摘要
阐述了引入时间距离权的传统GM(1,1)模型的建模过程以及灰色模型等级的判断方法,深入分析了白化背景参数λ取值对建模精度的影响;同时针对传统λ取值的缺陷,提出使用基于实数编码的改进遗传算法(IRCGA)对其进行优化处理,并用多个工程实例分析验证了优化后的GM(1,1)模型相对传统的灰色模型及其优化模型拟合效果更好,更加贴近真实数据序列。
The modeling process and methods for determining the model levels of grey system GM( 1,1 ) model which considered the weight of time -distance are elaborated in this article, and deeply analyzes the values of whitening background parameter' s influence for the modeling accuracy; meanwhile, according to the defects of traditional value of λ(whitening background parameter), a method is put forward by using improved genetic algorithm based on real -coding to optimize it, and the fitting effect of optimized GM( 1,1 ) model is proved more better and closer to the real data series compared with the classical GM( 1,1 ) and traditionally optimized model by several project cases.
出处
《测绘与空间地理信息》
2013年第12期42-45,共4页
Geomatics & Spatial Information Technology
基金
铁道部科技研究开发计划项目:高速铁路地质路基关键技术研究--沪宁城际铁路沉降控制效果的机理分析与对策研究项目(2012G009-C)
铁道部科技发展计划项目:区域地面沉降对(京沪)高速铁路工程的影响及对策研究项目(2008G031-5)
中央高校基本科研业务费专项资金:空间信息获取与建模及其在高速铁路中的应用项目(SWJTU10ZT02)资助
关键词
GM(1
1)模型
白化背景参数
建模精度
实数编码改进遗传算法
变形预测
GM (1,1) model
whitening background parameter
modeling accuracy
improved genetic algorithm based on real -cod- ing
deformation forecasting