This paper presents binary differential evolution based optimal reporting cell planning (RCP) for location management in wireless cellular networks. The significance of mobile location management (MLM) in wireless communication has evolved drastically due to tremendous rise in the number of mobile users with the constraint of limited bandwidth. The total location management cost involves signaling cost due to location registration and location search and a trade-off between these two gives optimal location management cost. The proposed binary differential evolution (BDE) algorithm is used to determine the optimal reporting cell planning configuration such that the overall mobility management cost is minimized. Evidently, from the simulation result the proposed technique works well for the reference networks in terms of optimal cost and convergence speed. Further the applicability of the BDE is also validated for the realistic network of BSNL (Bharat Sanchar Nigam Limited), Odisha.

This work describes the neural network technique to solve location management problem. A multilayer neural model is designed to predict the future prediction of the subscriber based on the past predicted information of the subscriber. In this research work, a prediction based location management scheme is proposed for locating a mobile terminal in a communication without losing quality maintains a good response. There are various methods of location management schemes for prediction of the mobile user. Based on individual characteristic of the user, prediction based location management can be implemented. This work is purely analytical which need the past movement of the subscriber and compared with the simulated one. The movement of the mobile target is considered as regular and uniform. An artificial neural network model is used for mobility management to reduce the total cost. Single or multiple mobile targets can be predicted. Among all the neural techniques multilayer perceptron is used for this work. The records are collected from the past movement and are used to train the network for the future prediction. The analytical result of the prediction method is found to be satisfactory.