臺灣地區平均年降雨量為2500公厘，但是降雨量分配極不均勻，雖然雨量充沛，但是卻無法儲蓄水源。由於近年來，工商業蓬勃發展，各個標的需水量增加，就現有之水源已經不夠使用。然興建水庫是增加水源最有效的方法，但是環保意識的高漲，致使興建水庫來開發水源更是難上加難。在不興建新水庫的前提之下，只有藉由改善水庫的運轉策略來增加水資源之利用效率。 本研究為淡水河流域之翡翠水庫及計畫中坪林水庫兩水庫串聯系統，以目標年119年需水量為供水標的，模擬翡翠水庫及坪林水庫進行聯合運轉，擬定兩水庫運轉規線及系統可再開發水量，再利用遺傳演算法來優選水庫最佳運轉規線。運轉規線形狀通常為一折線，本研究以規線之轉折點編碼進行優化，不同於一般優化規線之編碼方式。其主要為搜尋出一組缺水指數最少的運轉規線，在聯合運轉中得到整體之最佳解，使串聯系統之水資源能發揮充分利用的功效。研究結果顯示可使缺水指數減少，其優化的時間也較一般優化規線方法來的快速。 This study emphasizes on the joint operating rule for multiple reservoir system in series. The system, consisting of Feisui Reservoir and Pinlin Reservoir in Shindian Stream Basin, is employed as a case study. According to the inflow data of 43 years and predicted demand of the year 2030, we build reservoir simulation model, and then apply Genetic Algorithms (GAs) to optimize operating rule curves for reservoir system in series. In the processes of optimizing in GAs, the most important step is coding decision variable as chromosome. The coding method in this study is different from other researches. The most different point is that the number of turning point of operating rule curve is the decision variable. By the way of procedures in GAs, which are selection, crossover and mutation, the optimum operating rule curve of multiple reservoir system in series is searched for. The result shows that optimizing operating rule curve is obtained as the smallest shortage index appears. The feature of the optimizing operating rule curve is unnecessary to be smoothed.
