# -*- coding: utf-8 -*-

"""

Created on Mon Dec 18 17:08:12 2017

@author: SoliareOfAstora

"""

import numpy as np

class diagonalArray:

values = np.array([8, 2, 1])

offset = np.array([0, 1, 3])

def get(self, x, y):

if y < 0:

return 0

distance = x - y

if distance < 0:

distance *= -1

for i in range(np.size(self.offset)):

if self.offset[i] == distance:

return self.values[i]

return 0

def calculate(self, source, output):

for i in range(64):

output[i] = 0

for j in range(-3, 4, 1):

if 0 <= i + j < 64:

output[i] += self.get(self, i, i + j) * source[i + j]

arr = diagonalArray

output = np.ones(64)

while True:

inp = np.copy(output)

diagonalArray.calculate(arr, inp, output)

output /= np.linalg.norm(output)

if np.linalg.norm(output - inp) < pow(10, -7):

break

diagonalArray.calculate(arr, inp, output)

# ============================================================

np.linalg.norm(output) # Out[4]: 13.97516637314112

output # Wektor własny

# po obliczeniu powyższych można przejśc do poszukiwania drugiej co do wielkości wartości własnej

# ============================================================

output /= np.linalg.norm(output)

x = np.copy(output)

arr = diagonalArray

output = np.ones(64)

inp = np.copy(output)

inp /= np.linalg.norm(inp)

inp -= x * (x.dot(inp))

while True:

inp /= np.linalg.norm(inp)

diagonalArray.calculate(arr, inp, output)

inp = np.copy(output)

inp -= x * (x.dot(inp))

if np.linalg.norm(output - inp) < pow(10, -7):

break

# ============================================================

np.linalg.norm(output) # Out[6]: 13.900452697733499

inp # Wektor własny drugi co do wielkości

# ============================================================