三角函数选择算子的遗传算法在梯级水库优化调度中的应用

针对遗传算法求解水库优化调度时传统选择算子无法克服适应度非负问题，提出了一种三角函数选择算子遗传算法的梯级水库优化调度求解方法，并通过实例对正弦函数轮盘赌选择与单纯轮盘赌选择的自适应遗传算法做了比较．结果表明，由于基于正弦函数的选择算子克服了轮盘赌对适应值非负的要求，非常适用于求解水库优化调度，且能够很好地保持种群多样性，因此在进化后期能够找到更加优良的解．在实例的选择上，选用“以水定电”的三峡梯级调度模型，并采用由大量实测资料建立的三峡出库流量和葛洲坝入库流量变系数线性回归方程来处理水库间流量传播问题．本文提出的梯级水库优化调度求解方法，有效地改善了遗传算法的收敛性能，而且由于考虑了流量演进，提高了模型精度，具有较高的实用价值．

Application of Genetic Algorithm with Trigonometric Selective Operators in Optimization of Cascade Reservoirs Operation

For traditional selection operator of genetic algorithm(GA)optimizing reservoirs cannot meet the require-ment of non-negative fitness values,a new method for optimizing cascade reservoirs based on GA with trigonometric selective operators is proposed.A comparison is made between sine-roulette selection and simple roulette selection through an example of adaptive genetic algorithm.The results show that,because the sine-roulette selection can meet the requirement of non-negative fitness values and the diversity of population can also be well maintained,the new one can find more good solution at the later stage of evolution and be well suited for solving reservoir optimal schedul-ing problems.This paper chooses the "power determined by flow" scheduling model of the Three Gorges cascade,and uses the measured data obtained by linear regress equation of the outflow of the Three Gorges and the inflow of Gezhouba to deal with the close hydraulic coupling between reservoirs.The established method effectively improves the convergence performance of GA as well as the accuracy of the model for its consideration of temporal-spatial variation of flow.