摘要
参数识别是数学模型应用的前提.遗传算法是一种通用的全局优化算法,结构简单.一个实际应用问题能否利用遗传算法解决,关键在于遗传算法的设计和控制参数的选取.本文结合水质模型参数优化的特点,提出采用正交试验设计的方法来考察遗传算法不同控制参数对参数优化性能的影响,结果显示,正交法较好地识别了关键影响因素,并提出可能的最优方案.表明遗传算法能较好地应用于复杂多参数水质模型的参数识别研究.
Parameter identification plays an important role in environmental model application. As a commonly used global optimization method, genetic algorithm (GA) has very simple structure, the key related to whether a practical issue can be solved using GA or not is algorithm design and selection of the control parameters. Based on the feature of parameter optimization of water quality model, orthogonal test method was proposed for reviewing effects of different control parameters of GA on the performance of water quality parameter optimization. The results indicate that orthogonal method could identify key factors, and also provide possible optimized experiment plan. It is concluded that GA can be applied to the research on parameter identification of complicated water quality model.
出处
《环境科学》
EI
CAS
CSCD
北大核心
2005年第3期61-65,共5页
Environmental Science
基金
国家自然科学基金资助项目 (5 0 2 0 90 0 7)
关键词
参数识别
水质模型
遗传算法
全局优化
正交实验
parameter identification
water quality model
genetic algorithm
global optimization
orthogonal experiments
