Quantitative modelling of electricity consumption using computational intelligence aided design

High electricity consumption is of concern to the world for a variety of reasons, including its social-economic-environmental coupled impacts on well-being of individuals, social life and the federal energy policies. This paper proposes a quantitative model to examine the long-term relationship between annual electricity consumption and its major macroeconomic variables, including gross domestic product, electricity price, efficiency, economic structure, and carbon dioxide emission, using computational intelligence aided design (CLAD). It develops a firefly algorithm with variable population (FAVP) to obtain the parameters of the electricity consumption model through optimising two proposed trend indices: moving mean of the average precision (mmAP) and moving mean of standard derivation (mmSTD). The model is validated with empirical electricity consumption data in China between 1980 and 2012, based on which the error of approximations between 1980 and 2009 is +/- 15% and the error of predictions between 2010 and 2012 is [-8%, -5%]. The main contributions of this research are to develop: (1) a novel quantitative model that can accurately predict the social, economic and environmental coupled impacts on the annual electricity demands; (2) the conceptual CIAD framework; (3) FAVP algorithm; and (4) two new trend indices of mmAP and mmSTD. The findings of this research can assist the decision makers in resolving the conflict between energy consumption growth and carbon emission reduction without dooming the economic prosperity in the long run. (C) 2014 Elsevier Ltd. All rights reserved.