Advanced Multi-objective optimization of Energy management systems using Deep neural networks and HPC for real-time, multi-step energy forecasting

Robotina aims to revolutionize energy management in buildings (offices, hotels, shopping malls, public institutions) with a subscription service that combines real-time data and artificial intelligence (AI) to optimize energy consumption, reduce costs, and improve sustainability for end users. The service works with any building management system (BMS) and is based on a local EMS system with an EDGE computer, optimizing consumption based on weather conditions, energy prices, and tariffs. To efficiently forecast energy consumption (1–3 days in advance, with 15-minute accuracy), Robotina uses hybrid machine learning models, primarily BiLSTM and MLP. However, providing these forecasts to over 10,000 users is computationally demanding, as current CPU solutions require about a week to train a BiLSTM model with weekly data. Therefore, the use of high-performance computing (HPC) is crucial for efficiently processing data and implementing models at scale. HPC will enable Robotina to expand market opportunities and solidify its leadership in the field of energy management. The business experiment Advanced Multi-objective Optimization of Energy Management - AIMED-HPC has received funding through the FFplus project, which is financed by the European High-Performance Computing Joint Undertaking (JU) under grant agreement No. 101163317. JU receives support from the Horizon Europe program of the European Union.