A multiple-criteria model suggestion for Turkey energy planning

The energy which provides a basis for the social development has a vital role for the survival and development of humankind as an environmental factor. With the outrageous price increase at the fundamental energy source of the world and rapid development of Turkish economy, energy consumption of Turkey has become a major problem. In order to overcome such situation, the necessity of creating decision-making processes that will provide us to make correct predictions emerges. Correct predictions can provide great benefits for institutions on issues such as planning the commitments, productions and maintenance. For that reason, institutions need to forecast the demands related to future as closely to truth. Because deciding on future includes many uncertainties, implementing these decisions is fairly difficult. Artificial Neural Networks (ANN) as a regular data management method is very popular on energy demand forecasting. In this study, Adaptive Network Based Fuzzy Inference System (ANFIS) was suggested as an alternative approach. Recognition of the required amount of energy related to future also necessitates carrying out specific studies upon how Turkey will use the energy sources. For that reason, views of decision makers related to the relative importance of selection criteria was determined using a flurry Analytical Hierarchy Process (AHP) based upon type-2 fuzzy clusters, and multiple-criteria decision-making methodology based upon type-2 interval TOPSIS method was used in order to list the best energy alternatives. In this study, type-2 fuzzy clusters were used due to providing more independence on representing the uncertainty and fuzziness of the real world implementations. In order to forecast the energy demand of Turkey, ANFIS method combining the methods of Artificial Neural Network (ANN) and Fuzzy Logic (FL) were analyzed as the forecasting instruments. The suggested model forecast the energy demand using ANFIS application data sets processed with Principal Component Analysis (PCA) and obtained from Ministry of Energy and Natural Resources. ANFIS method was proved to reveal an efficient performance upon the accuracy of the obtained results. For that reason, ANFIS model has become an acceptable model as an efficient alternative method in forecasting the energy. Finally, the model was used in long-term energy forecasts, and the obtained results were interpreted. Moreover, in order to determine the most appropriate energy alternative for Turkey, type-2 interval fuzzy multi-criteria decision-making methodology was suggested, and efficient results were obtained. When test-purpose forecasts made for 2012-2035 years and the mistakes made for the projections were analyzed, ANFIS model was noticed to reveal better results. This result revealed the learning ability of ANFIS obtaining R=0.99998 value in ANFIS model. Besides, it was forecasted that energy needs of Turkey for 2020 will be (157.090, 65 PEKT) and for 2034 will be (266.689, 59 PEKT). When the results obtained through co-application of type-2 interval fuzzy AHP method and type-2 interval TOPSIS method employed for overcoming fuzzy multiple-criteria decision-making problems related to energy planning and investment; it was determined that the best energy alternative was wind energy. The rest alternatives were solar energy, hydraulic energy, geothermal energy, bioenergy, natural gas, petrol, coal-lignite, nuclear energy and hydrogen energy, respectively.