A Comprehensive Survey of Hybrid Whale Optimization Algorithm with Long-Short Term Memory: Applications, Improvements, and Future Perspective

Abstract

This paper systematically reviews the Whale Optimization Algorithm (WOA) in combination with Long-Short Term Memory (LSTM) networks. WOA algorithm is used as an efficient and robust metaheuristic algorithm in combination with LSTM networks to model temporal dependencies and extract sequence features to solve optimization and prediction problems. In this paper, improvements derived from WOA-LSTM models are identified and reviewed based on 116 papers published between 2020 and April 2025 from 8 reputable publishers. In the methodology section, the process of collecting, analyzing, and categorizing papers based on the year of publication and the publisher is carried out. It shows that the most significant number of papers belong to the year 2024, which were published by Springer (29%) and Elsevier (27%). This paper provides a coherent and in-depth perspective for researchers and practitioners interested in exploiting the potential of combining metaheuristic algorithms and deep networks, especially LSTM, in solving challenging real-world problems. It results show that time series forecasting and engineering systems design were identified as the most widely used areas. Also, the challenges in designing, implementing, and optimizing WOA-LSTM hybrid models are discussed, and future research directions in this field are proposed, focusing on improving performance and increasing scalability.