Design And Analysis Of Adaptive Clustering Algorithm

In order to ensure the scalability and quality of service (QoS), the wireless mobile network usually adopts the hierarchical structure. And the life cycle and bandwidth utilization of network are largely dependent on the load balancing of its hierarchical topology.This article is based on"theory and technology of amorphous flattening self-organization wireless network", a project of the national natural foundation of China.According to the actual needs of the military, disaster relief and other emergency scenarios, and In order to adaptively improve the load balance of wireless ad hoc networks, thereby extending the life cycle of network,and improve the bandwidth utilization of network, an adaptive clustering algorithm for homogeneous nodes: Adaptive Load Balancing(ALB) is proposed based on the Adaptive On-demand Weighting(AOW),which is a existing classical algorithms. And because of the specificity that the nodes move frequently, some practical functions of ad hoc networks need to be established on the fast positioning of nodes. A clustering-based localization algorithm without the anchor node network is proposed based on map-growing algorithms which is a existing classical algorithms. The above two algorithms are simulated by C ++ language platform. Their feasibility is verified, and their performance are compared with the existing classical algorithms.The main works and innovation points of this thesis are mainly reflected in the following three aspects:1?A scheme of cluster head selection is proposed.This scheme can effectively solve the cluster collision problem in non-overlapping cluster mode.Simulation results show that the scheme can enhance the robustness of the algorithm, reduce the randomness of clustering, and increase the load balancing factor to 4 to 10 times of the Adaptive On-demand Weighting algorithm, which effectively improves the overall performance of the network.2?A scheme of adding adaptive threshold to the weighted clustering method is proposed. Threshold method can effectively limit the number of nodes in a single cluster, improve the load balance of network.Adaptive threshold set more in line with the idea of weighted, and also the requirements of wireless ad hoc networks. The simulation results show that the threshold of the clustering algorithm based on weighting method can be set to a large interval, which is in accordance with the requirement of adaptive.3?A clustering-based localization algorithm without the anchor node network is proposed. The algorithm is a distributed algorithm,which reduces the recursion depth and accelerates the localization speed.Based on the clustering idea, the algorithm avoids the superposition of errors in the recursion process. And the algorithm uses the redundant information to correct the positioning result. Simulation results show that the algorithm can effectively speed up the positioning and reduce the positioning error by more than 10%.