梯级水库优化调度模型的蚁群系统(ACS)算法求解研究

Solution of ant colony system(ACS)algorithm to optimize reservoir operation

【目的】对梯级水库调度模型的动态、高维、非线性、复杂优化问题进行求解。【方法】在传统蚂蚁系统的基础上,将蚁群系统中的蚁密、蚁量系统的局部更新和蚁周系统的全局更新有机结合,提出了一种求解梯级水库优化调度模型的改进蚁群算法,即蚁群系统(ACS)算法,采用ACS算法对乌江梯级水库进行了优化调度实例研究。【结果】ACS算法兼顾了计算的时间和精度,优化得乌江梯级发电量为96.538亿kW·h,相比利用动态搜索算法求解的乌江梯级发电量95.882亿kW·h略大,但均接近于乌江梯级设计多年平均发电量100.21亿kW·h。【结论】采用ACS算法可快速求解乌江梯级水库优化调度模型,并可得到满意的结果,说明该优化算法是合理、可行的。

This study aimed to solve the dynamic,high-dimensional and nonlinear complex optimization problems of cascade reservoir operation model.【Method】Based on traditional ant system,an improved ant colony algorithm（the ACS algorithm）of solving the optimal operation model for cascade reservoirs was improved by combining the partial innovation of ant density system and ant quantity system with the entire innovation of ant circumference system.【Result】The ACS algorithm took into account of both computing time and accuracy.The cascade energy output of wujiang was 96.538 billion kW.h,which was slightly larger than dynamic search algorithm（95.882 billion kW.h）and close to the averaged capacity of Wujiang cascade reservoirs（100.21 billion kW.h）.【Conclusion】The ACS algorithm was reasonable and feasible to solve cascade reservoir optimal operation model quickly with satisfactory results.