Decomposition Principles and Online Learning in Cross-Layer Optimization for Delay-Sensitive Applications

TL;DR

This paper introduces a cross-layer optimization framework for delay-sensitive applications that accounts for dynamic network conditions and application dependencies, utilizing decomposition and online learning to adapt in real-time.

Contribution

It presents a novel decomposition approach combined with an online learning algorithm for real-time cross-layer optimization under uncertain network and application conditions.

Findings

The proposed online algorithm effectively adapts to unknown source and network dynamics.

Decomposition reduces complexity and facilitates message exchange between layers.

Numerical results confirm the algorithm's efficiency in delay-sensitive scenarios.

Abstract

In this paper, we propose a general cross-layer optimization framework in which we explicitly consider both the heterogeneous and dynamically changing characteristics of delay-sensitive applications and the underlying time-varying network conditions. We consider both the independently decodable data units (DUs, e.g. packets) and the interdependent DUs whose dependencies are captured by a directed acyclic graph (DAG). We first formulate the cross-layer design as a non-linear constrained optimization problem by assuming complete knowledge of the application characteristics and the underlying network conditions. The constrained cross-layer optimization is decomposed into several cross-layer optimization subproblems for each DU and two master problems. The proposed decomposition method determines the necessary message exchanges between layers for achieving the optimal cross-layer solution. 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 function in the Markov decision process (MDP) framework. Based on the dynamic programming solution to the MDP, we develop a low-complexity cross-layer optimization algorithm using online learning for each DU transmission. 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.