Yarramsetty, C.Moger, T.Jena, D.Rao, V.S.2026-02-052025IEEE Access, 2025, Vol.13, , p. 80871-80888https://doi.org/10.1109/ACCESS.2025.3567464https://idr.nitk.ac.in/handle/123456789/28230This paper provides a detailed review of reliability assessment methods for composite power systems, focusing on integrating renewable energy and advanced computational approaches. The study classifies existing research into three main areas: investigation studies, planning and optimization studies, and efficient evaluation studies. Findings indicate that machine learning techniques are increasingly important in improving accuracy and computational performance in reliability analysis. A comparative examination of conventional and Machine Learning (ML)-based methods demonstrates that deep learning models, such as Convolutional Neural Networks, offer substantial reductions in computational time while maintaining reliability assessment precision. The review also identifies key research gaps, such as the need for realistic test systems and enhanced hybrid ML strategies. Additionally, recommendations are proposed to address these challenges and strengthen the effectiveness of future reliability evaluation methodologies. The insights presented in this study aim to support researchers and industry professionals in developing more efficient and scalable reliability assessment models for evolving power systems. © 2013 IEEE.artificial intelligencecomputational efficiencyMonte Carlo simulationPower system adequacyrenewable integration