摘要
风险评估通过分析不确定的风险因素得到确定的风险评价,如果存在多个风险,需要实施风险聚合.传统风险聚合方法依赖风险权值,在分布式环境中较难具有客观性,尤其对于移动环境,因为移动节点具有移动性和随机性.提出补偿竞争风险聚合算法(CCRAA),CCRAA的基本思想是模糊聚类,对风险值进行补偿以减少其与聚类中心的距离,使补偿后的风险值向聚类中心聚集,取最大风险值和最小风险值的平均值为聚合风险.CCRAA使风险和值不变,不影响聚合风险的大小,但避免了传统风险聚合方法可能产生风险极值或对风险权值的依赖.使用实验证明CCRAA具有优于传统方法的聚合效果和稳定性.
Risk evaluation gave a certain result by analyzing uncertain risk factors and risk aggregation was requested if there were multiple risks. Traditional risk aggregation method, which rested on the risk weights, was difficult to be objective in the distributed environment, especially in the mobile environment because mobile nodes were mobile and random. Compensational competing risk aggregation algorithm （CCRAA） building on the basic concept of fuzzy clustering is proposed in this paper, where the risk value is compensated to shorten its distance from the clustering center so that the compensated risk value converges toward the clustering center and the aggregated risk is determined as the average of the maximal and minimal risk values. CCRAA averts the dependency in the traditional risk aggregation method on risk limits or weights while it leaves the sum of risks unchanged and has no effect on the magnitude of the aggregated risk. A test is described to have demonstrated that CCRAA is superior to the traditional method with respect to the effectiveness in aggregation and stability.
出处
《系统工程理论与实践》
EI
CSSCI
CSCD
北大核心
2015年第8期2137-2143,共7页
Systems Engineering-Theory & Practice
基金
国家自然科学基金(61370073)
湖南省自然科学基金(12JJ6056)
关键词
风险评估
风险聚合
风险权值
模糊聚类
补偿竞争
risk evaluation
risk aggregation
risk weight
fuzzy clustering
compensational competition
