A dynamic multiple attribute decision making model with learning of fuzzy cognitive maps. (September 2019)

• Record Type: Journal Article
• Authors: Baykasoğlu, Adil; Gölcük, İlker
• Is Part Of: Computers & industrial engineering. Volume 135(2019)
• Page Start: 1063
• Page End: 1076
• Publication Date: 2019-09
• Subjects: Decision analysis -- Dynamic multiple attribute decision making -- Fuzzy cognitive maps -- Jaya algorithm; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2019.06.032
• Languages: English
• ISSNs: 0360-8352

Abstract

Highlights: Learning from historical decision matrices by using fuzzy cognitive maps. Employing Jaya algorithm to train fuzzy cognitive maps. Generating future decision matrices for short-, medium-, and long-term. Utilizing past, present, and future information about alternatives. Ranking of suppliers in a real-life supplier performance evaluation problem. Abstract: This paper concerns with a new multiple attribute decision making (MADM) model to cope with temporal performance of alternatives during different time periods. Dynamic MADM problems has grabbed remarkable attention of decision analysis community in recent years. In parallel with the recent advances in information technologies, firms steadily recognize the importance of data, and business analytics solutions is about to become standard business practice. Majority of the dynamic MADM literature deals with combining past and present data by means of aggregation operators. There is a research gap in developing data-driven methodologies to capture the patterns and trends in the historical data, and provide decision makers with meaningful insights in decision making practices. Analogous with the fact that style of decision making evolving from intuition-based to data-driven, this study proposes a new dynamic MADM model by learning of fuzzy cognitive maps (FCMs) to support decision makers in making informed decisions by considering future performance of alternatives. According to proposed model, Jaya algorithm, a simple and effective metaheuristic optimization method, is used to train FCMs to capture the patterns in historical data. Then, short-, medium-, and long-term future decision making matrices are generated. Finally, past, current and future decision making matrices are taken into consideration, and ranking of alternatives are obtained based on closeness coefficients. The proposed model is realized in a real-life supplier performance evaluation problem.