Reverse spectrum auction algorithm for cellular network offloading

The explosion of mobile traffic and highly dynamic property often make it increasingly stressful for a cellular service provider to provide sufficient cellular spectrum resources to support the dynamic change of traffic demand in a day. In this paper, considering the dynamic characteristics of the cellular network traffic demands, we propose an optimal, truthful reverse auction incentive framework, which can minimize the leasing costs sustained by the mobile network operator at the premise of meeting the traffic demand of each time period. Such an issue is formulated as an Integer Programming (IP) optimization problem and we use an adaptive Lagrangian relaxation algorithm to solve the optimal reverse auction allocation problem. Besides, we propose a payment rule satisfying the truthfulness property (incentive compatibility) and the individual rationality property. Numerical results demonstrate that our proposed adaptive algorithm well captures the economical and networking essence of the reverse auction allocation problem, thus representing a promising approach to solve the optimal reverse auction allocation problem.