I'm absolutely sure that Rij > 2*r most of the time, yes. This is obvious if you consider that rand(n) returns values in the range [0, 1] and r = 0.01. I have also verified this empirically. I'll try to explore the view update idea further, I've never used that approach before.
On Friday, October 24, 2014 5:46:37 PM UTC+3, Andreas Noack wrote: > > Are you sure that Rij > 2*r in your calculations? > > You should use I instead of eye(3), but probably the best solution is to > avoid the explicit calculation of the outer product. You can probably write > your problem as a rank one update of a view of pry > > Med venlig hilsen > > Andreas Noack >
