標題: | 新興公共建築工程計畫審議與經費核定模式 Reviewing and Budgeting Model for New Public Building Construction Projects |
作者: | 王翰翔 Han-Hsiang Wang 王維志 Wei-Chih Wang 土木工程學系 |
關鍵字: | 計畫審議;經費核定;效用理論;分析層級程序法;電腦模擬;類神經網路;Plan Reviewing and Budgeting;Utility Theory;Analytical Hierarchy Process;Computer Simulation;Artificial Neural Network |
公開日期: | 2002 |
摘要: | 政府預算是國家政策的總藍圖,其形成必須經過籌編、審議、執行,與考核等四個程序,而每年公共工程預算佔中央政府總預算中之比例超過四分之一,其重要性不僅可反應在政府藉其達成特定政治政策之目的上,更可藉著公共工程預算的投入,來擴大總體需求,進而降低失業率與刺激總體景氣。
公共工程委員會在公共工程的預算程序中扮演著計畫與經費專業審議的角色,但在現行計畫與經費審議工作上,並沒有一個制度化的作法說明應針對哪些計劃內容進行審議、計劃經費應如何合理地決定。因此,本研究從預算籌編過程中工程會專業審議的角度出發,建構一協助進行公共建築工程計畫審議與經費核定工作之模式。在模式的計畫審議部份,透過問卷設計與調查,首先確認了22項計畫審議項目與各項目之權重,並藉由個別項目效用函數的建立,進一步建立計畫整體之效用函數,讓審議者可藉著評定計畫在各審議項目上表現之完整性高低決定出該計畫之效用值;在模式的經費核定部份,我們針對個案計畫的成本組成項目進行分析,利用電腦模擬建立出一條合理之計畫成本模擬曲線。將前述之計畫效用函數與此成本模擬曲線結合,即可透過計算出之計畫效用值求得一建議之經費核定值。此外,本研究亦根據假設之數據,利用類神經網路說明另一可用以計算建議經費核定值的模式概念。本研究最後將以一個工程會實際審議的案例說明模式運作的流程,並對模式輸出結果作一比較說明與分析。 In every year, more than one fourth of the government budget is allocated to public construction projects. The amount of construction budgets greatly influences on macroeconomic policies and the quality of people’s life. The specialized reviewing and budgeting job performed by the Public Construction Commission (PCC) on these construction projects plays an essential role. However, the current system can’t clearly illustrate what items should be reviewed and how much the amount of money should be budgeted. Therefore, this study proposes a utility-based model to support the reviewing and budgeting jobs. 22 project reviewing items and their weights are first identified through questionnaires and expert interviews. The PCC engineer can review the project by checking the completeness and integrity of these reviewing items. After establishing the utility functions for each items and the whole project, and deriving a simulated cost distribution of the specific project, a particular of amount of budget is suggested to the model user (i.e., PCC engineer). To apply to general cases, an alternative model to calculate the suggested amount of budget using neural networks based on hypothetic data is also proposed. The strengths of this reviewing and budgeting model are identified by applying it to a practical building project. |
URI: | http://140.113.39.130/cdrfb3/record/nctu/#NT910015010 http://hdl.handle.net/11536/69705 |
顯示於類別: | 畢業論文 |