Optimization of Location Management for PCS Networks with CTRW Mobility Model

TL;DR

This paper develops an analytical framework using CTRW mobility models to optimize location update areas in PCS networks, significantly reducing costs by considering initial position and distance thresholds.

Contribution

It introduces a novel diffusion-based framework for optimizing location management parameters in PCS networks using CTRW mobility models.

Findings

Optimal initial position depends on drift strength.

Cost reduction of up to 37% by optimizing initial position.

Analytical expressions for cost and parameters under various conditions.

Abstract

This paper considers the design of the optimal locationupdate area (LA) of the distance-based scheme for personal communication service (PCS) networks. We focus on the optimization of two design parameters associated with the LA: 1) initial position upon LA update; 2) distance threshold for triggering of LA update. Based on the popular continuous-time random walk (CTRW) mobility model, we propose a novel analytical framework that uses a diffusion equation to minimize the location management cost. In this framework, a number of measurable physical parameters, such as length of road section, angle between road sections, and road section crossing time, can be integrated into the system design. This framework allows us to easily evaluate the total cost under general call arrival distributions and LA of different shapes. For the particular case of circular LA and small Poisson call-arrival rate, we prove the following: (1) When the drift is weak, the optimal initial position approaches the center of the LA; when the drift is strong, it approaches the boundary of the LA. (2) Comparing the optimal initial-position and center-initial-position solutions (which is assumed in most prior work), when the drift is weak, the optimal distance threshold and the minimum total cost are roughly equal; when the drift is strong, the optimal distance threshold in the later is about 1.260 times that in the former, and the minimum total cost in the later is about 1.587 times that in the former. That is, optimizing on initial position, which previous work did not consider, has the potential of reducing the cost measure by 37%.