bmtgoncalves

bmtgoncalves

bmtgoncalves

def ErdosRenyi(N, m=N):
    net = np.zeros(2*m, dtype='int')

    for i in xrange(m):
        while True:
            node1, node2 = np.random.randint(0, N, 2)

            # No self loops allowed
            if node1 != node2:
                break

bmtgoncalves

def BarabasiAlbert(N):
    net = np.zeros(2*N, dtype='int')

    # Edges represented by sequence of nodes
    net[:6] = [0, 1, 1, 2, 2, 3]

    # Generate the network
    for i in xrange(3, N):
        pos = np.random.randint(0, i)
        net[2*i] = i

bmtgoncalves

# Perform the power method for "iter" iterations
def Power_Method(G, iter):
    N = G.shape[0]
    x0 = np.ones(N)/N

    for i in range(iter):
        x0 = np.dot(G, x0)

    return x0

bmtgoncalves

# Compute the "Google" Matrix from the adjacency matrix, A.
# For illustration purposes only. *Not* efficient!
def Google_Matrix(A, m):
    N = A.shape[0]
    v = np.ones(N)
    
    # Calculate the degree of each node
    KT = np.dot(A.T, v)

    # Normalize the columns

bmtgoncalves

def bak_sneppen(data, steps=10):
    N = len(data)

    for i in range(steps*N):
        pos = np.argmin(data) # Find the position of the minimum value
        
        # Replace the minimum valuea and both it's neighbours
        data[(pos-1) % N] = np.random.random()
        data[pos] = np.random.random()
        data[(pos+1) % N] = np.random.random()

bmtgoncalves

def uncorrelated(data, steps=10):
    N = len(data)

    for i in range(steps*N):
        pos = np.random.randint(N) # Select an element at random

        data[pos] = np.random.random() # Replace the value

    return data