On the Generalized Delay-Capacity Tradeoff of Mobile Networks with L\'evy Flight Mobility

Abstract

In the literature, scaling laws for wireless mobile networks have been characterized under various models of node mobility and several assumptions on how communication occurs between nodes. To improve the realism in the analysis of scaling laws, we propose a new analytical framework. The framework is the first to consider a L\'{e}vy flight mobility pattern, which is known to closely mimic human mobility patterns. Also, this is the first work that allows nodes to communicate while being mobile. Under this framework, delays ($\bar{D}$) to obtain various levels of per-node throughput $(\lambda)$ for L\'evy flight are suggested as $\bar{D}(\lambda) = O(\sqrt{\min (n^{1+\alpha} \lambda, n^2)})$, where L\'evy flight is a random walk of a power-law flight distribution with an exponent $\alpha \in (0,2]$. The same framework presents a new tighter tradeoff $\bar{D}(\lambda) = O(\sqrt{\max (1,n\lambda^3)})$ for \textit{i.i.d.} mobility, whose delays are lower than existing results for the same levels of per-node throughput.