Multi-layer optimization with backpressure and genetic algorithms for multi-hop wireless networks

Abstract

This paper presents an efficient scheme to optimize multiple layers in multi-hop wireless networks with throughput objectives. Considering channel sensing and power control at the physical layer, a non-convex throughput optimization problem is formulated for resource allocation and a genetic algorithm is designed to allow distributed implementation. To address link and network layers, a localized back-pressure algorithm is designed to make routing, scheduling, and frequency band assignments along with physical-layer considerations. Our multi-layer scheme is extended to cognitive radio networks with different user classes and evaluate our analytical solution via simulations. Hardware-in-the-loop emulation test results obtained with real radio transmissions over emulated channels are presented to verify the performance of our distributed multi-layer optimization solution for multi-hop wireless networks. Finally, a security system is considered, where links have their security levels and data flows require certain security levels on each of its links. This problem is addressed by formulating additional constraints to the optimization problem.