doiSerbia

Identifying key node in multi-region opportunistic sensor network based on improved TOPSIS

Liu Linlan (School of Information Engineering, Nanchang Hangkong University, Nanchang, China), [email protected]
Wang Wei (School of Information Engineering, Nanchang Hangkong University, Nanchang, China), [email protected]
Jiang Guirong (School of Software, Nanchang Hangkong University, Nanchang, China), [email protected]
Zhang Jiang (School of Information Engineering, Nanchang Hangkong University, Nanchang, China), [email protected]

The topology of multi-region opportunistic sensor networks is evolving, and it is difficult to identify the key nodes in the networks by traditional key node identification methods. In this paper, a novel method based on the improved TOPSIS method is proposed to identify the key node from the ferry node. The dynamic topology information is represented by the graph model which is modeled by the temporal reachable graph. Based on the temporal reachable graph, three attributes are constructed to identify the key node, which are average degree, betweenness centrality and message forwarding rate. The game theory with a combination weighting method is employed to combine the subjective weight and objective weight, so as to obtain the combined weight of each attribute. The TOPSIS method is improved by the combined weight. The key node is identified by the improved TOPSIS. The experiments in three simulation situations show that, compared with the TOPSIS method and MADM_TOPSIS method, the proposed method has better accuracy for the key node identification in the network.

Keywords: multi-region opportunistic sensor network, key node, combination weight, TOPSIS