Contracts for Joint Downlink and Uplink Traffic Offloading with Asymmetric Information

Motivating the potential offloading nodes (PONs) to cooperate with the offloading request nodes (ORNs) is an essential issue of traffic offloading in 5G. Nevertheless, in the coupled traffic offloading case where the PONs' uplink and downlink channel state information (CSI) of both the access and fronthault links is unknown to the ORNs, it is very difficult for the ORNs to provide an appropriate reward which is consistent with the PONs' effort. In this paper, we propose a contract-based framework to tackle this challenge, where the PONs are agents characterized by a two-tuple type, while the ORNs are the principles providing contracts in the form of (offloading quality, monetary reward). Since it is a tradeoff between offloading quality and monetary reward, when designing the contract, the ORN integrates these two factors into the offloading utility function, and then optimizes the utility under the constraints of individual rationality and incentive compatibility conditions. Through mathematical analysis, the necessary and sufficient conditions of the formulated problem for the one-ORN scenario is simplified. Then a gradient descent algorithm is proposed to find the optimal solutions, i.e., the optimal contract. Furthermore, the study on the PON cooperation stimulation problem is extended to the multi-ORN scenario, where the decision-making process of the ORNs and the PONs is formulated into a matching game. To find the stable solution to the matching game, a revised deferred acceptance algorithm is proposed and then proved to be convergent and have low computational complexity. Simulation results demonstrate that the proposed scheme achieves higher offloading utility and energy efficiency compared with the existing schemes.