Advances in AI and optimization for water quality index prediction and integrated water resources assessment: A review

Abstract

Accurate water quality assessment is critical for sustainable water resources management under growing environmental pressures. The Water Quality Index (WQI) provides a practical framework for summarizing complex water quality data into a single indicator. This review examines recent advances in artificial intelligence and optimization techniques for WQI prediction, with a focus on machine learning, ensemble models, deep learning, and hybrid approaches. Existing studies demonstrate strong predictive capabilities but remain largely model-centric and limited by localized datasets and weak system integration. This review identifies methodological limitations and outlines key components required for future integrated monitoring frameworks, including data acquisition, model interpretability, and uncertainty-aware decision support. The findings provide guidance for advancing toward scalable and transparent water quality assessment systems.