Cross-Layer Optimization with Complete and Incomplete Knowledge for Delay-Sensitive Applications

In this paper, we first formulate the cross-layer design as a non-linear constrained optimization problem by assuming complete knowledge of the dynamically changing application characteristics and the underlying time-varying network conditions. By decomposing the cross-layer optimization problem, we determine the necessary message exchanges between layers for achieving the optimal cross-layer solution and explicitly show how the cross-layer strategies selected for one data unit (DU, e.g. packet) will impact its neighboring DUs as well as the DUs that depend on it. However, the attributes (e.g. distortion impact, delay deadline etc) of future DUs as well as the network conditions are often unknown in the considered real-time applications. The impact of current cross-layer actions on the future DUs can be characterized by a state-value junction in the Markov decision process (MDP) framework. Based on the value iteration solution to the MDP, we develop a low-complexity cross-layer optimization algorithm using online learning for each DU transmission. This online optimization utilizes information only about the previous transmitted DUs and past experienced network conditions. This online algorithm can be implemented in real-time in order to cope with unknown source characteristics, network dynamics and resource constraints. Our numerical results demonstrate the efficiency of the proposed online algorithm.