水库防洪优化调度自适应拟态物理学算法

Adaptive artificial physics optimization of reservoir flood regulation

拟态物理学算法(APO)是近年来一种新颖的基于物理定律启发的智能算法,本文提出了利用智能算法求解水库防洪优化调度模型的通用启发式策略,阐述了拟态物理学算法的基本原理,尝试通过自适应方式合理设置算法的引力参数,同时改善标准算法的寻优进程,提出了自适应拟态物理学算法(AAPO)。将其应用于鲇鱼山水库实时防洪优化调度模型求解,采用最优值的均值和方差衡量算法的精度和稳定性,并对水库最优出流方案分段采用波动率衡量算法应用可操作性,同时统计计算机耗时,结果表明:自适应拟态物理学算法较标准算法搜索精度更高,寻优结果更稳定;且拟态物理学算法较其他算法计算的水库泄流方案的过程线波动变化更小,实时防洪调度应用可操作性较高。

Artificial physics optimization （APO） algorithm is a novel intelligent algorithm that based on physical laws and developed in recent years. This paper describes the basic principle of its calculation and presents a universal heuristic strategy for intelligent algorithms to solve optimal reservoir flood regulating models. In an effort to improve its virtual gravity parameters using the adaptive approach, we have developed an adaptive artificial physics optimization （AAPO） algorithm, expecting to improve its standard searching process. This AAPO algorithm was used to solve the optimal model of real-time flood regulation in operating the Nianyushan reservoir, with the average and variance of its optimal process used for evaluation of its accuracy and stability. In this calculation, the optimal outflow hydrograph of the reservoir was divided into two parts, and each part was calculated separately using a fluctuation ratio indicator to evaluate the feasibility of the new algorithm, along with an indicator of CPU cost. Results show that AAPO is more accurate and more stable than APO and fluctuations in the reservoir outflows optimized by AAPO or APO were weaker than those of other algorithms, indicating its superiority in real- time flood regulation.