【摘要】 针对传统节点重要性评估只考虑拓扑方面，主观赋权法与客观赋权法分别具有限制性，多评估指标如何准确有效融合的问题，建立面向服务的节点重要性指标体系，提出基于粒子群的主、客观综合赋权法与基于余弦距离和灰色关联的逼近理想排序算法.结果表明，与传统方法相比，优先移除或修复本文方法评估出的重要节点对网络性能的影响更大，即本文的节点重要性评估算法更准确.更多还原

【Abstract】 Traditional node importance assessment only considers topologies, subjective and objective assignment methods have limitations. In order to integrate multiple assessment indicators accurately and effectively, the service-oriented node importance indicator system was established and the subjective and objective comprehensive weighting method based on particle swarm and an approximate ideal solution sorting algorithm based on cosine distance and grey correlation were proposed. The results show that removing or repairing the important nodes evaluated by the method has a greater impact on the network performance compared with traditional methods. The node importance assessment algorithm is more accurate than traditional algorithms.更多还原