Neural network based adaptive radio resource management for GSM and IS136 evolution

With the evolution toward 2.5G bringing a wide range of new services, it is expected that the tele-traffic demand on current GSM and IS136 networks will further increase. In this paper we propose a new pro-active resource allocation method of increasing cellular network capacity by introducing an adaptive radio resource management system into a typical GSM/IS136 network. Adaptation is performed by using neural networks (NNs) to predict each cells future resource demands and adjusting the available resources accordingly. Results are presented which exhibit less resource requirements than existing fixed channel allocation (FCA) networks and performance that is comparable to previously proposed dynamic resource allocation (DRA) schemes, but with the advantage of significantly less complexity and no additional network signaling load