Development of dynamic and real-time fuzzy MCDM models for adaptive decision-making

Abstract

This research introduces a dynamic, real-time and hybrid intelligent fuzzy Multi-Criteria Decision-Making (MCDM) framework for supplier evaluation in the uncertain logistics con-text. The proposed framework is based on fuzzy logic, dynamic entropy weighting, temporal Basic Unit-Interval Monotonic (BUM) aggregation, Dynamic Fuzzy TOPSIS and Adaptive Neuro-Fuzzy Inference System (ANFIS) optimization to enhance the accuracy, robustness and real-time responsiveness of the decision-making process. The framework dynamically changes the importance of each criterion as a result of the logistics conditions; and assigns greater weight to recent operational information through temporal weighting mechanisms. The fuzzy inference modelling is very useful to place the uncertainty and nonlinear decision relationship in control, and the optimization of ANFIS has improved its adaptive learning capability and prediction reliability. The proposed model was implemented and validated in MATLAB and Simulink environments with a dynamic logistics dataset, including the sup-plier evaluation criteria like shipping cost, lead time, reliability, cargo status, fulfillment ef-ficiency, and route risk. The experimental results yielded 97.0% prediction accuracy, a cor-relation coefficient of 0.933 and low prediction error, which indicated good learning capabil-ity and computational stability. In addition, the real-time Simulink implementation proved adaptive decision behavior in changing continuously logistics conditions. The results ob-tained show that the proposed hybrid intelligent framework is an efficient, scalable and reli-able solution for the applications of intelligent supplier selection and real-time logistics deci-sion making.