Clustering schemes for energy conservation in two-tier mobile Ad-Hoc networks

Abstract

In this thesis, we study and propose clustering schemes for energy conservation in two-tier mobile ad-hoc networks. We propose seven new clustering schemes for two-tier mobile network. Proposed clustering schemes can be classified in four categories: Schemes for point-to-point connections and point-to-multipoint connections, and schemes using periodic control signaling and event-driven signaling for low signaling rate. Two distributed heuristic clustering schemes for point-to-point connections and two clustering schemes for point-to-multipoint connections are proposed which will minimize the required transmission power in two-tiered mobile ad-hoc networks. The proposed periodic clustering schemes are implementable and real-time executable. The performances of proposed schemes are simulated and compared with optimum configurations based on the mean transmission power and the call drop rate as performance measures. Numerical results show that the proposed schemes have performance similar to optimum results. Two distributed and event-driven heuristic clustering schemes for point-to-point connections and a distributed heuristic clustering scheme for pointto-multipoint connections are proposed that will minimize the required trans-mission power concerning signaling traffic in two-tiered mobile ad-hoc networks. These schemes can be implemented and executed in real-time. Performance of the proposed schemes are analyzed and simulated for comparing with a previous periodic scheme. The mean energy consumed for signaling and the mean transmission power are used as performance measures. Numerical examples show that the proposed schemes support performance similar to the previous periodic scheme with a much lower energy consumed for signaling. Proposed schemes are simulated and mathematically analyzed for numerical comparison with optimal results from binary integer programming (BIP) equations. All proposed schemes are distributed, time-limited energy con-serving clustering...