Research On Cross-layer Optimization Algorithms For Wireless Mesh Networks Based On Network Utility Maximization

As the economy grows and the society goes forward, higher rate, more different operations and more reliable service are required for wireless mesh networks. Though the layered design methods reduce the complex, and improve the level of modularization and standardization, the network performance only achieves local optimization, not the global optimization. This reduces the network efficiency. In wired networks, layered design methods covers up the disadvantage of negative effect due to the large capacity and high reliability of wired networks. But in wireless networks, especially in the wireless network which demands strictly QoS, since the time-varying status of wireless link and network topology, the network performance cannot meet the application demand if the network is designed by layered methods. So cross-layer design methods are suit for the wireless networks which are constrained by the strictly QoS.The cross-layer optimization based on network utility maximization is a systemic method for the cross-layer design. It formulates the network performance of each network layer as the objective function and constraints of the mathematics optimization problems, in which the function variables denote the network resource, such as time slot, frequency, power, link rate, energy etc.This paper mainly studies the cross-layer resource allocation of multi-radio multi-channel mesh networks, and formulates the problems as network utility maximization. To solve them, this paper also proposes the according distributed algorithms. The main contents include jointly rate control and power allocation optimization, joint congestion control, channel assignment optimization and the cross-layer optimization incorporating the stochastic effects. The details are as following:1) Jointly rate control and power allocation optimization in mesh networksBased on the convex optimization, a cross-layer optimization model for jointly rate control and power allocation in multi-radio multi-channel wireless mesh networks is proposed. The proposed algorithm attains the trade-off between the network utility and the total power by varying the intrinsic weight, and dynamically adjusts the injected rate and transmitting power to achieve the joint optimization of the network utility and power. 2) Jointly congestion control, scheduling, channel assignment optimization in mesh networksThis part studies the cross-layer approach with a joint consideration of multi-path routing, congestion control, scheduling, radio allocation and channel assignment. The problem is formulated as a mixed-integer non-linear programming (MINLP), which normally requires a centralized solution and hence has prohibitively high computation complexity. A two-phase distributed mechanism is proposed for reducing the computation load and communications overhead.3) Jointly rate control and power allocation optimization incorporating the stochastic effectsThis part studies the stochastic network flow and time-varying wireless environment in mesh networks based on the stochastic network utility maximization, and uses the genetic algorithm to solve the stochastic problem.The dissertation solves some key problems for mesh networks based on the network utility maximization. The works of this dissertation have some theoretical and practical significance.