A review of artificial intelligence techniques for optimising decentralised greywater treatment under operational constraints

Abstract Decentralised greywater treatment systems are increasingly recognised as a sustainable solution to water scarcity, yet their adoption is hindered by challenges such as fluctuating influent quality and limited technical oversight. This review systematically examines 90 studies on the application of artificial intelligence (AI) techniques, including fuzzy logic (FL), artificial neural networks (ANN), support vector machines (SVM), ensemble methods, genetic algorithms (GA), and deep learning (DL), for optimising these systems under operational constraints. A decision-support framework based on the Analytic Hierarchy Process (AHP) is developed to guide the selection of AI models, considering factors such as predictive accuracy, computational demand, interpretability, energy efficiency, and governance readiness. The review finds that while complex models like deep learning offer high accuracy, methods such as fuzzy logic and ensemble approaches are more balanced and better suited to decentralised systems due to their lower computational requirements and better alignment with real-world operational challenges, as indicated by a composite utility ranking under balanced weighting. Future research should focus on hybrid AI systems, real-time data integration, and addressing governance challenges to enhance the scalability and sustainability of decentralised greywater treatment.