摘要 良好的航空貨物收貨人員供給規劃，可幫助航空貨運公司有效率地完成收貨工作，以確保達成顧客的需求及維持良好的服務品質。以往航空貨運公司對於收貨人員的排班，大多以人工經驗進行，以預測之平均人力需求值，進行人員供給班次表之建立。然而此作法忽略了實務營運上，每日的人力需求的隨機變動情形，若隨機性過大，則可能使原規劃的人員供給班次表失去其最佳性。此等最佳化人員供給班次表在營運中受隨機世界擾動的影響，在過去甚少為實務界或學術界所探討。有鑑於此，為能反映真實之需求狀況，本研究以航空貨運業者的立場，考量現實環境中貨運量變動的不確定因素，針對長期規劃與短期規劃上不同之限制條件，各自發展多個確定性需求與隨機性需求之航空貨物收貨人員排班模式。本研究模式期能於未來實務的應用上，提供航空貨運業者一有效的工具，以期幫助航空貨運業者在隨機環境中，規劃最佳的人員供給班次表。 本研究考量C業者現行之收貨人員排班方式，以系統最佳化的觀點，輔以「彈性的人事安排」及「數量彈性化」的彈性策略，構建確定性及隨機性需求排班模式。此後，本研究進ㄧ步把確定性需求模式中之固定平均需求值修正為隨機需求值，建立以模擬為基礎之隨機性需求排班模式。此二模式皆為線性混合整數規劃問題，屬NP-hard問題。求解上，本研究利用模擬技巧及數學規劃軟體，以有效率的求解各模式。為評估隨機性需求排班模式所求得的人力供給規劃班次表，於隨機環境中營運的績效，本研究發展一模擬評估方法。最後，為測試本研究模式的實用績效，本研究以一C業者營運資料為例，利用C程式語言，配合CPLEX數學規劃軟體進行求解，並進行範例測試與分析，進而提出結論與建議。 ABSTRACT A good manpower supply plan of receiving officers helps terminals dealing with their cargoes efficiently and maintaining excellent service quality. In the past most of these plans depend on staffs’ experience to establish the scheduling. However, they neglect the random fluctuation of manpower demands so that can’t respond to unexpected situation optimally. Since studies in this field are quite lacking, this research considered actual manpower demands and aimed at establishing several deterministic-demand scheduling models and several stochastic-demand scheduling models for their long-term and short-term operations based on the perspective of terminals. The models are expected to be effective tools for terminals to establish their manpower supply scheduling. We developed the deterministic-demand scheduling model by using different flexible strategies such flexible manpower supply and numerical flexibility. Further, we established the stochastic-demand scheduling models by modifying the fixed demand parameters in the deterministic-demand scheduling models. These models are formulated as linear mixed integer program. Furthermore, we used the simulation-based skill and employ a mathematical programming solver to solve problems. To compare the performance of these models under stochastic demands in actual operations, we developed a simulation-based evaluation method. Finally, to evaluate the models, and the evaluation method in practice, we performed a case study using real operation data from a Taiwan air cargo terminal, with the assistance of C computer programs and a mathematical programming solver. Then show the conclusions and suggestions.
