運用蒙地卡羅模型進行海上搜救之分析-以台灣東部海域為例

The Analysis of Maritime Search and Rescue by Monte Carlo Model - Territory of Taiwan Eastern Seaboard

匡一忠(Khwang, I-Jung)

指導教授 ： 邱光輝

國立臺北大學/商學院/企業管理學系/碩士(2013年)

摘要

台灣附近海域之高密度海上交通與海洋環境成為中度風險環境，海難事故年均超過300件，使得海上救難的效率益形重要。海難發生時的天候通常不佳，因而救難船艦到達現場的遲到時間成為影響海上救難的最大因素之一。本研究之目的係針對海上救難船艦建立搜索模擬模型，分析搜救單位遲到時間對救難搜索效能之影響；海域則以台灣東部屬於黑潮體系的外海為研究範圍。以蒙地卡羅方法 (Monte Carlo method) 為建模主體，相關子模型包括：「落海目標之漂浮位移」、「搜索」、「搜救船艦偵測裁判」、「統計」等四項。其中，海流的流向以機率分配模擬環境行為。研究主要限制為搜救之船艦數量以單船為主，在東岸海域進行72小時海上救難之模擬。
本研究藉常態機率分布方式加以模擬。在分析中，以海上搜救模型 (MRSM) 求出11個設定遲到時間之後，進一步運用回歸方式求取線性迴歸模型，以具趨勢預判之能力，提供海上搜索與救助 (SAR) 決策人員之計畫執行。模擬結果中，若以最壞狀況而言，發現若將遲到時間控制在6小時之內落海人員存活率較高。

關鍵字

蒙地卡羅模擬法 ； 海上搜索 ； 偵測率 ； 遲到時間 ； 模型 ； 機率分配

並列摘要

The high density of maritime traffic with the marine conditions has made the sea area around Taiwan a moderate risk environment. More than the average of 300 shipwrecks annually has highlighted the issue of sea search and rescue. As shipwreck occurred, the live or death of survivals at sea are fully dependent on the efficient rescuers. However, the poor weather has been one of the major factors causing time delay of the rescue platform in execution of Search And Rescue (SAR) task. The purpose of this study is to model the searching vessel for its effectiveness on the SR task under the time of delay by various weather conditions. The scenario aims at Kuroshio area located on the eastern of Taiwan. With Monte Carlo method as fundamental, there are sub-models as the pillars include: “Floating target’s displacement”, “Search”, “Search and rescue ship detection referee” and “Statistics”. The behavior of current heading is represented by the normal probability distribution. The main limitation of this study is to simulate single ship’s SAR capability and with 72 hour time limit.
The research conducted the simulation under normal probability distribution. The simulation analyzed with maritime search and rescue (MRMS) model to compute 11 sets of delay time, and further utilized the linear regression model to train the model to have predictive capability. The capability can support to plan and operate on searching and rescue (SAR) decision making. The simulation result showed that the persons will get the higher survival rate if the delay time can be controlled within 6 hours.

並列關鍵字

Monte Carlo ； Maritime search ； Detection rate ； Delay time ； model ； Probability distribution

參考文獻

陳彥宏(2002)。臺灣海難搜救體系之分析與檢討。運輸計畫季刊，31(3)，635-662。

莊士賢（2008）。時空合域連續小波轉換應用於波流場影像分析。行政院國家科學委員會補助專題研究計畫。

Abi-Zeid, I., & Frost, J. (2005). SARplan: A decision support system for canadian search and rescue operations. European Journal of Operational Research, 162(3), 630-653.

Gentle, J. E. (2008). Random number generation and Monte Carlo methods (2nd ed.). New York, NY: Springer-Verlag.

International Maritime Organization [IMO] (1993). Merchant Ship Search and Rescue Manual. London: IMO.