One-dimensional location area design

Abstract

Location management cost in a wireless network is expressed in this work as the sum of signaling cost due to paging and signaling cost due to registration. We introduce and compare a heuristic method and an asymptotic method for selecting a disjoint partition of a one-dimensional service area to minimize location management costs. In special cases for which the best design is known, the asymptotic design rule outperforms the heuristic rule, including cases for which the number of location areas is small. With offered load /spl lambda/, per-event paging cost P, per-event registration cost R, and fixed traffic patterns, optimal location area edges depend on /spl radic//spl lambda/P/R and location management costs are proportional to /spl radic//spl lambda/PR.