環保設施選址線上決策支援系統

Abstract

環保設施選址由於需要評估許多選址準則與因子，以及處理大量的地理資訊，導致選址過程複雜度頗高，因而有必要發展有效率的工具輔助選址決策分析。本研究因而發展一個線上決策支援系統輔助環保設施選址及進行相關決策分析工作，此系統主要包括選址模式求解模組、選址平行化運算模組與選址空間分析模組等三大部份。選址模式求解模組主要包括網格式及向量式選址模組，由於現有向量式選址優選模式針對實際問題的求解效率不佳，且必須使用套裝軟體求解，本研究因而發展一個向量式選址的專屬演算法及求解模組。選址模式求解模組雖已大大改善求解效率，但對於網格式案例而言，隨案例區域增大，求解時間會顯著增長，本研究因而發展一個平行運算環境，將大案例區分為數個子區域分別求解，進一步改善選址模式求解效率。本研究亦發展環保設施選址空間分析模組，用以進行空間性因子分析及展示所得的成果，增進在選址決策支援上的效率與實用性。並藉由本研究所發展的環保設施選址決策支援系統，有效輔助進行相關決策分析及場址評估，並進行案例測試，以示範本系統如何提昇選址效益與決策支援品質。依案例實測結果證實優選分析模組可效率提供較佳的優選方案，平行求解模組可有效改善求解效率，約在3倍到7倍之間，而所發展的向量式掩埋場選址電腦程式更可增進至少29倍的求解效率，且可獲得適宜性更好的候選場址。本研究亦分別示範應用各系統進行掩埋場、水質監測站與空氣品質監測站的選址案例分析，展示各系統輔助選址決策的實用性，讓使用者有效率且較深入的分析不同選址方案，提昇選址決策分析的效率與品質。

Abstract Environmental facility siting (EFS) must consider various criteria and factors for evaluating the suitability of a candidate site. EFS also needs to process a massive amount of spatial information before a proper EFS decision can be made. This study was thus initiated to develop a decision support system (DSS) to facilitate EFS analyses. The system includes three core modules: optimization analysis, spatial analysis, and parallel computing modules. This study developed four sub-systems for raster-based landfill siting, vector-based landfill siting, water monitoring station siting, and air monitoring station siting. The raster-based landfill siting sub-system includes three major modules for landfill siting optimization, spatial analysis, and parallel computing. For establishing the vector-based landfill siting sub-system, a specific algorithm and data format were developed to improve its computational efficiency. The vector-based landfill siting sub-system includes an optimization module using the specific algorithm and a spatial analysis module. The water quality monitoring station siting sub-system consists of an optimization module to determine a monitoring network, a spatial analysis module, and a siting summary and comparison module. The air monitoring station siting sub-system aims to assist users in designing an appropriate air monitoring network and comprises an optimization module, a spatial analysis module, and a siting summary and comparison module. The results of parallel computing show that the parallelized algorithm and computing environment can increase the speed by about three to seven times while compared to the original algorithm implemented on a single computer. The vector-based landfill siting program can significantly improve computational efficiency, at least 29 times faster than the time used by a previous model. The sub-system is thus able to find candidate sites with good suitability. The system and all sub-systems have been applied to several illustrative cases to demonstrate their applicability in improving the quality of decisions for EFS problems.