Fuzzy Machine Learning Applications in Environmental Engineering: Does the Ability to Deal with Uncertainty Really Matter?

Abstract

Statement of Problem: Environmental engineering confronts complex challenges characterized by significant uncertainties. Traditional modeling methods often fail to effectively address these uncertainties. As a promising direction, this study explores fuzzy machine learning (ML) as an underutilized alternative. Research Question: Although the potential of fuzzy logic is widely acknowledged, can its capabilities truly enhance environmental engineering applications? Purpose: This research aims to deepen the understanding of the role and significance of fuzzy logic in managing uncertainty within environmental engineering applications. The objective is to contribute to both theoretical insights and practical implementations in this domain. Method: This research performs a systematic review carried out in alignment with PRISMA guidelines, encompassing 27 earlier studies that compare fuzzy ML with other methods across a variety of applications within the field of environmental engineering. Results: The findings demonstrate how fuzzy-based models consistently outperform traditional methods in scenarios marked by uncertainty. The originality of this research lies in its systematic comparison and the identification of fuzzy logic’s transparent, interpretable nature as particularly suited for environmental engineering challenges. This approach provides a new perspective on integrating fuzzy logic into environmental engineering, emphasizing its capability to offer more adaptable and resilient solutions. Conclusions: The analysis reveals that fuzzy-based models significantly excel in managing uncertainty compared to other methods. However, the study advocates for a case-by-case evaluation rather than a blanket replacement of traditional methods with fuzzy models. This approach encourages optimal selection based on specific project needs. Practical Implications: Our findings offer actionable insights for researchers and engineers, highlighting the transparent and interpretable nature of fuzzy models, along with their superior ability to handle uncertainties. Such attributes position fuzzy logic as a promising alternative in environmental engineering applications. Moreover, policymakers can leverage the reliability of fuzzy logic in developing ML-aided sustainable policies, thereby enhancing decision-making processes in environmental management.