Nature-inspired optimization of Moving Access Point-based radio networks

Abstract

Wireless communication systems have penetrated into almost all parts of human life. They have become essential and pervasive and do affect both personal and professional aspects of our lives. While most types of wireless networks, such as cellular mobile communication networks, have been exhaustively studied and optimized, in order to handle various situations, there are also contexts which require efficient handling, e.g., moving hotspots, areas that have lost their infrastructure, and areas with hard morphology. The motivation for this work is the efficient handling of these important contexts by the use of Moving Access Points (MAPs). MAPs are capable of autonomously moving and establishing a radio network in short time, with limited centralized management. The radio network provides wireless access to users, is based on ad hoc connectivity (self-adapting mesh network concept), and has some elements acting as gateways to a wide-area infrastructure. The main purpose of this paper is to find the optimal position of the MAPs, i.e., the ones which require minimum movement and telecommunication cost. In order to achieve this goal, an innovative algorithm that combines the well-known simulated annealing algorithm with the characteristic of pheromone of the ant colony optimization is proposed. This is done in order to exploit ant colony optimization concepts in the related problem of the optimization of the MAP-based radio network design. The operation of the algorithm is validated by being applied to three most indicative scenarios, in which it is compared to the pure simulated annealing and a brute-force method, respectively. Our scheme exhibits a 5% improved performance in terms of solution quality in medium and large size problems, a 9% improved speed for the estimation of the suboptimal solution and 15% quicker adaptation to new context, without impacting the computational complexity.