基于粒子群算法的非等距GOM(1,1)模型

Non-equidistant GOM(1,1) model based on particle swarm optimization algorithm

等间距灰色GOM(1,1)模型是一种基于反向累加生成的灰色预测模型.为了拓广适用范围,提高GOM(1,1)模型的拟合和预测精度,给出了非等间距灰色GOM(1,1)模型的建模方法,并利用粒子群优化算法对非等间距灰色GOM(1,1)模型的参数进行优化.最后,利用一个仿真实例,表明基于粒子群优化算法的灰色GOM(1,1)模型具有较高的拟合和预测精度.也说明该方法是有效的和可行的,具有重要的理论意义.

The equidistant GOM（1,1） model is a grey forecast model based on the opposite direction accumulated generating. In order to widen the application range and improve the precision of model, the particle swarm optimization algorithm is used to solve the models parameters based on the average relative error function minimization. The basic theory and approach for establishing a GOM（1,1） model for non-equidistant sequence are presented. Finally one application example shows the precision of the equidistance GOM（1,1） with parameter identification based on particle swarm optimization algorithm is higher than the GOM（1,1） model. So this method is feasible, effective and has important theory significance.