需求變動下之航空貨運網路規劃

Airline Freight Network Planning under Demand Variation

航空公司營運路網為影響營運成本的重要因素，而飛航網路規劃會隨預設的需求水準有所不同。大多數的做法是以規劃期間內的平均需求為基準，設計相對應的營運路網；然而，航空網路市場的需求常有明顯的季節性變動，且尖峰、離峰差異頗大，以平均值為依據的結果並無法適用於大部分的需求情形。本研究直接將需求變動納入路網規劃模式的考量因素，以隨機數學規劃方法，建構變動需求下之飛航網路規劃模型。所提模式可決定最適網路型態（直航、軸輻式或兩者混合），以及不同需求水準下，相對應的飛航路徑、貨物分派於各路徑的流量。最後利用海峽兩岸主要機場的貨運資料對所提模式進行測試，並利用隨機數學規劃中的評估指標，比較確定性模型（給定需求平均值）與隨機性模型（考量需求變動）間的差異，結果顯示隨機模式可獲得較佳的網路結構。

The operational costs of airline companies are highly dependent on their service networks. The airline network is usually designed based on an estimated demand. Therefore, the network design would be different according to different given demand levels. The most commonly used method is to employ the average demand as the design basis. However, the demand has obvious seasonal fluctuations in air freight markets and the gap between the peaks and off-peaks are usually significant. Hence, the network design based on the average demand cannot guarantee to provide the most efficient service in various demand levels. This study addresses the airline network design under demand variation by applying stochastic programming technique. The proposed model can determine the optimal network structure (direct flight, hub-and-spoke, or mix of both), flight routes, and flow fractions of routes for different demand levels. The data collected from the air freight market between Taiwan and Mainland China were used to test the proposed model. Finally, quantitative indices were used to compare the traditional deterministic model (based on a given demand level) and the proposed stochastic model (considering demand variation). The results show that the stochastic model can obtain a better solution.