Unfortunately to discuss under what
circumstances a splitting of A into M minus N has the property that for some norm
M inverse N is less than one requires us to learn a little bit about
eigenvalues and the spectral radius of the matrix comes in. And that’s a future
topic in this course. So we’re going to defer on that. However, you can sorta
contemplate under what circumstances is M inverse N in norm small. Okay? What is
the best way of splitting A into M and N such that you converge in the fewest
iterations? Think about it.