Towards Integration of Multi Objective Optimization and Multi Criteria Decision Making

Multi Objective Optimization (MOO) is the procedure of arriving at the best solution, where the objectives are in competition with one another. The idea is to classify such solutions that optimize the value of objectives without compromising any of the objectives’ performance. Multi-Criteria Decision-Making (MCDM) approaches provide systematic approaches to estimate and rank the options based on the criteria or attributes. This research paper presents a critical review of MOO and MCDM approaches applied in the current literature. Out of the 41 articles that were reviewed, it was identified that 70.7% of the articles that were being used were metaheuristic optimization methods. The first, and the most frequently applied, approach identified in 41.5% of the articles was NSGA-II, which is an evolutionary algorithm. As for MCDM methods, 48.8% of the articles used subjective weights, and the remaining ones were Analytic Hierarchy Process (AHP) and Best-Worst Method (BWM). Besides, the ranking method that was frequently employed was TOPSIS, accounting for approximately 43.9% of the articles. The use of MOO and MCDM was mainly performed sequentially (post-hoc) in 78.1% of the analyzed articles. In addition, uncertainty was accounted for by 80.5% of the MOO models through Monte Carlo Simulation and triangular fuzzy numbers. Conversely, the aspect of uncertainty was addressed in only 29.3% of the MCDM models, where the most common methods were the fuzzy set approach and probability distributions. The ability of the model to be less delicate to the changes in the input parameters was tested in 39% of the analyzed papers through sensitivity analysis.