系統流程轉換成動力流圖之研究－以水資源分析為例

Abstract

傳統上對水資源問題，常會應用系統圖與流程圖做為分析輔助，而目前系統動力學亦逐漸被應用於分析水資源問題，其中之系統動力流圖可更完整的表達問題的內部流程與結構，輔助傳統流程分析方法之不足。本研究目的即在提出一整合系統圖與流程圖成為系統動力流圖的新方法，並以五個水資源案例來驗証此方法之有效性，及透過各種情境模擬來進一步分析系統的反應。 研究結果顯示，系統圖與流程圖各自只表達解題所需的部份訊息，因此解題時仍需整合兩者的資訊，但是這種整合資訊經常隱藏於分析者之中，造成後續對於問題瞭解分析的困難。而系統動力學中之系統動力流圖，不但可表達系統架構及演算流程，而且可進一步將系統圖與流程圖所無法表達的各變數間的影響關係展現出來。本研究所提出的整合流程主要包含四大步驟：(1)資訊傳遞之連結；(2)實體流與資訊流之釐清；(3)系統圖與流程圖之整併；(4)動力流圖之修正。而以水庫水量調配為主題的各種情境模擬結果，亦驗證了因果回饋圖之正負回饋迴路對系統反應的解釋與預測，本研究更進一步証明若有邏輯判斷變數存在，將使回饋環的結構隨著時間推演變化。本研究提出之整合流程可連貫傳統分析方式與系統動力學的分析邏輯，幫助一般人掌握系統動力學的基本理念，而動態回饋結構的發現將有助於分析者更深入的觀察及解釋系統的反應。

The system diagram and flowchart are conventional ways to assist in analyzing water resources problems. On the other hand, system dynamics has got more attention in water resources analysis. The stock-flow diagram is the most important concept in system dynamics. To facilitate understanding the merits of system dynamics, this research proposes a novel procedure to integrate the system diagram and flowchart into a stock-flow diagram of system dynamics. Five hypothetical case studies on reservoir operations were performed to demonstrate the effectiveness of integrating procedures and scenario studies to analyze system response. This research shows that only the system diagram and flowchart cannot present all the information. To solving a problem, the hidden relations in the two diagrams must be linked. However, the hidden information only existes in the mind, and this causes a difficulty of understanding how the problem is solved. On the other hand, the stock-flow diagram of system dynamics can clearly describe all the information required for solving a problem including the inter-relationship between system variables. The proposed novel integration procedure involved four steps: connecting the variables, identifying the material and information flow, disassembling the system diagram and flowchart and reassembling that into a stock-flow diagram, and then modifying the stock-flow diagram. Scenarios studies focusing on reservoir operation have demonstrated the integration procedure and verified the effectiveness of predicting system response by the causal feedback diagram. This study also reveals that the structure of the causal feedback diagram may vary depending on the evaluation of logical conditions. The proposed novel integrating procedure bridges the gap between conventional system analysis and system dynamics and facilitated understanding the merits of system dynamics. The concept of dynamic structure of the causal feedback diagram also helps researchers to understand system responses in the deep.