摘要
为了寻求精度更高的暴雨强度公式,提出嵌入共轭梯度的自适应小生境遗传优化算法,该算法基于绝对均方差最小准则和相对均方差最小准则构建暴雨强度公式参数优化模型.通过标准遗传算法中引入小生境技术提高了种群多样性,同时与共轭梯度算法相结合增强了算法的局部搜索能力,使得遗传算法全局搜索能力强和共轭梯度法局部搜索能力强这2个优势有效结合,并应用于暴雨强度公式参数的优化计算.以北京、广州和郑州的降雨资料为基础进行暴雨强度公式优化研究,并将本文算法与传统方法和标准遗传算法进行比较.结果表明:运用本文算法对暴雨强度公式参数进行优化时,优化结果较好且能满足规范要求.与传统方法和标准遗传算法相比,优化结果具有更高的精度,为暴雨强度公式的推求提供参考依据.
To get a higher accuracy of rainstorm intensity formula, an optimization algorithm (NicheAGA- CGA) was proposed based on niche adaptive genetic algorithm (NicheAGA) and conjugate gradient algorithm (CGA). The rainstorm intensity formula parameters optimization model was constructed by the rule of absolute and relative mean square deviation minimum rule. The niche technology was introduced to improve the variety and the conjugate gradient was combined with the genetic algorithm to improve the local searching. The advantages of global searching in genetic and local searching in conjugate gradient were combined effectively. The rain data of Beijing, Guangzhou and Zhengzhou were used to optimize the rainstorm intensity formula, and NicheAGA was compared with traditional algorithms and standard genetic algorithm. Results show that the optimized results of NicheAGA are better and meet the standard requirements. NicheAGA can get a higher accuracy and provide reference for the study of rainstorm intensity formula.
作者
唐颖
张永祥
王昊
刘宇
TANG Ying;ZHANG Yongxiang;WANG Hao;LIU Yu(College of Architecture and Civil Engineering,Beijing University of Technology,Beijing 100124,China;Key Laboratory of Beijing for Water Quality Science and Water Environment Recovery Engineering,Beijing 100124,China;China Institute of Water Resources and Hydropower Research,Beijing 100038,China;State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin,Beijing 100038,China)
出处
《北京工业大学学报》
CAS
CSCD
北大核心
2019年第3期292-298,共7页
Journal of Beijing University of Technology
基金
国家科技支撑计划资助项目(2011BAC12B00)
关键词
暴雨强度公式
小生境
遗传算法
共轭梯度法
参数优化
rainstorm intensity formula
niche
genetic algorithm
conjugate gradient algorithm
parameter optimization
