建置以“航段”為基之空服員整合型排班決策支援系統

Abstract

過去受限於電腦速度慢與空服員排班問題複雜性，傳統上分成組員勤務組合產生與組員指派兩階段求解空服員排班問題，導致人員指派效果不佳，且因短程航線問題特性造成排班單位過大，指派不易，因此本研究考量台灣航空業實務運作之限制，將組員勤務組合產生與組員指派兩階段整合爲一階段求解，建立一個以「航段」爲排班基礎之最佳空服員排班整合模式，以提升排班接續性和人員利用率。本研究利用基因演算法求解空服員排班規劃與指派問題，並與分枝界限法所得之最佳解比較，結果顯示本研究所提出之排班整合模式與求解演算法，可有效地處理空服員排班問題。最後本研究以Java 程式語言爲基礎，設計一套整合型排班決策支援系統，可規劃符合經濟效益的空服員排班表，並有效地解決空服員調度與空載和過夜問題。

To improve the accuracy of cabin crew scheduling, we present a new solution model that integrates the crew pairing problem and the crew rostering problem into one stage, conducts the scheduling by leg, and searches for the global optimal solution. Referring to the objectives and constraints of Taiwan's airline business, we decide the levels of parameters via the design of experiment and then apply the genetic algorithm to solve the concerned problems. We compare the result of genetic algorithm with that of the branch and bound algorithm to show the model's fidelity. At the end, an integrated crew scheduling decision support system in JAVA language is also proposed. We claim the system can potentially assist to plan a crew timetable and to solve the crew arrangement and the deadhead and layover problems efficiently.