def test_2():
fss = FishSchoolSearch(
lower_bound_point=[-100, -100],
higher_bound_point=[100, 100],
population_size=50,
iteration_count=100,
individual_step_start=10,
individual_step_final=0.01,
weight_scale=50,
func=lambda x: 10 * (np.sin(0.1*x[0]) + np.sin(0.1*x[1])) + 20,
)
f, x = fss.max()
cmp(np.array([10*np.pi/2, 10*np.pi/2]), 40, x, f, fss)
c = convergence.Convergence(fss.history(), 40)
c.show()
def test_1():
fss = FishSchoolSearch(
lower_bound_point=[-10, -10],
higher_bound_point=[10, 10],
population_size=50,
iteration_count=100,
individual_step_start=2,
individual_step_final=0.01,
weight_scale=50,
func=lambda x: -0.1 * (x[0] * x[0] + x[1] * x[1]) + 20,
)
# v = visualization.Visualization(fss.history())
# # v.start(saved=True, filename="test_1.mp4")
f, x = fss.max()
cmp(np.array([0,0]), 20, x, f, fss)
c = convergence.Convergence(fss.history(), 20)
c.show()
def test_6():
def rosenbrock_func(x):
return sum((1 - x[i])*(1 - x[i]) + 100*(x[i+1] - x[i]*x[i])*(x[i+1] - x[i]*x[i]) for i in range(1, len(x) - 1, 1))
fss = FishSchoolSearch(
lower_bound_point=[-1, -1, -1, -1],
higher_bound_point=[2, 2, 2, 2],
population_size=50,
iteration_count=500,
individual_step_start=1,
individual_step_final=0.01,
weight_scale=50,
func=lambda x: 200 - rosenbrock_func(x),
)
f, x = fss.max()
cmp(np.array([1, 1, 1, 1]), 200, x, f, fss)
c = convergence.Convergence(fss.history(), 200)
c.show()
def test_3():
def rosenbrock_func(x, y):
return (1 - x)*(1 - x) + 100*(y - x*x)*(y - x*x)
fss = FishSchoolSearch(
lower_bound_point=[-1, -1],
higher_bound_point=[2, 2],
population_size=50,
iteration_count=100,
individual_step_start=1,
individual_step_final=0.01,
weight_scale=50,
func=lambda x: 200 - rosenbrock_func(x[0], x[1]),
)
f, x = fss.max()
cmp(np.array([0, 0]), 200, x, f, fss)
c = convergence.Convergence(fss.history(), 200)
c.show()
def test_5():
def holder_func(x, y):
return np.abs(np.sin(x)*np.cos(y)*np.exp(np.abs(1 - np.sqrt(x*x + y*y)/np.pi)))
fss = FishSchoolSearch(
lower_bound_point=[-10, -10],
higher_bound_point=[10, 10],
population_size=50,
iteration_count=200,
individual_step_start=5,
individual_step_final=0.01,
weight_scale=50,
func=lambda x: holder_func(x[0], x[1]),
)
f, x = fss.max()
cmp(np.array([8.05502, -9.66459]), 19.2085, x, f, fss)
c = convergence.Convergence(fss.history(), 19.2085)
c.show()
def test_4():
def ackley_func(x, y):
return -20 * np.exp(-0.2*np.sqrt(0.5*(x*x + y*y))) \
- np.exp(0.5*(np.cos(2*np.pi*x) + np.cos(2*np.pi*y))) + np.exp(1) + 20
fss = FishSchoolSearch(
lower_bound_point=[-4, -4],
higher_bound_point=[4, 4],
population_size=50,
iteration_count=100,
individual_step_start=3,
individual_step_final=0.01,
weight_scale=50,
func=lambda x: 20 - ackley_func(x[0], x[1]),
)
f, x = fss.max()
cmp(np.array([0, 0]), 20, x, f, fss)
c = convergence.Convergence(fss.history(), 20)
c.show()