Joint optimization for base station density and user association in energy-efficient cellular networks

Abstract

In the research of green networks, considering the base station (BS) density from the perspective of energy efficiency is very meaningful for both network deployment and BS sleeping based power saving. In this paper, we optimize the BS density for energy efficiency in cellular networks by the stochastic geometry theory and optimize the user association matrix by the Quantum Particle Swarm Optimization (QPSO). On one hand, we model the distribution of base stations and user equipment (UE) as spatial Poisson point process (PPP). Based on such model, we derive the closed-form expressions of the average achievable data rate, the network energy consumption and the network energy efficiency with respect to the network load. Then, we optimize the BS density for network energy efficiency maximization by adopting the Newton iteration method. On the other hand, we build a user association matrix to present the connection state between BSs and UEs, and then optimize it by QPSO. Our study reveals that we can improve the network energy efficiency by switching on/off proportion of the BSs according to the network load. The simulation results validate the theoretical analysis, and show that when the right amount of BSs is deployed according to the network load, the network energy efficiency can be maximized and the maximum energy efficiency is a fixed value once the network parameters are given.