A Machine Learning-Based Multi-Criteria Decision-Making Approach Utilizing D-Numbers for Water-Energy-Food Nexus Assessment

Abstract

Interdependency between the infrastructure of the water and the energy represents the core challenge of the management of resources. Effective decision-making for Water-Energy-Food (WEN) scenarios requires robust decision-making tools. The traditional Multi-Criteria Decision-Making (MCDM) tools are undermined by uncertainty due to the reliance of the latter upon complete and perfect knowledge and conditions that are rarely attained under WEN scenarios. The traditional models are simplifications of the interdependency of the water and the energy systems and are responsible for generating nonoptimal decision-making strategies. While the fuzzy and intuitionistic techniques attempt to cope with the situation of uncertainty, they are inefficient in addressing conflicting and uncertain information that hinders the practical implementation of these techniques. Besides that, the non-existence of a platform that unites MCDM with the integrated uncertainty management increases decision-making complications. In bridging the above gaps, the current study proposes a new framework that integrates D-number-based Multi-Criteria Analysis with Dempster-Shafer Theory (DST) for WEN decision-making. The integration of DST rigorously enhances the ability of DST for processing complete, uncertain, and conflicting information for WEN decision-making. The study also compares the performance of the Random Forest and Optimized Artificial Neural Network models. The new framework enhances decision robustness with the use of D-TOPSIS, D-VIKOR, and D-Entropy techniques with the assistance of D-number theory for more efficient handling of uncertainty and providing decisionmakers with a credible basis for the management of resources.