灰色系统模型的优化岭回归算法

文献[1]指出了目前用普通最小二乘法估计灰微分方程参数的方法由于方程组的病态问题很难求解得合理的参数;文献[2]指出了根据初值求解灰色系统模型的时间响应式的方法由于初值的误差使所求得时间响应式产生系统误差.为了克服灰色模型的上述两个缺点,本文设计了一种求解灰色系统模型的优化岭回归算法,计算一个广泛引用的算例演示了这种算法的优越性.

The Optimized Ridge Regression Algorithm for Grey System Models

Literature [1] shows that, by means of the parameter identification of grey differential equation by ordinary least square approach reasonable parameters are hard to obtain due to ill-posed problem of the equation; literature[2] points out that systematic error occurs due to the error of the initial value while solving grey time response formula. A kind of new optimized ridge regression algorithm is designed to obtain optimal grey time response formula while overcoming the two above shortcomings. A widely cited example is computed again to demonstrate the advantage of the algorithm.