Kalman filter based capacity adaptation for network profit maximization

Estimation of traffic demand is a major requirement in numerous telecommunication network applications. As traffic level typically varies with time, online applications such as dynamic routing and dynamic capacity allocation need to accurately estimate traffic in real time to optimize network operations. While optimizing network capacity, the effect of estimation error on Grade of Service must also be considered. In this paper, we propose an estimation approach based on the Kalman filter where the model parameters are based on historical data. This estimation is used to adapt the network capacity with the objective of network profit maximization under the connection blocking constraints. Performance of proposed approach is compared to estimation based on adaptive exponential smoothing. The results show that our approach gives better network profit together with enhanced Grade of Service.