以限制理論為基礎之專案排程及執行控制研究

Abstract

專案為一有明確目的、生命週期及特殊性的為了完成特定目標的組織活動。專案工作排程的方法自要徑法及計畫評核術之後，發展至今已有數十年的基礎，PERT/CPM利用網路圖的技術，將專案中的要徑求出，根據專案中的要徑來排每一個工作的排程，對於專案中資源的衝突都假設資源充分供應的情況。現今專案愈趨於複雜化，專案的工作排程不只要考慮工作間的前、後置關係，還要將資源限制及專案中的不確定因素考慮在內。專案的不確定因素包括專案工作路徑的變異、會合路徑的變異以及一些人為的因素（如學生症候群、多重工作現象、不會提早將工作完成報告出來），而PERT/CPM所發展的專案排程及控制方法常忽略不確定因素考慮的影響，使得專案排程在執行時並不能提供一準時的完工時間。本研究將以限制理論為基礎，將不確定因素考慮在專案排程內，推導出單專案及多重專案的排程與執行控制的方法。 再者，針對排除專案中資源的競爭問題，已有很多最佳的專案排程方法被使用，本研究將使用一資源限制下專案排程最佳解方法（ACTIM），求得一專案資源限制下之最佳排程，並將其與限制理論專案排程結合，利用程式模擬實際專案執行情況，並與一般任意專案資源競爭下之排程相比較，分析專案執行結果，以作為限制理論專案排程是否需先費力求得專案最短排程的依據。本研究專案排程模擬結果，專案最短排程對於工期為偏態機率函數（如Lognormal）時，其工期及會合路徑的變異性大，因而吸收最短排程所節省之排程規劃時間，從模擬結果中可看出並未增加限制理論專案執行的效率。

A project can be defined as a specific, finite, and unique task with a well-defined purpose and life cycle. Project scheduling methods have been developing for decades since the PERT/CPM technology of network. PERT/CPM focus on critical path of project. Every project's job is arranged in project scheduling according to the Critical Path. The PERT/CPM assumption of infinite capacity solved resource contention problem. Now the project is more complex. Project scheduling is not only considering relation with the jobs between forward and backward position on scheduling. But also be considered something about resource contention problem, project path variability, variability of the path merged point and some working behaviors ( example student syndrome、 multiple tasking、early report etc. ). PERT/CPM have neglected to do with these uncertainties in the project scheduling and controlling. Thus, project scheduling executed failed to provide realistic completion dates. In the past, the scholars had written a lot of articles about find the optimum project scheduling on resource limited. The study will based on Theory of Constrain ( TOC ) to add uncertainty in project scheduling and then develop a scheduling method for single project and multi-project. Furthermore, there are many methods exploring optimum scheduling regarding the resource contention problem in project scheduling. The study will use a method that can find the optimum project scheduling under resource limited. And then, the study finds common another project scheduling. Combined they with TOC project scheduling method. We use computer program to simulate the actually condition of executing project scheduling. The study will compare optimum with common result of scheduling simulations. The report can provide people with decision standard when they used TOC project scheduling. We find that optimum project scheduling used skew distribution of task duration. The skew distribution has more variability of task duration and merged path. It attracts optimum scheduling reduce time. In the process, it tells us that optimum project scheduling doesn't make TOC project scheduling more efficient.