A novel learning mechanism for traffic offloading with small cell as a service

The densification of mobile networks with small cells is seen as the most promising solution to the explosive data traffic increase. Due to their financial implementation requirements, which could not be met by the service providers, the emergence of third parties that deploy and lease small cell networks opens up new business opportunities. In this paper, we study a proportionally fair auction scheme as an efficient way of small cell capacity distribution, both in network and financial terms. To improve the bidders´ strategies, we propose a novel learning mechanism that alleviates the uncertainty incurred by variations in the traffic and the lack of information in the auctions. Extensive simulations prove the efficiency of our proposal, which also performs in equal terms with the ideal case of complete information.