Wireless networks are progressively displacing service provisioning towards the edge to accommodate the increasing traffic demands. This paradigm shift calls for smart policies to efficiently share network resources and ensure service delivery. In this talk, a concept of bidding and tipping for spectrum auctioning will be presented. In this concept users iteratively revise their strategies, which include bidding for spectrum, route selection and relaying incentives, to achieve their quality of service requirements. Further, a reinforcement learning algorithm will be discussed to automatize the bidding process based on users¹ previous experiences. Then, a collaborative cognitive dynamic network architecture (CDNA) will be outlined where mobile devices share their connectivities acting as access points for other users and harvest available spectrum.