Title: A Hidden Resource in Wireless Channel Capacity: Dependence Control in Action

Abstract: This paper aims to initiate the research on dependence control, which transforms the dependence structure of a stochastic process in the system through dependence manipulation, to improve the system performance. Specifically, we develop a dependence control theory for wireless channels, focusing on three principles in dependence control: (i) the asymptotic decay rates of delay and backlog in the system are the measures for dependence comparison and ordering, (ii) the dependence in the arrival process and the service process have a dual potency to influence the system performance, and (iii) the manipulation of the dependence in the free dimensions of the arrival or service process transforms the dependence structure of the arrival or service process. In addition, we apply the theory to the Markov additive process, which is a general model for a class of arrival processes and a versatile model for wireless channel capacity, and derive a set of results for various performance measures, including delay, backlog, and delay-constrained capacity. To demonstrate the use of the theory, we focus on dependence manipulation in wireless channel capacity, where we use copula to represent the dependence structure of the underlying Markov process of wireless channel capacity. We show that, based on a priori information of the temporal dependence of the uncontrollable parameters and the spatial dependence between the uncontrollable and controllable parameters, we can construct a sequence of temporal copulas of the Markov process and obtain a sequence of transition matrices of the controllable parameters to achieve the demanded dependence properties of the wireless channel capacity. This dependence manipulation technique is validated by simulation.