A new distributed optimization framework for hybrid ad-hoc networks
Georgios Tychogiorgos, Athanasios Gkelias, Kin Kwong Leung · 2011
The continuously increasing demand for resources in modern networks urges for more efficient resource allocation. Such an allocation of resources to network users can be formulated as an optimization problem. However, the existence of wireless links in modern networks and the competition for resources by multimedia applications turn the optimization problem into a non-convex one, which is in general difficult to solve. This paper presents a non-convex optimization formulation to describe the Network Resource Allocation problem in hybrid ad-hoc networks, i.e. networks with both wired and wireless links. To find the optimal solution to this problem, a novel general optimization framework, for non-convex optimization problems, is presented and the necessary and sufficient condition for the convergence of a distributed algorithm to the optimal solution is also proven. Moreover, based on this framework, a distributed joint power and rate adaptation algorithm is proposed to calculate the optimal solution, and finally, the convergence and optimality of the algorithm are verified by simulation.