CONGESTION CONTROL FOR EFFICIENT RESOURCE ALLOCATION IN AD HOC NETWORKS WITH MULTIPATH SESSIONS

In this paper, based on the inherent characteristic of the contention relation between flows in ad hoc networks, we introduce the notion of the link's interference set, extend the multipath utility maximization problem formulating the multipath congestion control problem in wireline networks to ad hoc networks. As the objective function is not strictly concave in the path rate, we apply the penalty function method to take into account the constraints and transform the primal problem into a new formulation, obtain the optimal solution based on the subgradient method, and propose the multipath congestion control algorithm MPCC. MPCC is distributed, scalable and low-overhead. The simulation results illustrate that it can quickly close to the globally optimal solution. Furthermore, since its high-speed convergence can cope with changes in network conditions in most cases, it can be used to the practical ad hoc networks.