改进离散灰色模型在基坑周围地面沉降预测中的应用  

Improved discrete grey prediction model of ground settlement around excavation

作　　者：张永磊[1] 何秀凤[1] 

机构地区：[1]河海大学卫星及空间信息研究所,南京200098

出　　处：《工程勘察》2013年第2期63-65,70,共4页Geotechnical Investigation & Surveying

基　　金：国家自然科学基金(40974001);江苏省研究生科研创新计划项目(CXZZ11-0451)

摘　　要：离散灰色模型解决了常规灰色模型从离散形式到连续形式直接跳跃的问题,该模型估计模型参数时采用最小二乘法,而在模型检验时常采用平均相对误差指标,因此导致建模与模型检验的脱节。鉴此,本文基于平均相对误差最小准则,利用果蝇算法的快速全局优化能力对模型参数进行优化,从而提高离散灰色模型的建模与预测精度。将本文提出的果蝇算法优化离散灰色模型应用到工程实例,结果表明该模型具有较好的有效性和可行性。Discrete grey model solves the jump problem of conventional grey model when it takes the transformation from discrete to continuous. The model uses least square approach to estimate the model parameters, but generally takes the average relative error between restored value and real value as the criterion to evaluate the simulation precision, so the estimation method is inconsistent with the evaluation method in the model. This paper uses the fruit fly optimization algorithm to get the optimal parameters by minimizing the average relative error between restored value of the model and the real value. It improves both simulation and prediction precision. The result acquired from a real example shows that the improved model is effective and practicable.

关 键 词：离散灰色模型 果蝇优化算法 平均相对误差 地面沉降 

分 类 号：O141.4[理学—数学]