Analysis and Optimization of Cellular Network with Burst Traffic

TL;DR

This paper models cellular network performance with bursty Poisson traffic, analyzing SINR and delay, and proposes an iterative algorithm to optimize base station density for minimal power consumption.

Contribution

It introduces a stochastic geometry-based framework for cellular networks with burst traffic and develops an iterative method to find the optimal base station density.

Findings

Optimal BS density reduces power consumption

SINR depends on transmission probability solved via bisection

Proposed algorithm converges to the global optimum

Abstract

In this paper, we analyze the performance of cellular networks and study the optimal base station (BS) density to reduce the network power consumption. In contrast to previous works with similar purpose, we consider Poisson traffic for users' traffic model. In such situation, each BS can be viewed as M/G/1 queuing model. Based on theory of stochastic geometry, we analyze users' signal-to-interference-plus-noise-ratio (SINR) and obtain the average transmission time of each packet. While most of the previous works on SINR analysis in academia considered full buffer traffic, our analysis provides a basic framework to estimate the performance of cellular networks with burst traffic. We find that the users' SINR depends on the average transmission probability of BSs, which is defined by a nonlinear equation. As it is difficult to obtain the closed-form solution, we solve this nonlinear equation by bisection method. Besides, we formulate the optimization problem to minimize the area power consumption. An iteration algorithm is proposed to derive the local optimal BS density, and the numerical result shows that the proposed algorithm can converge to the global optimal BS density. At the end, the impact of BS density on users' SINR and average packet delay will be discussed.