A functional Link Artificial Neural Network for location management in cellular network

Mobility management is one of the major issues in mobile networks to provide an efficient and low-cost service. In this paper, we intend a prediction-based location management scheme for locating a mobile host (MH) or mobile station, which depends on its history of movement pattern of a mobile subscriber. A multilayer neural network (MNN) model for mobile movement prediction is designed to predict the future movement of a mobile host. For predicting the location of a mobile host the MNN is trained with respect to the data obtained from the movement pattern. The difficulties associated with the location management can be solved by nonlinear neural network that is computationally efficient. The major issue in feed forward neural network such as Multilayer perceptron (MLP) trained with back Propagation (BP) is that it requires a large amount of computation time for learning the network. Functional Link Neural Network (FLNN) is proposed here and that is simpler than MLP-BP. This is basically a single layer structure in which nonlinearity is introduced where the input pattern is enhanced with nonlinear functional expansion. The novelty of the proposed work is it requires less computation than that of MLP-BP. With proper choice of functional expansion in case of FLANN, this network performs better than multilayer perceptron with back propagation. It is observed from the simulation result that FLANN outperforms MLP-BP in terms of performance error. It is also shown that proposed network is computationally cheap and gives better classification accuracy than that of MLP classifier.