基于多指标综合的复杂网络节点重要性评估

摘要

为全面综合评估节点的重要性,提出一种基于D-S证据理论的多指标综合的复杂网络节点重要性评估方法.通过对复杂网络基本理论的研究,定义以度中心性、接近中心性、介数中心性为代表的显著性评估指标和以韧性度、凝聚度为基础的破坏性评估指标,从各个角度描述节点的重要性;在此基础上,利用D-S证据理论在融合多数据源方面的优势,建立多指标综合的辨识框架和概率分配函数,用Dempster合成规则实现节点多指标的综合评估;综合评估结果排名越靠前,节点越重要.实例分析结果表明,该方法综合了多指标的优势,全面准确地刻画了节点的重要性,为进一步有效利用节点奠定了基础.%To estimate the node importance comprehensively and synthetically,an estimation method based on multi-index and D-S evidence theory was proposed.Through studying the basic theories of complex networks,the significance index represented by degree centrality,closeness centrality and betweenness centrality as well as the destructiveness index represented by tenacity de-gree and condensation degree were defined,which described the importance of node in every aspects.On this basis,taking advan-tage of D-S evidence theory in fusing multi-data,the frame of discernment and basic probability assignment for the multi-index were built.The node multi-index comprehensive estimation was realized using the Dempster fusion rule.The comprehensive esti-mation result ranked higher meant the node was more important.The case analysis shows that the proposed method describes the node importance comprehensively and synthetically with the advantage of multi-index,which establishes the foundation for fur-ther utilizing the node effectively.