MOST: Mobile Broadband Network Optimization Using Planned Spatio-Temporal Events

Continuous optimization of Mobile Broadband Networks (MBNs) is done to simultaneously optimize and balance some performance factors such as Quality of Service (QoS) and energy usage. Initial MBN optimization approaches used various heuristics, while more advanced approaches have investigated the MBN optimization potential using predictions of future node loads based on historical load data. However, using only historical load data for prediction often fails when events occur. This paper presents MOST (Mobile broadband network Optimization using planned Spatio-Temporal events) approach that additionally uses external spatio-temporal information about planned events, e.g., the time and place of big concerts or football matches, to improve the prediction quality, thus enabling more correct MBN optimization. A simulation using experimental data from a mobile broadband network in Denmark shows that MOST is able to improve the prediction quality very significantly, and thus enables significantly better MBN optimizations to be performed.