Constrained Opportunistic Power Control in Wireless Networks

In the opportunistic power control algorithm (OPC), designed in [1], [2], each user tries to keep the product of its transmit power and its experienced effective interference to a constant, called the target signal-interference product (SIP). This increases the transmit power when the channel is good and reduces it when the channel is poor (opportunism). It has been shown that the OPC always converges to a fixed point irrespective of whether the power is constrained (where there is no upper bound on transmit power) or unconstrained (where an upper bound on transmit power is taken into account). It has also been shown via simulation that the throughput achieved by the unconstrained OPC is significant (as compared to other existing distributed schemes), and is an increasing function of the target-SIP set by users, in the sense that a higher target-SIP results in a higher throughput. In this paper, we show that in contrast to the unconstrained OPC, when the constrained OPC is applied, not only the throughput is not necessarily increased as the target-SIPs increase, but it may even decrease if some users set their target-SIPs at high values. Furthermore, we propose a heuristic solution to determine the target-SIPs by users in a distributed manner. Our simulation results show that the throughput achieved by our distributed setting of the target-SIPs very closely approaches the maximum achievable throughput by the constrained OPC, and is very close to the global optimum value.