Research On Routing Technologies Based On Load Balancing For Wireless Mesh Networks

With the rapid development of the wireless communication and information networks, wireless mesh network has become a new development direction of the wireless communication field which has the advantages of flexibility, high reliability and better performance. Thus wireless mesh networks have become the key technology for solving the "last mile" bottleneck problem. And they have broad application prospects in the field of military, environment, health care, family and so on. Routing technologies based on load balancing is a key technology that makes nodes have higher throughput and improve the bandwidth utilization, and thus can increase the network capacity and balance network resources. So research on routing technologies based on load balancing for wireless mesh networks has a great significance. There are three main aspects on load balancing technologies in wireless mesh networks:(1) Load balancing among gateways in the wireless mesh network, (2) Single-path routing technologies based on load balancing, and (3) Multi-path routing technologies based on load balancing. In this dissertation, these three main aspects are the starting points to achieve load balancing in wireless mesh networks. Aiming at the above problems, the main research works achieved are as follows.(1) A gateway placement algorithm based on load balancing is proposed. Aiming at the load balancing problem, a force-based greedy heuristic clustering algorithm is proposed in this dissertation. The algorithm selects the least number of nodes in a given candidate gateway node set as gateways, and uses the gateways as the cluster head to clustering. The attractive force in the algorithm indicates the relationship between the ordinary members in a cluster and the cluster head. The repulsive force in the algorithm indicates the relationship between the ordinary members within each other’s interference range in a cluster. So, while calculating the relationship between the ordinary members in a cluster and the cluster head, not only the attractive force but also the repulsive force needs to be considered. On this basis a gateway placement algorithm based on hybrid particle swarm optimization algorithm for achieving the optimization between the number of gateways and the load balancing is proposed. Simulation results show that the algorithm proposed in this dissertation can achieve load balancing among gateways effectively.(2) A hybrid routing algorithm based on load balancing is proposed. In order to solve the load balancing problem in wireless mesh networks, a routing algorithm based on the leftover load rate is proposed at first. And then it is applied to the routing protocol based on the combination of genetic algorithm and ant colony algorithm. The algorithm takes advantage of ramdomly searching, and fast and global convergence of the genetic algorithm to generate the initial solutions, and then transform them into the initial pheromone distribution of the ant colony algorithm. Then the algorithm takes advantage of the features of parallelism, positive feedback mechanism and high problem-solving efficiency to find the optimal solution. Simulation results show that the algorithm can achieve load balancing among mesh routers effectively, improve the network throughput and the network transmission performance.(3) A multi-channel routing algorithm based on load balancing is proposed. Load balancing, link interference and channel switching delay are not considered in the WCETT routing metric. Aiming at the above problem, a centralized load-aware multi-channel routing algorithm is proposed. The algorithm builds a conflict graph according to the network topology first, and then calculates a weight value for each node in the conflict graph according to the initial traffic load and the potential interference link set. At last, an improved routing metric LBI_WCETT based on the WCETT routing metric is proposed, which considers link interference, load balancing, channel distribution and the delay brought by channel switching during routing. Simulation results show that the multi-channel routing algorithm and the LBI_WCETT routing metric can help increase the network capacity effectively, reduce the average end to end delay, and improve the network performance.(4) A multi-path routing algorithm based on clustering is proposed. Aiming at the high complexity and large overhead problems, a multi-path routing algorithm based on clustering is proposed. In this algorithm, aiming at the without cluster head and large overhead problems in the k-hop clustering algorithm, a new concept of node connectivity is proposed. Then by selecting the cluster head according to the node connectivity, meanwhile, with the idea of maximum connectivity, the k-hop clustering algorithm is improve. In the clustering process, the selection of border gateway nodes becomes critical, so the node which has smaller expected load is selected as a border gateway node. At last, according to the inter-cluster and intra-cluster routing policies, the algorithm selects multiple paths for data transmission. In the traffic assignment stage, the algorithm assigns traffic according to the virtual multi-path model. Simulation results show that the multi-path routing algorithm based on clustering proposed in this dissertation can reduce the network overhead effectively, improve the packet switching rate, and implement load balancing among mesh routers effectively.In this dissertation, the load balancing technologies in wireless mesh networks are studied in detail and some achievements are made, but further improvement and exploration are still necessary.