基于多目标遗传算法的渠系配水优化模型

针对渠系配水传统人工调度难以统筹优化的问题,为使优化模型更加贴近实际,体现渠系配水优化的时空特点,充分考虑配水持续时间的不确定性,建立以下级渠道配水净流量、配水开始时间和结束时间作为决策变量,以保证上级渠道水流平稳和下级渠道渗漏最小为模型目标的多目标多变量渠系优化配水模型,采用向量评估遗传算法进行求解,将第一目标即上级渠道水流平稳作为优先满足目标,第二目标为次要满足目标。结果表明:在保障实际运行对于渠道流量要求和水量要求的前提下,优化结果与该时段甘州区配水计划相比,配水时间由轮期的25d缩短为15d,干渠平均配水流量由1.298m~3/s提高到1.414m~3/s,渠系水利用系数由0.651提高到0.706,田间配水量提高了14.25%。本研究模型能够实现集中、高效配水,且具有普遍适用性,可为渠系优化配水决策提供理论和技术支持。

Multi-objective genetic algorithm optimization model for canal scheduling

Aiming at the complication of optimizing water delivery scheduling,and to reflect its spatial and temporal characteristics,a multi-variable and multi-objective optimization model for canal scheduling is proposed,which considers the time uncertainty.Decision variables are flow rates,start time and end time of the distributary canals.The objectives are to minimize the standard deviation of flow rates of superior canal and the seepage of the distributary canals.The model is solved by multi-objective genetic algorithm with the first objective taking the top priority.The results show that:On the premise of satisfying the need of actual operation,compared with planned conditions at the same period,the rotation decreases from 25 to 15 d,the average flow rate of superior canal increases from 1.298 to 1.414 m³/s,the water use efficiency grows from 0.651 to 0.706 and the amount of water into farmland rises by 14.25%.In conclusion,the model can realize concentrated and high-efficient water distribution,and it can be applied to many contexts in order to provide theoretical and technical support.