A distributed pricing algorithm for achieving network-wide proportional fairness

摘要

Proportional fairness (PF) scheduling achieves a balanced tradeoff between throughput and fairness and has attracted great attention recently. However, most previous work on PF only considers the single cell scenario. This paper focuses on the problem of achieving network-wide PF in a generalized multiple base station multiple user network. The problem is formulated as a maximization model and solved using the dual method. By decomposing the dual objective function, we get a distributed pricing based algorithm. Optimality of this algorithm is presented. Although the algorithm is derived using fixed link rate assumption, it can still apply in the presence of time-varying rates. The proposed algorithm is suitable for distributed systems in the sense that it does not need any inter base station communication at all. Simulations illustrate that the proposed distributed network-wide PF scheduling algorithm achieves almost the same performance as the centralized one. Compared with traditional local PF (LPF) scheduling, the network-wide PF scheduling achieves higher throughput, lower throughput oscillation, and greater fairness.