دانشگاه آزاد اسلامی واحد تهران جنوب

صفحه اصلی / دومین همایش دوسالانه پژوهش های نوپدید در حسابداری،مالی و مدیریت

Optimal Real-Time Decision Making with Predictive Artificial Intelligence and Fuzzy Logic in Big Data Environments

نویسندگان :

Felor Beikzadeh Abbasi¹

1- Technical University of Crete

کلمات کلیدی :

Optimal real-time Decision-Making،Predictive AI،Random Forest،Fuzzy Logic

چکیده :

Decision-making is a critical and inseparable component of modern business enterprises, where achieving optimal real-time decisions requires rigorous prediction and evaluation of uncertain outcomes in big data environments. Artificial intelligence techniques, particularly machine learning algorithms, have emerged as powerful tools to support and enhance these processes. This study proposes an enhanced Random Forest algorithm incorporating dynamic feature selection at the node level of each tree. A dedicated preprocessing pipeline prepares big data, introducing purposeful diversity among trees and transforming the ensemble into a more adaptive dynamic forest, thereby overcoming limitations of traditional static approaches in high-dimensional and volatile settings. To integrate predictive outputs with strategic considerations, the framework combines Dynamic Random Forest results (upward, stable, or downward trends) with risk management level (high, medium, low) and resource type (strategic or tactical) within a fuzzy multi-criteria decision-making (FMCDM) system. The fuzzy approach effectively manages vagueness, subjectivity, and linguistic uncertainty inherent in business decisions by using fuzzy sets, linguistic variables, and Mamdani-type inference. This hybrid methodology enables optimal real-time decision making—recommending to continue, cancel, or modify the current trend—in high-stakes applications such as bid response projects, investment evaluation, and risk-adjusted strategic planning, providing a robust, interpretable solution for adaptive corporate decision support in complex and uncertain environments.