Efficiency in energy decision support systems using soft computing techniques

Abstract

In the recent years, the advent of globalization has caused a steep rise in energy consumption especially for the most developed countries. At the same time, there is an important need for energy planning especially from the energy distributors’ point of view as, it is vital for setting energy wholesale prices and profit margins. The same argument also holds for the case of energy transition from fossil fuels to more environmentally friendly forms since, it is necessary to predict future energy demands in order to achieve sustainability by balancing energy supply and demand. Even though a lot of research has been produced towards the development of energy models that try to optimize sustainability, we claim that, there is a need for multi-criteria Decision Support Systems, (DSS) that integrate a variety of econometric and computational intelligence methodologies in order to evaluate the impacts of a mix of sometimes conflicting factors such as climatological conditions, global prices, availability etc. In this paper, we propose an integrated DSS that involves Adaptive Neuro Fuzzy Systems (ANFIS), Neural Networks, (NN) and Fuzzy Cognitive Maps, (FCM) along with Econometric Models, (EM) in a hybrid fashion to predict energy consumption and prices. Such a system defines the basic structure in building energy networks that use combined heat and power. Even though the system can be applied for the totality of energy sources, we focus only on the prediction methodologies for the natural gas consumption for the country level. We provide a comparative analysis of the aforementioned methods by obtaining the MSA, RMSE, MAE and MAPE errors, as well as R2 metric. © Springer International Publishing AG 2017.