- 學位論文
火警初期派遣決策支援系統反饋機制研究-以新北市為例
The Refinement Mechanism of Preliminary Dispatch Fire-alarm Decision Support System for Fire Department of New Taipei City
摘要
臺灣地狹人稠、發展快速且住宅密度高，其中新北市幅員遼闊且地理類型多樣，又為一工、商、住綜合型的都市，在複雜都市結構、人口密集的情況下，使得新北市的火警案件發生比例高居全國之冠。火警案件發生時，119救災救護指揮中心的執勤員面對必須即時決策的情況，其必須在最短的時間內作出最適當的判斷，決定需派遣的資源以因應該案件發生之情形。正確性的決策，可以達到安全、效率、適當且預期的搶救效果，反之，可能造成災情擴大、且需要耗費更龐大的資源及時間，使財產損失及人命傷亡的可能性增高。 有鑒於此，新北市政府消防局於民國100年積極推動火警救災派遣車組輔助系統規劃與建置案，目的在於運用資訊科技建置先進且符合現況運用的火警救災派遣決策輔助功能。 本研究係以新北市政府消防局為研究對象，進行「火警初期派遣決策支援系統反饋機制」之研究與建置作業，根據會影響火警的相關因素，設計出一模型，並且設計其模型運作程序，再利用過去發生的歷史案件，藉由研究出適當的統計量方法以計算出建議派遣車組總量，以使派遣建議能夠與時俱進，最後，將決策樹模型以及模型的反饋機制作為系統核心，設計其實際運作的演算邏輯並加以實作出火警初期派遣決策支援模組。 透過本研究設計的模型，能夠分析找出派遣資源的規則與趨勢，活化系統提供的派遣車組建議，使決策支援機制能夠因應環境的變遷，提供建議派遣車組相關資訊予派遣人員參考。透過派遣工作的標準化與系統化，有效改善超派、誤派或是必須改派的狀況，促使火警救災派遣相關資源能夠更加的精確與快速調度及有效運用，提升救災之工作效率。
並列摘要
Taiwan is cramped and crowded. And Taiwan’s development is so fast and residence is high-density. Particularly, New Taipei City is a big city with a vast territory, large population and geographical types of diversity, and its fire caseload for most of Taiwan. When the fire event happened, the dispatcher of the Fire Department 119 Dispatch Center must make a right decision in short time to decide the appropriate resources for fire event. A right decision can be safe, efficiency, appropriate rescue effect. Conversely may be the scale of the fire event make larger and need to more cost, time and resources to rescue and may be make more property damaged, people injured and dead. Therefore, New Taipei City created a project to practice an assistant system on fire-alarm dispatch in 2011. The objective of the project is use of information technology, to set appropriate fire-alarm dispatch modules to the new needs of the fire-alarm dispatch and relief. This research is about the research of the refinement mechanism of preliminary dispatch for fire-alarm decision support system for Fire Department of New Taipei City. We designed a model based on the factor of influence rescue and use the history cases to calculate the total dispatch engine group. It can make the system catch up the trends. We use decision tree model and refinement mechanisms to design the algorithm of the system, and implement the decision support module of the preliminary fire dispatch. The model can analyze, find the rules and trend of the dispatched resources. It also can make the suggestions of quantum of the dispatched resources to activate the system in the future and assist 119 dispatchers to dispatch fire engine groups of various fire units, to make more precise and fast scheduling and effective use.