204 ANALYSIS OF SUSTAINABLE DEVELOPMENT USING FUZZY LOGIC PREDICTION MODELS AND ARTIFICAL NEURAL NETWORKS Daniel-Petru, Ghencea 1 Mihaela, Asandei 2 Miron, Zapciu 3 Abstract: Sustainable development is a priority of policies in countries all over the world, regardless of their level of development; this is a dynamic and complex concept based on indicators with vague and difficult to measure characteristics such as resources, labor, education, infrastructure, the existence of modern equipment to ensure manufacturing performance and flexibility. A model of approach and analysis of sustainable development using these indicators with vague characteristics can be achieved by combining prediction models: artificial neural networks and fuzzy logic. Artificial neural networks are used in the study, as they have the advantage of working with hidden layers, and recursive backpropagation algorithms to predict the size of indicators for a certain period, while fuzzy logic is used for three-dimensional interpretation of interdependencies and trends of indicators. The model provides long-term, flexible management decisions by eliminating bottlenecks and assessing deviations from a target defined so that the final result ensures a fast and flexible solution through fast and durable reconfiguring. Keywords: neural networks, fuzzy logic, sustainable development, management, decision-making, sustainability JEL Classification: A12, C63, D89 1. Introduction Rapid economic development and improved quality of life are achieved in close relation to sustainable development, but require an efficient management of natural and technological resources at all levels: global, regional, national or local. The constant emergence of new challenges and sustainable development indicators requires deciding on the priority of each problem. Because these indicators are characterized by uncertainty, a vague vision on new issues that arise, and mutual influences between indicators, it is preferable that they be analyzed using prediction models that bring forth hidden information that cannot be perceived through a classical analysis. The advantage of digital-linguistic dual analysis is the extensive area of research due to the linguistic variable. The paper addresses such a dual analysis using artificial neural networks (ANN) employed for forecasting and fuzzy logic employed for the three- dimensional interpretation of interdependencies between prediction data. Different authors analyzed the use of artificial neural networks and fuzzy logic to evaluate sustainable development indicators, but the dualism of these analyses has not been addressed. Specific indicators, such as population density, GDP per capita, water, soil and air quality were analyzed to evaluate urban development (Daniela Hîncu, 2011; Marius Pislaru et all, 2011; Lucas Andrianos, 2015). Energy resources, the risk of using pesticides and nitrates, air pollution, income and employment rate were used as indicators to analyze sustainable agriculture in Iran (Moslem Sami et all, 2013). Nenad Stojanović (2011) analyzed the tourism sector in natural reserves (protected areas) in terms of sustainable development by minimizing the impact of tourism on components such as economy, society, culture, and tourist satisfaction. For a qualitative 1 PhD. Eng.-Ec. Student, POLITEHNICA University of Bucharest, daniel.ghencea@blackseasuppliers.ro 2 Assoc. Prof., PhD., „Constantin Brâncoveanu” University of Pitești, mihaela.asandei@yahoo.com 3 Professor at POLITEHNICA University of Bucharest, Corresponding member of the Academy of Romanian Scientists, miron.zapciu@upb.ro