This study aims to assess and compare the predictive performance of advanced forecasting approaches, including stochastic models, like the Geometric Brownian Motion (GBM), Fractional Brownian Motion (FBM), and Brownian Bridge (BB), as well as a Machine Learning (ML) model, the Long Short-Term Memory (LSTM), in predicting the price movements of energy commodities such as Brent oil, crude oil, gas oil, heating oil, and natural gas across eight geopolitical and macroeconomic events. To reassess the role of persistence in energy commodity prices, the study strengthens the Hurst analysis by combining classical rescaled range analysis with detrended fluctuation analysis, the Geweke-Porter-Hudak estimator, Lo's modified rescaled range statistic, bootstrap confidence intervals, and Generalized Autoregressive Conditional Heteroskedasticity (GARCH)-filtered residual diagnostics. The forecasting results further indicate that no single model dominates uniformly across all commodities and crisis regimes. BB records the most frequent best performance across the main accuracy criteria, while LSTM, GBM, and FBM remain competitive in specific commodity-period combinations. During the most recent period, with the Russia-Ukraine Conflict, GBM, FBM, and BB demonstrated strong forecasting performance across different time horizons, achieving an average Mean Absolute Percentage Error (MAPE) of less than 10% for crude oil and Brent oil and below 20% for heating oil and gas oil, based on 1,000 simulations. These results highlight the comparative effectiveness of both ML and stochastic methods, offering valuable insights and practical tools for policymakers and market agents to manage energy price risks more effectively.
The increasing penetration of renewable energy sources has intensified the need for intelligent and adaptive energy management in decentralized grid-connected systems. While solar photovoltaic (PV) energy is cost-effective and modular, its intermittency limits supply reliability, whereas geothermal energy provides a stable and dispatchable baseload but requires effective coordination in hybrid systems. Integrating solar PV and geothermal energy offers a promising solution; however, renewable variability, dynamic load behavior, power-electronic interactions, and grid constraints necessitate advanced Intelligent Energy Management Systems (IEMS). The present paper introduces a new hierarchical intelligent energy management system for a decentralized, grid-connected solar-geothermal hybrid energy system using a Swin transformer-based spatio-temporal perception model and an Mamba state-space decision-making model. To generate accurate short-term forecasts of renewable generation, load demand, and system operating states, the Swin transformer is used to extract hierarchical spatiotemporal features. The Mamba model then uses these high-level contextual representations to make low-latency, long-horizon optimal energy dispatch decisions, which have linear computational complexity and are thus appropriate for real-time control. Extensive simulation results demonstrate that the suggested Swin-Mamba system outperforms conventional control and optimization approaches, exhibiting a lower voltage deviation (+/- 1.2%), reduced frequency deviation (+/- 0.10 Hz), a shorter settling time of 3.1 s, higher renewable energy utilization (92%), and reduced grid dependency (10%) under dynamic operating conditions. The findings affirm the fact that the proposed Swin Transformer-Mamba-powered IEMS is an efficient, scalable, and cost-effective solution for managing intelligent operations of hybrid renewable energy systems, facilitating the development of low-carbon and resilient smart grid systems.
