The path toward decarbonization in energy systems: A review on optimization, machine learning and spatiotemporal approaches

The decarbonization of energy systems requires analytical methods that combine optimization, machine learning (ML), and spatiotemporal analysis to capture the complex dynamics of energy production, distribution, and consumption used in the energy policy agenda under the transition toward a net-zero emissions energy system. The analysis is based on 82 peer-reviewed papers that cover the period 2001–2024 from Scopus, Web of Science, and Google Scholar databases. Based on the analyzed publications, the review focuses on four research strands (i) multiobjective optimization algorithms in energy modeling, (ii) spatial decomposition and environmental impact assessment techniques, (iii) geospatial data interpolation and autocorrelation analysis, and (iv) advanced econometric and ML approaches for spatiotemporal forecasting. The review offers a three-fold contribution (i) it systematically compares empirical spatiotemporal advances used to analyze decarbonization in the energy sectors, (ii) critically synthesizes spatiotemporal econometric techniques, optimization, and ML techniques in capturing spatial heterogeneity and temporal dynamics for sustainable energy development, and (iii) highlights methodological gaps relevant for energy decision-making under net-zero targets. Overall, the present review provides a holistic methodological framework that can support decarbonization efforts toward sustainable energy development.