A review on the integrated optimization techniques and machine learning approaches for modeling, prediction, and decision making on integrated energy systems

Advances, Synergy, and Perspectives of Machine Learning and Biobased Polymers for Energy, Fuels, and Biochemicals for a Sustainable Future

“…141,142 Significant enhancements in the efficiency and sustainability of these systems have been realized by harnessing the predictive and analytical capabilities of ML algorithms. 141,142 The optimization of renewable energy production, particularly from solar and wind sources, stands as a striking illustration of this trend. 143,144 Real-time data are analyzed to predict energy output, allowing for more efficient grid integration and energy distribution.…”

Automated deep reinforcement learning for real-time scheduling strategy of multi-energy system integrated with post-carbon and direct-air carbon captured system

“…Carbon neutrality has become an ambitious goal; achieving the milestone on or before 2060 is of utmost importance [1]. The passion for its feasibility is evident with the rapid transformation of the energy sector.…”

“…In fact, to the best of the authors' knowledge, our study is the first to integrate these two CDR technologies as part of ZCMES. Conventionally, the optimal scheduling strategy has been implemented using mathematical programming in extant studies [1,30]. In contrast, the conventional control strategies suffer from two major setbacks, which are 1) the inability to be deployed in real-time; and 2) local optimum results due to a reduction in model complexity that sacrifices accuracy for computational cost.…”

Boosting operational optimization of multi-energy systems by artificial neural nets

“…Considering multi-energy systems in particular, only a few studies combine machine learning and optimization (Alabi et al, 2022a). Taheri et al (2021) used a deep recurrent neural net for the longterm planning and design of multi-energy systems.…”

“…However, none of the approaches uses a combined approach that tackles the problem of high and unreliable computation times in operational optimizations of multi-energy systems. In fact, it is observed that applying a combination of machine learning and optimization to the optimal decision-making on multi-energy systems is at an early stage that requires further research (Alabi et al, 2022a). In particular, a solution method is still missing that is applicable to largescale MILP multi-energy system models and provides feasible solutions for operational optimization in a reliably short time.…”