hellomrrobot
- 10
- 0
I am having trouble showing that ##(A_{ik}B_{kj})_{mm} = (A_{ki}B_{jk})_{mm}## Wouldn't the right side end up having a different outcome? Or can we assume its symmetric?
It is true for the anti-symmetric part also because the sign change cancels.hellomrrobot said:I am having trouble showing that ##(A_{ik}B_{kj})_{mm} = (A_{ki}B_{jk})_{mm}##Wouldn't the right side end up having a different outcome? Or can we assume its symmetric?
Mentz114 said:It is true for the anti-symmetric part also because the sign change cancels.
Yes, I think so. A and B can be decomposed into symmetric and anti-symmetric parts.hellomrrobot said:anti-symmetric?
It would help if you would explain your notation.hellomrrobot said:I am having trouble showing that ##(A_{ik}B_{kj})_{mm} = (A_{ki}B_{jk})_{mm}##Wouldn't the right side end up having a different outcome? Or can we assume its symmetric?
gill1109 said:It would help if you would explain your notation.
I suppose that you are using Einstein summation convention. And that when you write (A_{ik}B_{kj}) you mean, the matrix whose i,j element is sum_k (A_{ik}B_{kj}). In other words, the ij element of the matrix AB. Now you want to sum over m the m,m elements of that matrix, so you are talking about trace(AB).
Similarly on the right hand side, replace i and j both by m and m, sum over m and sum over k, and you see that what you have written is just trace(AB).
Thank you! You are right.stevendaryl said:Well, I would say that the right hand side is trace(BA), but that is always equal to trace(AB).