Interference pricing for self-organisation in OFDMA femtocell networks

We propose a distributed technique for downlink interference management in overlay femtocell-macrocell deployments operating in the same channel under OFDMA access. The proposed target function is the weighted sum rate, which preserves relative QoS between users. Being the overall problem non-convex in terms of transmitted power, we propose to solve it distributedly, where each femto access point (FAP) optimizes a convex function for some measured interference levels. If the optimization is done independently at each FAP, we can formulate the overall problem as a non-cooperative game where each action is taken as a selfish response to the own FAP channel states and to the interference generated by other FAP decisions. The resulting iterative waterfilling algorithm provides a local optimum solution which is known to be far from optimal efficiency, in general. Alternatively, a better solution can be found by coordinating the FAPs in terms of exchanging interference prices. Assuming that the air interface could provide much larger spectral efficiency than the Internet Service Provider (ISP) wired connection, the problem solved at each FAP must include restrictions on the sum rate perceived by the femto users (FUEs). Numerical results compare gains obtained by pricing exchange and from pure competition in dense FAP deployments.