Time-Clustering of Load in Mobile Networks

Mobile network operators are faced with congestion problems on cell capacity that degrade the quality of service (QoS). Must often these congestions appear during few hours in the daytime. A pricing policy per period is presented there as a solution to the temporal redistribution of the network load. For that purpose, we use a time pricing model to smooth the load during the day. The problem is then to determinate the right time pricing. The contribution of this paper consists in proposing a method to partition the day in several periods of different prices. This work is based on the analysis of the current state of the network load and its new state after the time pricing optimization. We present this method and we test it on a small urban network of 140 cells