The global shift toward renewable energy is driven by the dual imperatives of rising energy demand and the need to reduce environmental harm caused by fossil fuels. However, renewables like wind and solar power pose unique challenges, particularly due to their intermittent generation and current limitations in energy storage technologies. Battery banks, commonly used to store surplus energy, degrade over time, making accurate forecasting of their remaining usable lifecycles critical for maintaining system reliability and efficiency. This study proposes a novel approach for forecasting battery health using an optimized long short-term memory (LSTM) network. To address the complexity of deep learning hyperparameter selection, a modified metaheuristic optimization algorithm is developed and integrated into a broader optimization framework aimed at improving model performance while minimizing overfitting. The method is benchmarked against several state-of-the-art optimizers, with results validated through comprehensive simulations and statistical analysis. This work contributes a scalable forecasting methodology, an effective optimization strategy, and interpretable results to support sustainable energy storage solutions.

Electroencephalography (EEG) represents an essential neurological assessment technique that captures the brain’s bioelectrical signals via electrodes positioned on the scalp. Although artificial intelligence has shown substantial potential across various areas of medical analysis, its application within neurodiagnostic procedures is still not widely examined. This study tackles that shortfall by introducing a novel methodology based on time-series categorization of EEG recordings, utilizing gated recurrent unit neural architectures to pinpoint irregular neural patterns, with a focus on seizure detection. To further boost the accuracy of the proposed framework, metaheuristic optimization strategies are applied for refining hyperparameter configurations. Moreover, a customized variant of particle swarm optimization is developed, designed specifically for this neurodiagnostic context. The performance of the approach is assessed using a carefully selected dataset containing authentic EEG traces from neurologically healthy subjects as well as individuals diagnosed with epilepsy. The resulting software-driven solution yields impressive outcomes, demonstrating strong capability in identifying anomalies even when operating with comparatively limited sample counts.