Manet On-demand Weighted Clustering Algorithm Based On Mobility Prediction,

Now, in the field of the research of Mobile Ad Hoc Network (MANET), the study of clustering algorithm has become one of the focuses. The paper analyzes and compares several typical clustering algorithms in MANET, especially on-demand Weighted Clustering Algorithm (WCA). WCA is a clustering algorithm which considers several factors synthetically and has better performance. But there are some flaws in selecting values about node's mobility and energy. It does not consider the relations of relative locations among nodes and the actual situation of energy consumption.For the flaws of WCA, we propose a Mobility Prediction based on-demand Weighted Clustering Algorithm (MPWCA). The algorithm adopts the link expiration time (LET) mobility prediction model and node energy model to solve the problem about selecting values about node's mobility and energy respectively. So it can improve the stability and survivability of the network. Then, it gives the process of foundation and maintenance of cluster in detail. We select the LET as main criterion in cluster head electing and adopt on-demand cluster maintenance mechanism. So that we can reduce the times of cluster head's update, keep the structure of cluster steadily and guarantee the performance of the network.By extending Network Simulation 2 (NS2), we design and implement MPWCA. Then we structure the network environment and compare MPWCA and WCA. In the experiments, we mainly observe the situation of the changes about performance of cluster when the transmission range of the node is changing. The simulation test shows, compare to WCA, MPWCA can reduce the times of rejoin of node and update of the dominion set which consists of cluster head. And the number of cluster head is more reasonable. The algorithm achieves the goal of improving the stability and survivability of the network.