A neural network demand prediction scheme for resource allocation in cellular wireless systems

Third generation cellular networks provide users with a variety of services such as multimedia and interactive applications that have ever increasing bandwidth requirements. In wireless systems, bandwidth is a scarce resource that needs to be used in an effective manner. Subscriber mobility also makes resource allocation a challenging task. In this paper, we propose an adaptive dynamic resource allocation policy based on a neural network demand prediction scheme for cellular wireless systems. Unlike some other proposed schemes, we do not require knowledge of the position and velocity of every mobile subscriber in the network. This greatly reduces signaling and computational overhead and results in a policy that is simple to implement. We perform a simulation study to compare the proposed scheme with other adaptive resource allocation schemes.