Index operation: shapes do not match

General
1

Hello, I am learning to operate with indices.

In the following example I try to calculate a 4-order tensor and then perform a double contraction but when I use inner I get the error “shapes do not match”.

from dolfin import *
from ufl import indices
import numpy as np

mesh = UnitSquareMesh(2,2) 
i,j,k,l,m = indices(5)
I  = Identity(2)
a = as_tensor([[1.0,2.0],[0.5,0.5]])
b = as_tensor([[1.5,1.0],[0.3,0.8]])
R = as_tensor(a[i,l]*a[j,k]+a[i,m]*a[j,m]*I[k,l],(i,k,j,l))

c = inner(R,b)

I know I can do it by doing:

c2 = as_tensor(R[i,k,j,l]*b[j,l],(i,k))

but, is there any way to make it work with inner?

Thanks!

2

Consider the documentation for ufl.inner. If this doesn’t satisfy your needs you’ll need to define the product yourself as you have done already.

1 Like