1. A new classification of multi-objective optimization methods is proposed, distinguishing the pruning method from a priori, a posteriori, and interactive methods.
2. The pruning method is defined as an alternative approach that produces a focused subset of Pareto optimal solutions that are easily comprehensible to the decision maker.
3. A comparative analysis of the pruning method and other MOO approaches allows for insights into current trends in the field and recommendations on potential research directions.
The article provides a comprehensive review of Pareto pruning methods for multi-objective optimization, offering a novel classification scheme for MOO methods according to the resolution process, role of the DM, and characteristics of the final output. The authors provide precise definitions of related terms and discuss different pruning methods in four subclasses: preference-based, diversity-based, efficiency-based, and problem specific methods. Furthermore, indicators employed to assess the performance of the pruning method in previous literature are compiled and discussed.
The article is generally reliable and trustworthy due to its comprehensive coverage of existing literature on Pareto pruning methods for multi-objective optimization. It presents both sides equally by providing an overview of MCDA methods for a priori decision-making as well as discussing evolutionary algorithms to address MOO problems in the a posteriori mode. Additionally, it offers insights into current trends in the field and potential research directions based on comparative analyses across different MOO classes.
However, there are some points that could be improved upon in order to make this article more reliable and trustworthy. For instance, while it provides an overview of MCDA methods for a priori decision-making, it does not provide any examples or case studies illustrating how these methods can be applied in practice. Additionally, while it discusses evolutionary algorithms to address MOO problems in the a posteriori mode, it does not provide any details on how these algorithms work or how they can be used effectively in practice. Finally, while it offers insights into current trends in the field and potential research directions based on comparative analyses across different MOO classes, it does not provide any evidence or data to support its claims about these trends or directions.