Cooperative communication has been shown to havegreat potential in improving wireless link quality. Incorporatingcooperative communications in multi-hop wireless networks hasbeen attracting a growing interest. However, most current research focuses on either centralized solutions or schemes limited to specific network problems. In this paper, we propose a distributed framework that uses Network Utility Maximization(NUM) to optimize the following joint objectives: flow control,routing, scheduling, and relay assignment; for multi-hop wireless cooperative networks with general flow and cooperative relay patterns. We define two special graphs, Hyper Forwarding Graphs (HFG) and Hyper Conflict Graphs (HCG), to represent all possible cooperative routing policies and interference relations among the cooperative relays respectively. Based on HFG and HCG, a stochastic mixed-integer non-linear programming problem is formulated. We then propose lightweight algorithms to solve these in a fully distributed manner, and derive the theoretical performance bounds of these proposed algorithms.Simulation results verify our theoretical analysis and reveal thesignificant performance gains of our framework, in terms ofthroughput, flexibility, and scalability. To our knowledge, thisis the first distributed cross-layer optimization framework formulti-hop wireless cooperative networks with general flow andcooperative relay patterns.
展开▼
