考慮公平性之後艙組員派遣模式 ---以座艙長為例

Abstract

航空公司組員排班問題可分為「勤務組合產生」和「組員派遣」兩個階段，前者以最小成本及效率性為考量，後者則應該以組員的滿意度為考量。而兩者皆必須滿足航空公司所規劃之班表與機隊，並遵守民航法規以及工會所訂定之各種限制。然而過去相關文獻多是探討組員派遣成本的最小化，甚少以公平性之觀點來探討之。在公平性越來越受重視的趨勢下，如何讓組員的工作量能均等，進而提高組員的滿意度與士氣，亦是一項重要之課題。本研究之特點是從後艙組員的觀點出發，來建構一公平性成本函數，並以最小公平性成本為目標，產生組員公平之班表。 由於後艙組員派遣問題實屬於一複雜之數學規劃問題，因此實務上皆是利用啟發式解法來增加求解的效率。本研究將後艙組員派遣問題列式為一集合分割模式，並建構啟發式解法，分為三個階段來求解此問題。第一階段為長勤務指派，將飛行天數大於(或等於) 3天的勤務組合，在滿足各種限制以及最小公平性成本的目標下，全數指派給後艙組員；第二階段為2日短勤務指派，將所有飛行天數為2天的勤務組合，同樣在法規限制與最小公平性成本的目標下指派給組員。第三階段為1日短勤務指派，由於此階段問題變數過大，故依照本研究特別提出之解法，將飛行天數為1天的勤務組合全數指派給組員。相關求解演算方法，皆使用C語言撰寫程式，並以CPLEX 7.0數學規劃軟體為求解工具。 本研究並以國內C航空公司座艙長之原始班表資料作實例驗證，該問題原始變數高達1022，而本研究所提出的分階段啟發式解法，除了可以大規模的降低問題變數的規模外，其結果與C航空公司相較之下，總公平性成本有25 %的改進比率。而且在組員的個別飛航時間上，亦未超過C航實際指派的最大值78小時，因此在實務上並不會增加航空公司在飛行加給成本上的負擔。本研究提出之公平性派遣模式，不僅能產生出合法可行之公平性班表，進而增進組員的滿意度，也極適合作為航空公司在實際組員派遣上之參考。

Generally, the airline crew scheduling problem can be separated into two sub-problems of lower difficulty. One is “crew pairing problem” and the other is “crew rostering problem”. For the crew pairing problem, the object is to minimize the cost and construct an efficient set of pairings; for the crew rostering problem, it must consider the satisfaction of crew members. Both of them have to obey the flight schedules、flight regulations、operation restrictions and union rules. Nowadays, some airlines put more emphasis on equitability. So in order to increase the satisfactions of the crews, it’s an important issue to assign even workload to crew members. The feature of this research is that we generate the cabin crews’ schedules for the purpose of equitable rostering. The cabin crew rostering problem is a complicated MP problem. This research formulates it as a set partition problem and constructs a heuristic method to solve the problem in three phases. The first phase is long-haul pairing allocation. In this phase, we assign all pairings each with flight-day greater than two. The second is 2-day short-haul pairing allocation. In this phase, we assign the pairings each with flight-day equal to two. The third is 1-day short-haul pairing allocation. In this phase, we present a unique heuristic method to assign the pairings each with flight-day equal to one (including stand-by duties). In each phase, the objective is to minimize equitable crew rostering cost under all the restrictions and laws. The heuristic program is written in C language, and the CPLEX 7.0 is used as the MP solver. With a case study of 169 pursers and 1274 pairings of a Taiwan’s major airline, the results show that the total equitable cost decreases 25% comparing to the original equitable cost of the airline. The Block Hour in our roster is less than the airline’s maximal BH, so it won’t increase the airline’s real cost. It also proves that the new methods we propose can not only decrease the complexity and variables of the problem, but also construct new equitable and good quality rosters.