企業資訊部門之軟體專案接案決策模式─一個多重專案與多期之專案分派模式

Abstract

隨著資訊時代與科技的普及之下，企業內持續的資訊化需求已成為企業資訊部門重要及繁重的業務。在軟體及資訊系統開發專案化之環境中，資訊部門在面臨源源不斷、大小不一的專案需求，如何以有限人力進行專案人員的指派及調度，是資訊部門重要之管理議題。因此，本研究提出一個多重專案與多期專案分派模式（Multiple-project Selection, Multiple-period Assignment Model，簡為MPP模式）來處理上述的資訊部門中軟體接案管理問題。在考量軟體專案當中多重及多期的人員指派情況，本模式在執行上為事件導向，在每一期中依據上期的指派結果加上本期的新進專案加以重新計算，並搭配田口損失函數來表示專案因延遲或未排入開發所造成的機會成本，透過混和整數規劃計算出最佳化的開發指派結果，以達多期再規劃的目的。此外，本研究舉出一個企業應用實例說明本研究模式的使用情形與討論研究限制。

With the popularity of the information technology, the needs of continuous, ad-hoc information system (IS) development in enterprise have become critical and heavy work to IT departments. In the context of project-oriented software and IS development, IT departments have been facing endless and variety of projects requirements. How to continually and effectively arranging the projects with limited manpower becomes an important issue of IT department. Therefore, this study proposes a multiple-project selection, multiple-period assignment (MPP) model prototype to handle the problem of software project selection and assignment. The MPP model is event-driven in operation. In each period, the model re-computes based on the results of last period and the incoming projects of current period, together with Taguchi Loss Functions which represents the costs from past project delay and the possible delay due to un-assignment, all of which is calculated through mixed integer programming (MIP) for the optimized results of developer assignments and corresponding development rates. Furthermore, this study illustrates a case of multiple-project selection, multiple-period assignment for software projects. The actual usage patterns are discussed.