Multi-community Opportunistic Routing Algorithm Based on Machine Learning In the Internet of Vehicles

In view of the limitations of the opportunistic routing algorithm that divides communities based on interest only considering single community at present, this paper proposes a multi-community opportunistic routing algorithm based on machine learning (MORAML). According to the characteristics of small range and high density of places visited by private cars, this paper analyzes the resident area of users based on clustering method, and takes the center of the area as the hot spot location. Aiming at the problem that K-means algorithm is difficult to select the K value, this paper combines the advantages of DBSCAN and K-means algorithm to design a hybrid clustering algorithm to calculate the hot location of users, and then the clustering method is used to achieve community division of users. Because users have multiple hot spots, they can belong to multiple communities. In the process of message transmission, MORAML makes full use of the multi-community characteristics and centrality of nodes to decide the next hop route, and sets up redundant copy removal mechanism to reduce message transmission and avoid resource waste. Experiments show that MORAML has higher delivery rate, smaller network load and better performance compared with the existing classical routing algorithms.