近年來油價不斷的上漲,其約佔航空公司營運成本50%~60%。因此,貨運機隊排程與班表規劃對航空業者的貨運營運績效而言,甚為重要。本研究以航空業者立場,在給定的營運資料下,包括機隊規模、起降額度、可用時間帶、相關飛航成本等,以總成本最小化為目標,並考量相關營運限制,建構貨機排程及班次表建立的整合規劃模式。 本研究利用網路流動技巧建立模式,此模式主要包含多重物流時空網路與機流時空網路,用以定式貨物與機隊在時空中的流動。物流時空依起迄點的不同構建多重起迄時對(OD-time pair)時空網路。機流時空網路則以整數流動方式定式機隊的週期排程。在物流時空網路與機流時空網路中及其間,再加上實務的營運限制,以符合實際的飛航作業。 此模式可定式為一混合整數多重網路流動問題,屬NP-hard 問題,問題規模龐大。因此,本研究以C++語言撰寫,利用啟發式解法-基因演算法求解。藉由本研究可求解出:1.最小總成本2.貨機最佳飛航路徑3.貨機班表之建立。本研究模式能於實務的應用上,提供一有效的工具,以輔助航空貨運業者在短期營運中規劃合適的航點、排程與班次表。最後本研究以一國籍航空公司之國際貨運航線營運資料為例,進行範例測試與分析,進而提出結論及建議。
In recent years the oil price unceasing rise, it has soared which occupies about 50%~60% operating costs to an airline, therefore, the result of air freighters fleet routing will affect carriers’ profitability in the market. Therefore, given the operating data, including fleet size, airport flight quota and available time slots, related flight cost, on the basis of the carrier’s perspective, this research tries to develop a scheduling model by integrating, cargo and freight flight schedules, with the objective of minimizing the operating profit, subject to the related operating constraints. The model is a useful planning tool for cargo airlines to determine suitable service airports, fleet routes and timetables in their short-term operations. We employ network flow techniques to construct the model, which include multiple cargo- and fleet-flow networks in order to formulate the flows of cargos and fleet in the dimensions of time and space. In the cargo-flow networks, different from that in the past research, we construct multiple OD-time-pair time-space networks on the base of cargos’ timeliness. In the fleet-flow networks, we use an integer flow network to formulate the periodical fleet routes. The model is formulated as an integer multiple commodity network flow problem that is characterized as an NP-hard problem. Since the real problem size is huge, this model is harder to solve than the conventional passenger flight scheduling problems in the past. Therefore, this research composes by the C++ language, develops the algorithm solution with the heuristic solution – genetic algorithm. Finally, to evaluate the model, we perform we perform a case study using real cargo operating data from a major Taiwan airline.