275), (pi, 2.275), (9.42478, 2.475)
with a function value of
f(x,y) = 0.397887
Parameters
----------
problem: iwopy.Problem
The underlying optimization problem
ana_deriv : bool
Switch for analytical derivatives
name: str
The function name
"""
[docs] def __init__(self, problem, ana_deriv=False, name="f"):
super().__init__(problem, name, n_components=1, has_ana_derivs=ana_deriv)
# Parameters of branin function,
# (a, b, c, r, s, t)
self._pars = (
1,
5.1 / (4 * np.pi**2),
5 / np.pi,
6,
10,
1 / (8 * np.pi),
)
self._ana_deriv = ana_deriv
[docs] def f(self, x, y):
"""
The Branin function f(x, y)
"""
a, b, c, r, s, t = self._pars
return a * (y - b * x**2 + c * x - r) ** 2 + s * (1 - t) * np.cos(x) + s
[docs] def g(self, var, x, y, components=None):
"""
The derivative of the Branin function
"""
a, b, c, r, s, t = self._pars
if var == 0:
return 2 * a * (y - b * x**2 + c * x - r) * (-2 * b * x + c) - s * (
1 - t
) * np.sin(x)
else:
return 2 * a * (y - b * x**2 + c * x - r)
[docs]class BraninProblem(SimpleProblem):
"""
Problem definition of benchmark function Branin.
Parameters
----------
name : str
The name of the problem
ana_deriv : bool
Switch for analytical derivatives
initial_values : list of float
The initial values
Attributes
----------
initial_values : list of float
The initial values
"""
[docs] def __init__(self, name="branin", initial_values=[1.0, 1.0], ana_deriv=False):
super().__init__(
name,
float_vars={"x": initial_values[0], "y": initial_values[1]},
min_values_float={"x": -5.0, "y": 0.0},
max_values_float={"x": 10.0, "y": 15},
)
self.add_objective(BraninObjective(self, ana_deriv=ana_deriv))