A Fast Convergent Cross-layer Resource Optimization Allocation Algorithm in Wireless Multi-hop Networks

In order to improve the performance of the large queue backlogs and low convergence rate in back pressure routing algorithm, the cross-layer optimization of joint congestion control, multi-path routing and power allocation in wireless multi-hop networks is investigated. The system is modeled as a network utility maximization problem under the constraints of flow balancing condition and power. Based on the Newton's method, the problem is solved and an algorithm with superlinear convergence speed is proposed. With matrix splitting technology, the algorithm can be implemented distributedly further. The simulation results show that the algorithm can effectively increase the energy utility while achieving the maximum network utility, and can keep the queue length at a very low level to decrease the packet transmission delay.