Practical protocols for the control of distributed networks are based on a collision avoidance rationale, tailoring the separation
communications in the time and frequency domain. During the last years, considerable advancements have been achieved in the
development and implementation of transmitter-receiver architectures reducing the impact of the mutual interference among
communications sharing the same time/frequency resource. Therefore, the use of collision avoidance protocols for the control of future
distributed networks appears to be a simplistic and inefficient solution. Nevertheless, the study of interference networks mostly focuses
on small topologies and fixed and known channels, and often neglects the impact of interference on the long-term evolution of transmissions.

The fast and unpredictable variations of the channel conditions in asynchronous networks with simultaneous access due to the start and end of
interfering transmissions make rate and error control critical for the efficiency of the network. In this scenario, there exists a strong mutual
interdependency between the statistics of the interference and the rate and error control algorithms. For instance, different error control protocols
(e.g., FEC, ARQ, Hybrid ARQ) not only result in a different average interference load, but also in a different correlation of the stochastic process
modeling the interference. The latter effect influences the effectiveness of the transmission parameters' control. In such a network, the implementation
of cooperation protocols produces an effect on the interference load and characterization as well.

In cognitive networks, the interference generated by the secondary users' activity impacts the stochastic process modeling primary users. In fact, secondary users'
transmissions increase the failure probability at primary users' receivers; thus, it biases the long-term evolution of the primary transmitters' queue,
destination and ARQ state. This effect, which we call process distortion, affects the performance of the primary users, but also defines a strategy for
the secondary users' transmissions. For a simple two-link network, in which the primary transmitter adopts a finite retransmissions ARQ protocol, the
structure of the optimal transmission policy for the secondary transmitter can be analytically determined. For larger networks, and more complex scenarios,
we devised an optimization framework based on a constrained infinite horizon Markov Decision Process.