我國位處於亞洲大陸板塊擠壓帶，地處地震、颱風、洪水等天然災害發生頻繁地區，其對國人生命、財產安全造成嚴重威脅。921集集大地震及918 納莉颱風造成空前重大災難，其中921 地震所造成經濟損失約3000 億元，已引發政府與社會大眾對防災與風險管理之重視。 地震事件之發生為大自然現象，以人類現有之科技尚無法預測或阻止其發生，惟有以事前之天然災害風險管理措施，來減低其可能導致損失及其對個人、企業或政府之衝擊。災害風險的主要管理措施有三種，包含自留(Retention)、減災(Mitigation)及移轉(Transfer)等三種，而該採用何種災害風險管理策略，則可依風險分析之資訊，如災害之危害度、標的物的易損性及風險曝露值等，進行風險管理決策。 本論文初探國內天然災害風險管理環境，提出天然災害風險管理措施之經濟效益分析架構；分別為風險管理決策流程、風險控制效用分析和風險控制財務機率分析。對政府或私人企業而言，有數種建物風險控制措施，如：補強結構、購買保險、發行巨災債券及實施緊急應變計畫等。吾人將定義何謂最適化決策且評估之，並提出經濟效用條件最適化分析，此研究之目的在於建立最適化的決策以決定保險和風險控制計畫，此決策同時考慮各措施之災害防治經濟效益(如:降低災害損失)已知條件平衡。此模型分析之結果可進一步比較各風險控制措施以決定措施優先性，並幫助評估其效益性。 This dissertation presents extended frameworks of strategic analysis for natural disaster risk control. Natural disasters have properties of low frequency and severe loss, especially catastrophic flood and earthquake. For government or a private company, there are several kinds of actions for flood and earthquake risk control, i.e. structure retrofitting, insurance policy purchasing, catastrophe bond issuing, emergency planning and subsidiary reserve (for workers compensation and etc.). The direct and indirect economic losses caused by the simulated disasters can be estimated using the engineering and financial analysis model. Basing on the model, we can calculate how much loss will be ceased or transferred to other entities, if some budgets spent on risk control actions. In addition, from various points of view, I try to define the optimal strategy and evaluate it. There are two bundles of models proposed: deterministic analysis models and stochastic analysis models. Deterministic analysis models include two kinds of concepts: interpretive structural modeling (ISM) and economic utility constrained-maximization. Stochastic models start with EP (Exceeding probability) curve into various fascinating optimal models. The background and purpose of this study are to establish a strategy to determine the risk control plan in which consideration is given to the balance with the economic effects (e.g. decrease in damage cost) due to disaster prevention. ISM developed acts as a tool for top management to understand the variables of natural disaster risk control. Though ISM is developed on the basis of perception of the experts of natural disaster risks, the results are quite generic and helpful for the top management to drive the efforts towards the roots of the problem. Furthermore, in economic utility constrained-maximization and stochastic models, these values were compared between those risk control actions to determine the priority and provide data to help evaluate the profitability of each risk control action. Several aspects of risk and uncertainty are discussed within the context of economic utility constrained-maximization models with a major focus on the importance of risk and uncertainty in research evaluation, how strategy determines insurance and risk control plans, and a real-world example dealing with risk and uncertainty. The core of a default risk-based probabilistic decision model is the development of an "exceeding probability curve" that synthesizes stochastic dominance test results to rank the ordering of risk control and accounts for different decision makers' biases with respect to risk. Some further analyses are described that complement stochastic dominance in significant ways.
