Investigations on Location Prediction and Resource Management of Cellular Network

In this era of smart applications, the advancement of mobile technology has improved the style of living. Specifically, confining the focus to mobile communication, mobile phones have become an inevitable part of human being. With the invent of cellular communication technology, the world looks smaller and connected to every nook and corner. The modern day technology has made improvised cellular communication in terms of portability, quality of service, reliability, affordability and various other features offered to users. This led to scarcity in network resources in terms of bandwidth and signalling cost, specifically finding the location of a mobile user in the network. One critical feature that enables the ubiquitous communication in the location management which is perceived to provide standard quality of service even under very harsh and unexpected conditions. The main objective of this thesis is to examine as well as access the e-ciency of the cost effectiveness of the cellular network location management by the use of soft computing techniques and to ensure e-cient bandwidth management in the mobile computing network. This work focusses mainly on reduction of location management cost on the basis of MSC's (Mobile Switching Center) prediction of user's next location using his/her past histroy. Many pros and cons of the different techniques are also highlighted in this thesis. In the recent years, even though many prediction techniques are adopted in various engineering fields, still statistical analysis and artificial intelligence methods have not been explored much in the implementation of mobile location management. So in this dissertation, statistical analysis and dynamic nonlinear predictor models have been explored to validate the techniques of location management for forecasting the location of a mobile user to easily forward the call or SMS without degradion of quality of service. Further, based on the user's mobility, it is required to optimise the network so as to reduce the signalling cost that incurs during the operation of location update and paging, two basic component of mobility management operations. Some nature inspired softcomputing algorithms such as Particle Swarm Optimization (PSO), Binary Bat Algorithm (BBA) are used to defend the accuracy of the location management issues. Similarly, the hybrid BPSO-BGA technique has also been verified for the optimization of network compared to the genetic algorithm under identical situation. Performances of all these techniques have been tested for static networks of different sizes to claim the e-ciency of the models. Dynamic management like profile based paging has also been investigated with the real data to protect the bandwidth issue in our work. So, the results of the proposed pro le based paging approach are analysed and compared with Hyper location area (Hypo-La) paging which is found to be encouraging. This research work has a significance in identifying the limitations of the previous work as well as overcoming of issues by proposing novel ideas. The analytical and simulated results show the applicability of the proposed methods on both the test network and the real networks