摘要
依据飞行系统实际,以CPC因子的内在依赖性和同类因子的非负相关性为基础,通过在免疫克隆选择算法中,增加抗体净化环节,设置抗体群亲和力函数,建立了模糊规则库优化模型,用于去除FCREAM中的冗余规则,从而完成对方法的改进。采用回看分析法,通过改进前后的FCREAM对"韩亚航空214事故"进行预测,并对预测结果进行对比分析。结果表明,改进的FCREAM与实际较为匹配,且具有较好的连续性,同时敏感性较改进前得到增强,是一种有效的飞行员人因可靠性定量预测分析方法。
FCREAM is improved for quantitative prediction of pilot＇s reliability.On the basis of flight system,the inter-dependency and the same kind of factors＇ uncorrelation of CPC are analyzed.Antibody cleaning of CSA is added and the objective function of the antibody-set is made,along with CPC＇s quality,the rule set of FCREAM is optimized.The retrospective analysis method is adopted to analyze and predict＂Asiana Airlines OZ214 flight accident＂by FCREAM before and after the improvement and then the predicted results are compared and analyzed.The results show that the prediction in the improved method is uniform with the reality and has strong continuity,meanwhile its sensitivity becomes stronger and is one kind of effective prediction method.
出处
《工业安全与环保》
北大核心
2016年第8期21-24,共4页
Industrial Safety and Environmental Protection
基金
国家自然科学基金(61102120)
