目前台灣的橋樑管理系統多依橋樑狀態計算指標排序作為維修橋樑評選的基準,較缺乏系統性效益的衡量,且在有限的經濟資源下,僅以排序評選橋樑將無法使資源得到最佳配置而造成浪費。本研究提出一系統性橋樑維修評選架構,可作為各橋樑管理單位評選維修橋樑、編列年度預算、乃至採用續進式決策之參考依據。此架構首先利用劣化預測評估橋樑未來狀態,同時藉以計算維修橋樑後,決策者與用路人可得的效益。最後在系統總效益最大的前提下,針對一般性橋樑的維修,考慮橋樑對間的相關性,並利用數學規劃的方法,構建一組合最佳化評選模式,再利用傳統鄰近搜尋及一新式啟發式演算法求解。為說明此架構之應用,本研究進行一範例測試,發現初步效果良好。
Most of bridge management systems in Taiwan choose the bridges for maintenance by ranking an index counted from its condition state. Because the index may not completely represent the system benefit, the resources may be inefficiently used. In this research, we develop a systematic framework for evaluating bridge maintenance. The framework is divided into four stages. At first, we apply the Markov Chain theory to forecast the future condition of a bridge. According to the bridge's condition, we calculate both the agency and the user benefits that can be obtained from maintaining the bridge. Thereafter, we develop an optimization model that determines the best bridges to be maintained with the objective of maximizing the system benefit, considering the budget constraint, the correlation constraint between every bridge pair, and other related constraints. A local search algorithm and a new heuristic algorithm are then employed to solve the model. Finally we show an example to demonstrate our framework. The preliminary test results show that the framework could be useful.