Abstract:

To solve the problems of significant communication overhead and limited real-time response capability in the traditional flooding-based state diffusion mechanism, an efficient diffusion algorithm based on regional partitioning and information compression was proposed. A multi-dimensional feature clustering method integrating geographic distance and link quality was used to achieve high-cohesion network partitioning through weighted Euclidean distance and the K-means algorithm; a dynamic selection mechanism for representative nodes was designed to conduct comprehensive scoring bycombining the link quality and geographic centrality of nodes, realizing the efficient compression and aggregated transmission of regional information. Simulation results indicate that the algorithm significantly reduces the redundant communication overhead of state information transmission in wireless ad hoc networks and improves diffusion efficiency, providing a reusable algorithm model and performance evaluation framework for the simulation modeling and situation generation of dynamic wireless ad hoc networks.

Key words: wireless ad hoc network, network state awareness, information diffusion, clustering algorithm