def algorithm(point, lmbda):
fhelper.func = "x**2" # raw_input("Enter function: ")
x_range = np.arange(-50, 50)
fig = plt.figure()
ax = fig.add_subplot(111)
plt_func, = ax.plot(x_range, [fhelper.f(value) for value in x_range], label=fhelper.func)
plt.legend(handles=[plt_func])
# Algorithm
df_result = fhelper.df(point)
while abs(df_result) > abs(lmbda * 0.1):
df_result = fhelper.df(point)
point -= lmbda * df_result
print("Extremalstelle bei x = %f" % point)
# Plot approximation
point1 = -40
x_range_tan = np.arange(point1 - 20, point1 + 20)
ax.plot(x_range_tan, fhelper.tan(x_range_tan, point1))
point2 = -20
x_range_tan = np.arange(point2 - 20, point2 + 20)
ax.plot(x_range_tan, fhelper.tan(x_range_tan, point2))
x_range_tan = np.arange(point - 20, point + 20)
ax.plot(x_range_tan, fhelper.tan(x_range_tan, point))
plt.grid(True)
ax.set_xlabel("x")
ax.set_ylabel("y")
plt.savefig("doc/gradient.png")
def bisect(a, b):
c = (a + b) / 2
ax.plot((c, c), (0, fhelper.f(c) + 30), 'r-')
if abs(fhelper.df(c)) < 0.001:
return c
elif fhelper.df(c) < 0:
return bisect(c, b)
else:
return bisect(a, c)
