國內一家知名的國際航空公司,委託淡江大學開發了一套飛航操作風險評估系統(Flight Operations Risk Assessment System, FORAS),將航班在機組人員、飛航設備、航線狀況三大類的各種風險因素,以模糊推論(Fuzzy Inference )的方式,對航班飛航安全的風險進行評估,運算出決策風險值,提供飛行員以及安管人員作參考。影響一航班的飛航安全因素有很多,如果能找出其中影響力較大的關鍵因素以及在推論結構裡的路徑,可以在起飛之前對航班進行風險因素的調整,亦可將分析的結果進行統計與分析,無疑對於航空公司在飛航安全事故的預防,以及未來的趨勢規劃上都有很大的幫助。 本研究經過分析風險因素的特性以及對決策風險值的影響方向,歸納出四類的關鍵風險因素,並以Branch and Bound演算法為基礎,開發出可在FORAS推論架構下,搜尋可讓決策風險值變得更好或是變得更差的關鍵因素與路徑的演算方法,並開發出推論的系統,進行系統實作的驗證,經過與合作航空公司的測試之後得到了令人滿意的結果。
The Flight Operations Risk Assessment System (FORAS) developed by Tamkang University is a risk assessment tool that will estimate risk associated with a flight operation. This risk is estimated using a fuzzy inference model that synthesizes a variety of inputs, including information on crew, aircraft, and flight sector. This study provides a methodology based on the Branch and Bound algorithm to identify those inputs that contribute most significantly to the estimated flight risk. According to the proposed methodology, the risk inference chain can be highlighted and be able in some cases to suggest possible interventions for the flight safety managers to prevent high flight risk. To verify the feasibility and usability of this methodology, a tool has been developed and tested in an international airline in Taiwan. The testing results are very impressive.