摘要
通过对dxd(t1)+ax(1)=u的通解^xk(1)=ce-ak+au的参数a、u、c直接求解,避免了灰微分方程参数辨识时选取合理背景值的问题,构建了适应性更强的不需构造GM(1,1)模型的背景值而直接求解灰微分方程参数的模型,并且在求解这些参数的过程中,应用了在求解非线性问题中具有全局寻优能力的粒子群算法(PSO)。提出了基于粒子群算法优化的电力负荷灰色预测模型PSOGM(1,1,a,u,c),通过在电力负荷实例中的应用并与传统的GM(1,1)预测模型进行了效果比较,验证了基于粒子群算法优化的电力负荷GM(1,1)模型具有很好的预测精度和适用性。
A new method which is used to solve the parameter a, u and c of GM ( 1, 1 ) is discussed. Then particle swarm optimization is adopted to solve the value of a, u and c as this algorithm has the virtue of optimum - seeking and high - quality solution. Therefore, a GM ( 1, 1, a, u, c) based on PSO is finally built. And the result of the traditional GM ( 1 , 1 ) compares with the result of the new GM ( 1, 1 ) model. The practical example indicates that the GM ( 1, 1, a, u, c) based on PSO model has the characteristic of better precision and wider application field.
出处
《四川电力技术》
2009年第1期32-35,共4页
Sichuan Electric Power Technology
关键词
灰色模型
粒子群算法(PSO)
电力负荷预测
背景值
grey model
particle swarm optimization (PSO)
power load forecasting
background value