基于模糊免疫策略的机场航班延误自适应实时预测方法

摘要

To make an accurate and real-time prediction for flight delays, referring to the thought of fuzzy membership grade, a fuzzy immune prediction method is proposed. The concepts of "self", "nonself", "antibody" , "antigen" and a membership degree function for antibody to the delayed flight set are defined. During the prediction process, through real-time monitoring changes of the environment and self-adaptive adjusting A-cut sets, the paper chooses detectors according with the current confidence level, and overcome the limitations of the deterministic model. The simulation experiment on a domestic airport shows that the method is adaptive and can accurately predict delayed flight numbers of each period all day.%为了对航班实时延误情况做出准确的预测,在生物免疫检测的基础上,借鉴模糊隶属度的思想,提出了模糊免疫预测方法.给出了自体/非自体、抗体/抗原的定义以及抗体对延误航班集合的隶属度函数.在预测过程中,实时监测外界环境变化,自适应调整λ-截集,选取符合当前置信度的检测器,克服了基于确定性模型的局限性.对国内某大型机场做仿真实验表明,该方法比简单使用生物免疫方法更具适应性,能较准确地预测全天各个时段航班的延误数量.