def model_2_Grad_Decent_unconstrained_nice_problem():
x = [3, 1, 3, 0, 0]
x0 = [1, 0, 1, 0, 0]
z, inner = generate_points(x, size=500, scale=0.7)
z = generate_noise(z, 1 * 10 ** -1)
plot_solution(x0, z, inner, rxy_tilde, 0, Metode=2)
xf, nf, gradsf = BFGS_model_2(x0, z, inner, TOL=10 ** (-6), gradient_decent=1)
plot_solution(xf, z, inner, rxy_tilde, nf, Metode=2)
convergence_plot(gradsf, 2)
return 0
def model_2_BFGS_unconstrained_not_so_nice_problem():
x = [0.008, 1, 0.008, 0, 0]
x0 = [1, 0, 1, 0, 0]
z, inner = generate_points(x, size=500)
z = generate_noise(z, 1 * 10 ** -1)
plot_solution(x0, z, inner, rxy_tilde, 0, Metode=2)
xf, nf, gradsf = BFGS_model_2(x0, z, inner, TOL=10 ** (-10), gradient_decent=0)
plot_solution(xf, z, inner, rxy_tilde, nf, Metode=2)
convergence_plot(gradsf, 2)
return 0
def model_1_BFGS_unconstrained_nice_problem():
x = [3, 1, 3, 0, 0]
x0 = [1, 0, 1, 0, 0]
z, inner = generate_points(x, size=500, scale=0.7)
z = generate_noise(z, 1 * 10 ** -1)
plot_solution(x0, z, inner, rxy, 0, Metode=1)
xf, nf, gradsf = BFGS_model_1(x0, z, inner, TOL=10 ** (-10), gradient_decent=0)
plot_solution(xf, z, inner, rxy, nf, Metode=1)
convergence_plot(gradsf, 1)
return 0
def model_2_BFGS_constrained_nice_problem():
x = [3, 1, 3, 0, 0]
c = [c1, c2, c3, c4, c5]
x0 = np.array([4, 1, 3, 0, 0])
points, inner = generate_points(x, size=300, scale=1)
points = generate_noise(points, 2 * 10 ** (-1))
plot_solution(x0, points, inner, rxy_tilde, 0, 2)
xf, itr, b_vals = beta_optimization(x0, B_func, grad_B, f2, grad2, 1, c, points, inner, 0.1, 1000, n=0, TOL=10 ** (-6))
plot_solution(xf, points, inner, rxy_tilde, np.sum(itr), 2)
convergence_plot_constr(b_vals)
return 0
H = (np.eye(5) - rho * temp1) @ H @ (np.eye(5) -
rho * temp2) + rho * temp3
print('n = ', n, "\t x=", xnew)
n += 1
funks = np.append(funks, f2(xnew, z, inner))
return xnew, n - 1, funks
if __name__ == '__main__':
#x = [0.01, 1, 0.1, 0, 0] #kult problem
x = [3, 1, 3, 0, 0]
x0 = np.array([5, 1, 0.1, 0, 0])
points, inner = generate_points(x, size=500)
Af, cf = constructproblem(x0)
points = generate_noise(points, 2 * 10**(-1))
plot_solution(x0, points, inner, rxy_tilde, 0, Metode=2)
xf, nf, gradients = BFGS_model_2(x0,
points,
inner,
TOL=10**(-10),
gradient_decent=0)
plot_solution(xf, points, inner, rxy_tilde, nf, Metode=2)
convergence_plot(gradients, 2)
if __name__ == '__main__':
#x = [3, 1, 3, 0, 0]
#x = [3, 1, 0.2, 0, 0]
#x = [0.008, 1, 0.008, 0, 0] #not so nice problem
x = [0.008, 1, 0.008, 0, 0]
c = [c1, c2, c3, c4, c5]
x0 = np.array([4, 1, 3, 0, 0])
points, inner = generate_points_constr(x, size=200)
points = generate_noise(points, 2 * 10**(-1))
plot_solution(x, points, inner, rxy_tilde, 0, 2)
print(B_func(x, f2, 0.1, c, points, inner))
print(B_func(x0, f2, 0.1, c, points, inner))
#xf = BFGS_constr(x0, B_func, grad_B, f2, grad2, 5, c, points, inner, 0.1, 1000, n = 0, TOL = 10 **(-3)) #general_BFGS(x, f, gradf, n=0, TOL=10**(-6))
xf, itr, b_vals = beta_optimization(x0,
B_func,
grad_B,
f2,
grad2,
1,
c,
points,
inner,