Compressed Control of Complex Wireless Networks

TL;DR

This paper introduces a joint optimization framework for complex wireless networks with multiple operators and technologies, using parameter clustering to simplify control and improve network utility based on traffic and topology dynamics.

Contribution

It proposes novel parameter compression techniques for joint optimization in multi-operator wireless networks, enabling scalable and adaptive control strategies.

Findings

Network capacity can increase up to 50% with multi-operator cooperation.

Optimal offloading prices vary significantly with traffic patterns.

Dynamic strategies are justified due to traffic-dependent cooperation decisions.

Abstract

Future wireless networks are envisioned to integrate multi-hop, multi-operator, multi-technology (m3) components in order to meet the increasing traffic demand at an acceptable price for subscribers. The performance of such a network depends on the multitude of parameters defining traffic statistics, network topology/technology, channel characteristics and business models for multi-operator cooperation. So far, most of these aspects have been addressed separately in the literature. Since the above parameters are mutually dependent, and simultaneously present in a network, for a given channel and traffic statistics, a joint optimization of technology and business model parameters is required. In this paper, we present such joint models of complex wireless networks and introduce optimization with parameter clustering to solve the problem in a tractable way for large number of parameters. By parameter clustering we compress the optimization vector and significantly simplify system implementation, hence the algorithm is referred to as compressed control of wireless networks. Two distinct parameter compression techniques are introduced mainly parameter absorption and parameter aggregation. Numerical results obtained in this way demonstrate clear maximum in the network utility as a function of the network topology parameters. The results, for a specific network with traffic offloading, show that the cooperation decisions between the multiple operators will be significantly influenced by the traffic dynamics. For typical example scenarios, the optimum offloading price varies by factor 3 for different traffic patterns which justifies the use of dynamic strategies in the decision process. Besides, if user availability increases by multi-operator cooperation, network capacity can be increased up to 50% and network throughput up to 30-40%.