“…. (43) For all channel states , ̅ ( , , π(1)) can be approximated as ̅ ( , , π(1)) = ̅ ,0 −1 ( , , π(1)) [40], where ̅ ,0 is a constant and independent on the channel state, ( , , π(1)) represents the theoretical achievable rate of user under channel matrix . In the following, we derive ( , , π(1)).…”