Consider a network of multiple independent stochas-tic linear systems where, for each system, a scheduler collocatedwith the sensors arbitrates data transmissions to a correspondingremote controller through a shared contention-based communi-cation network. While the systems are physically independent,their optimal controller design problems may, in general, becomecoupled, due to network contention, if the schedulers triggertransmissions based on state-dependent events. In this article wepropose a class of probabilistic admissible schedulers for whichthe optimal controllers, with respect to local standard LQG costs,have the certainty equivalence property and can still be determineddecentrally. Then, two subclasses of scheduling policies withinthis class are introduced; a non-event-based subclass, so calledpurely stochastic transmission (PST) policy, and an event-basedsubclass, both with easily adjustable triggering probabilities atevery time step. We then prove that, given a PST policy withgiven triggering probabilities and an associated closed-loop per-formance with optimal control law, we can find an event-basedscheduler from the proposed subclass and with the same trigger-ing probabilities for which the associated closed-loop performancewith optimal control law is strictly superior. Moreover, we showthat, for each closed-loop system, the optimal state estimatorsfor both scheduling policies follows a linear iteration. Finally, weprovide a method to regulate the triggering probabilities of theschedulers by maximizing a network utility function.