Application and adaptation of Genetic Algorithm in optimal Eco-friendly reservoir operation

Abstract

Optimal reservoir operation is a complex problem that involves multiple objectives, multiple constraints as well as considerable risk and uncertainty. Eco-friendly reservoir operation makes it more complicated by taking into account a conflicting objective or highly nonlinear constraint related to ecosystem requirement. The study developed a model to optimize reservoir operation in an Eco-friendly manner by using Genetic Algorithm(GA) and applied it to two cascade reservoirs of Yalongjiang River in the Southwest of China. In order to improve its performance, GA was adapted in transferring objective function and operating mutation dynamically. In addition, a time-nested model was proposed to optimize monthly-based data to daily one, thereby avoiding too much state variables being involved when reservoir require daily operation policy. It is shown that the adapted GA can certainly fulfill the goal of eco-friendly reservoir operation and it was enhanced in search accuracy and global searching ability with objective function transfer and dynamic mutation operator. Moreover, the time-nested model was greatly help to build a daily-based optimization model which can cut computing times dramatically and improve the GA efficiency.