Question

I’m using scipy.linalg to solve a matrix equation A*x = b
The following code does not work:

from scipy import *
from pylab import *
from scipy.sparse  import lil_matrix 
from scipy.sparse.linalg import spsolve, factorized 
from numpy.random import rand 
from numpy import ones 

def build_matrix(n): 
        A_lil = lil_matrix((n, n)) 
        A_lil.setdiag(rand(n)) 
        A_lil.setdiag(rand(n-10), k = 10) 
        A_lil.setdiag(rand(n-10), k = -10) 
        A_csr = A_lil.tocsr() 
        return A_csr 

A = build_matrix(500)
fsolve = factorized(A)
b = rand(500,1)

fsolve(b)

Traceback (most recent call last):
  File "<pyshell#70>", line 1, in <module>
    fsolve(b)
SystemError: gstrs was called with invalid arguments

(Edit: actually, the console does not say fsolve(b,trans), it says self.solve(b,trans), so the part below is probably wrong)

If I type in fsolve manually into the console I see that fsolve requires two arguments ‘b’, and ‘trans’. This for some reason is not mentioned in the documentation for factorized here: http://docs.scipy.org/doc/scipy/reference/generated/scipy.sparse.linalg.factorized.html. Adding in the argument ‘N’ gives the following error

>> fsolve(b, 'N')
Traceback (most recent call last):
  File "<pyshell#55>", line 1, in <module>
    fsolve(b,'N')
SystemError: gstrs was called with invalid arguments

Any help?

Answer 1

You seem to have an error in:

b=rand(500,1)

it needs to be:

b=rand(500)

else the array gets ‘too deep’. Hope that helps.