Adjustment of the analytic hierarchy process indicators using AI tools

This study aimed to enhance the analytic hierarchy process (AHP) by integrating artificial intelligence (AI) algorithms for the automatic adjustment of its indicators, thereby improving the method’s accuracy, consistency, and adaptability. A conceptual analysis of both the traditional and AI-oriented approaches was conducted. The research methodology included a systematic literature review, identification of the key limitations of the classical method, and testing of AI capabilities to improve the consistency and precision of weighting coefficients. The findings demonstrate that the integration of AI into AHP significantly reduces the subjectivity of expert evaluations, lowers the need for manual adjustment of pairwise comparison matrices, and enhances the consistency of decision-making. Specifically, optimisation algorithms automatically identify conflicting judgements and correct them without human intervention, thus reducing decision-making time. The use of clustering methods facilitates the automatic grouping of criteria and alternatives based on similar characteristics, thereby reducing the number of required pairwise comparisons. The application of machine learning-based algorithms for predicting weighting coefficients enables the AHP to adapt to dynamic changes in data, enhancing the stability and reproducibility of results. Furthermore, the incorporation of Explainable AI methods improves the transparency of the decision-making process by allowing the influence of each criterion on the final outcome to be clearly explained. The analysis also demonstrated that the application of AI in multi-criteria analysis significantly reduces the cognitive load on experts, minimises the impact of human factors, and increases the accuracy of calculations. However, despite these substantial advantages, the integration of AI into AHP requires careful model configuration, as the effectiveness of such systems depends on the quality of the input data and the explainability of the outcomes. The practical significance of these findings lies in the potential to apply the proposed approaches to optimise decision-making processes in business, public administration, and the technical sciences, thereby contributing to the improved efficiency of analytical systems