Model parameter estimation
Optimization principle
Accurate estimation of the parameters of the logistic model is
essential for successful modelling and forecasting. In order to estimate
the parameters Dmax, tmax and a, we fitted logistic model to
actual natural gas consumption of China during the period
2001–2010 through minimization of Sum of the Square of the
Errors (SSE) between the actual and forecasted data. For this, we
use SSE technique, as an objective function to estimate the
unknown parameters. The mathematical expression of this objective
function is described below as [32]:
Minimize objective function ðSSEÞ ¼ Min xðDF;t; pÞ
¼ Xn
i¼1
DAi fðDF;t; pÞi
2 ð3Þ
where DA is the actual natural gas consumption, DF, t are the
dependent and independent variables and p is number of parameters
to be estimated by the logistic function DFðtÞ.
The minimum value of SSE leads to optimal value of the parameters
of the logistic model. The objective function described in Eq.
(3) cannot be solved in the ordinary way, because of the nonlinear
characteristics of logistic model. Therefore, numerical or iterative
method based on the least square technique is considered appropriate
to solve such types of objective functions.