Data Mining Assisted Resource Management in Wide WLANs

Abstract

WLANs are currently being considered for use in the context of a larger geographical area such as a city or a campus due to their convenience, cost efficiency, and ease of integration with other networks. Due to support large numbers of portable devices and their dynamic relocation, wide WLANs must face problems of location management and network resource allocation. In order to solve these challenges, future mobility information of all mobile users in the network is required to accurate estimation of network resource demands at future time toward more efficient network resource management. Therefore, mobility prediction has played a crucial role in the resource management of wide WLANs and it has attracted recently a great deal of research interests. However, since most of the current approaches are based on personal movement profile for predicting the next location of mobile users, these techniques may fail to make a prediction for new users or ones with movements on novel paths. In this paper, we propose a prediction model which is based on group mobility behaviors to deal with such the lack of information of individual movement histories. Our proposed prediction approach makes use of clustering techniques in data mining to classify mobility patterns of users into groups. Experiments will be performed to demonstrate that using group mobility behaviors may significantly enhance the accuracy of the mobility prediction.