On Movement-based Location Update in Nonhomogeneous Topologies of Multitier PCS Networks

We investigates the performance of movement-based location update in multitier wireless networks with sparsely correlated low-tier microcells. The topologies are considered to have random overlays of different microcell cluster sizes, with distributions spatially of two forms: homogeneous and nonhomogeneous. Permission for a mobile station (MS) to report to either a low-tier mi-crocell or a high-tier macrocell is determined by the state of macrocell-based movement counter compared with two thresholds. The higher threshold is fixed but the lower threshold is either static or adaptive with a value optimized for an MS roaming everywhere or locally, respectively. Results from simulations on basis of user mobility of memoryless random walk between microcells suggest that given an average density of accessible microcells, each location update scheme in topologies of nonhomogeneous distributions or random overlays of large cluster sizes yields a higher optimal total location update cost per call arrival; that the adaptive scheme is able to gain cost advantages over a corresponding static one only in nonhomogeneous topologies; and that the schemes by using one more buffer space for the previous visited macrocell ID effectively reduce many unnecessary location updates