2016-01-012024-05-18https://scholars.lib.ntu.edu.tw/handle/123456789/697711摘要：運輸設施生命週期管理(transportation infrastructure life-cycle management)之目的乃是輔助管理者在有限的天然資源及維修管理經費下，最小化生命週期維修成本以及營運成本(包括使用者成本、損壞風險…等)，以期延長設施的使用年限，並且在使用年限內提供更優良的服務績效或設施狀態/狀況。維修門檻值為最常見的設施維護策略之一，其優點為維修方案與施工單位之工作流程相容，其形式直觀且便於執行，易被施工單位理解及接受，但缺點則是門檻值之訂定多依據經驗，缺乏全面性考量。本計劃擬進行四年度的研究計劃，以提出完整的門檻值維修策略運輸設施生命週期管理架構。本計劃於第一及第二年度採用混合動態模式描述門檻值維修策略下之運輸設施生命週期，該模式可同時考量多設施、多維修工法、以及設施相關性，具有極高的創新性，未來若繼續延伸以考量更多元且真實之設施特性與管理策略，可望更貼近運輸設施生命週期管理之實際運作，由於確定性模式之建構與最佳化複雜度較隨機性模式相對為低，納入實際設施特性與不同管理策略可行性較高，因此確定性模式之研究仍具有顯著之理論與應用價值，值得更進一步深入研究，基於此一考量，本計劃在前兩年擬先對確定性模式予以深入瞭解，於第三及第四年考量隨機性模式。 由於混合動態模式最佳化之複雜度甚高，本計劃第一年及第二年度擬先針對確定性模式之問題特性，建立有效率之求解演算法，預計採用之方法包括「混合整數規劃」、「分支界限法結合問題分解策略」、以及「非線性規劃」，同時將針對設施數目、預算大小、門檻值可變與否、不同等級之門檻值、狀態門檻與時間門檻、不同規劃區間長度…等屬性，分析門檻維修方法之效益，以提供運輸設施管理單位在評估選擇策略時的參考，另外也計畫考量其他種類之運輸設施(如：機場跑道、橋梁、軌道、管道、電力設施、通信網路…等)進行模式之延伸，以擴展混合動態模式於運輸設施生命週期管理之應用價值。第三及第四年度計劃擬將運輸設施生命週期中的隨機性納入最佳化模式中，預計由「隨機性混合動態模式」與「隨機規劃」中擇一，亦可能進行兩方法之比較。由於應用混合動態模式於門檻維修策略生命週期管理最佳化為目前最新且最完備之模式架構，其後續相關研究將具有時效性以及創新性，可望有良好的研究成果產出。<br> Abstract: The objective of transportation infrastructure life-cycle management is to support the decision-making of resource allocation for the infrastructure maintenance. Under the constraints of natural resources and maintenance budgets, maintenance optimization extends the life of the infrastructure and provides better services and facility conditions. Threshold-based maintenance is one of the most common strategies because it is compatible with the workflow of the maintenance agencies. However, the determination of the maintenance thresholds is often made based on expert experience and can be improved with more rigorous approaches. I plan to conduct a four-year research and develop a framework for threshod-based transportation life-cycle management. The framework will adopt hybrid dynamic modeling (HDM) to formulate the lifecycles of transportation infrastructure. Adopting HDM is beneficial because it considers threshold-based maintenance, multiple facilities, multiple maintenance actions, and interdependency between facilities simultaneously. The framework will also extend the methodology of HDM to consider more complicated management strategies and deterioration mechanisms, in order to describe the lifecycles of transportation infrastructure in a more realistic way. I plan to study deterministic models in the first two years of the project and consider stochastic models in the final two years. Due to the complexity of the optimization problem, efficient solution algorithms are necessary. Three solution methods will be developed and tested in the first two years: mixed-integer programming, branch-and-bound combining problem decomposition and nonlinear programming. The framework will also be evaluated to understand the performance of the threshold-based strategy under different conditions of number of facilities, budget availability, level of thresholds, planning intervals, etc. The study will strengthen the understanding for the strategy and will be beneficial for the maintenance agencies. Other types of transportation infrastructure might also be considered to further increase the value of the framework. In the final two years, the stochastic components of the life cycles of transportation infrastructure will be considered. Currently, the candidate methods for this task include stochastic hybrid dynamic models and stochastic programming. Because the application of hybrid dynamic modeling on life-cycle management is the latest and the most complete framework so far, the following studies will be timely and innovative and positive outcomes can be expected.生命週期管理維修決策養護門檻混合動態模式隨機模式life-cycle managementhybrid dynamic modelmaintenance decision-makingmaintenance thresholdstochastic models