Algorithm Design and Complexity Analysis of the Adaptive Trust- Aware Multi-Criteria Decision Method for Artificial Intelligence Systems

Abstract

The increasing complexity of artificial intelligence (AI) systems requires decision- making methodologies that are not only adaptive but also capable of incorporating dynamic trust into the decision process. While the previously proposed Adaptive Trust-Aware Multi-Criteria Decision Method (ATMCDM) established a theoretical and mathematical foundation for trust- aware intelligent decision making, its practical effectiveness has not yet been comprehensively validated. This study presents an extensive experimental evaluation of the ATMCDM framework through comparative performance analysis, sensitivity analysis, and multiple case studies. The proposed validation framework compares ATMCDM with widely used multi-criteria decision- making methods, including AHP, TOPSIS, VIKOR, and ELECTRE, under dynamic and uncertain decision environments. Several quantitative evaluation metrics are employed to assess decision accuracy, adaptability, trust calibration, computational efficiency, and robustness. In addition, an ablation study is conducted to investigate the contribution of the trust adaptation mechanism to the overall decision quality. Experimental results demonstrate that ATMCDM consistently outperforms conventional MCDM approaches in environments where decision criteria, contextual information, and trust levels change over time. The adaptive weighting mechanism improves decision stability and robustness while maintaining computational feasibility for real-time applications. The findings confirm that trust-aware adaptation represents a significant enhancement to traditional decision- making models and provides a practical foundation for next-generation intelligent decision support systems. This study bridges the gap between theoretical modeling and practical implementation, contributing empirical evidence for the effectiveness of adaptive trust-aware decision methodologies in artificial intelligence.