隨著全球經濟的發展，各國的企業為了獲得更高的利潤必須更有效地利用有限的資源，隨著這些專案的規模逐漸成長，專安排程問題逐漸複雜化，在過去主要是利用要徑法與計劃評核術進行專案作業排程，但近年來資源限制的增加，單憑考量時程的控管而未考量有限的資源，將導致決策缺乏整體最佳化分析。因此本研究將應用網路流動技巧，並考量作業流程與資源限制及現金流量時間價值及相關作業流程，有效的以淨現值最大求解出最佳化專案排程，同時也把過去文獻求解僅求出近似解的缺點改善。 在過去的文獻中僅以確定性的觀念針對客戶端 (1)於作業完成付款(Payments at Activities′ Completion Times, PAC)及(2) 全額付款(Lump-Sum Payment, LSP)從事最佳化排程，忽略隨機工期對排程的影響或雖有考量隨機工期但非預期淨現金流量計算有誤，本研究納入考量作業工期的隨機性，並以最大化淨現值為排程目標，利用數學規劃方法建構PAC及LSP兩者的數學模型。此兩種模式均為一含有額外限制的網路流動問題，可利用(CPLEX)數學規劃軟體求解。本研究以國際測試題庫(PSPLIB)所提供之專案資料，進行範例測試，兩者的測試結果均良好，顯示在PAC及LSP兩種不同付款方式下，本研究所建構的隨機模式可成為學術界及實務業者之參考。 ;As companies started to run business globally, scale of projects grew larger. These projects became much complex because limitation of resources increased. In the past, researchers used to conducted project scheduling by critical path method and program evaluation and review technique. However, these researches may fail to be applied to huge-scale projects due to negligence of the limitation of resources raise. As a result, In this research, integer network flow technique are applied. Cash flow, sequence of works and limitation of resources were considered as factors influencing the results of optimization.
The purpose of this research is to optimize cash flow of projects. By building two models of project payment: first, Payments at Activities′ Completion Times, PAC, and second, Lump-Sum Payment, LSP. Researchers before hypothesized that the stochastic activity time does no influence to projecting scheduling, or considering the stochastic activity time ad a factor but failed to calculate the correct figures. Thus, this research utilizes network flow technique on two project payments. Two models are tested by the data referred to the Project Scheduling Problem Library(PSPLIB).
