Skip to Main Content
Governments are interested in energy supply industry and so consider it essential to predict energy demand, as well as the private sector. Solution of the problem depends on availability of a model. This paper proposes a systematically developed model. It is based on a previously performed exogeneity investigation of various quantified variables. A certain nonlinear model among a collection of 100 models with different inputs is chosen as the most appropriate model. Structure of all nonlinear competing models is established according to logical conjunctive and disjunctive relationships between variables. Different combinations of the exogenous variables generate these models. At first, unacceptable models are put away applying coefficient sign significance criterion and error validity. An automated fuzzy decision-making process determines the winner model, which is a hybrid nonlinear model, among the other remaining models.
Date of Conference: 5-8 April 2004