离散勒让德多项式序列预报控制原理与算法

DISCRETE LEGENDRE POLYNOMIAL SERIES FORECASTING CONTROL PRINCIPLE AND ALGORITHM

  • 摘要: 本文根据数据拟合曲线的外延特性和勒让德多项式的正交性,提出了离散勒让德多项式序列(Discete Legendre Polynomial Series,简记为DLPS,以下同)预报控制原理,给出了LDPS的一系列性质和实时在线递推预报算法,其计算量远远小于已有各种方法,解决了预报控制的实时性问题。经实验和生产现场验证,这一预报原理与算法具有很好的控制效果。

     

    Abstract: According to the extrapolation property of data fitting curve and the orthogenality of the Legendre polynomial, Discrete Legendre Polynomial Series (DLPS for short) is developed in this paper. A series of properties and an real-time on-line iterative algorithm are proposed. The amount of calculation is much smaller than that of all kinds of original algorithm, and the problem of calculation time in forecasting control is solved. Experiments and applications in production floor show that the forecasting principle and algorithm proposed is powerful.

     

/

返回文章
返回