The main objective of this project is to develop an innovative Neurone Network Architecture of the Model Predictive Management (MPV) for the management of heating systems in multi-apartment buildings and to assess the efficiency of its use. The project is aimed at increasing energy efficiency, improving the accuracy of forecasting and increasing the comfort of the population. It will compare different neural network architectures for buildings of different sizes, optimise network configuration parameters and create scalable methods for switching from traditional control systems to data-based MPVs. The aim of this transition is to achieve significant energy savings and reduce the carbon footprint of the European Union (EU) building sector, including. in Latvia. Given that most buildings in the EU are energy inefficient, the introduction of optimised management systems can significantly reduce energy consumption and the associated greenhouse gas emissions. This study will systematically assess neural network models across different building types, taking into account building specificities and the impact of data volume and model configuration on performance. The project also provides for the integration of neural network-based MPVs into the building system, moving from the traditional proportionally integrative differential (PID) system management. The results of the development will provide scalable, data-driven solutions that will increase energy efficiency and reduce costs, while maintaining the comfort of residents. The main activities of this industrial research project include: - an in-depth analysis of the current situation, as well as the selection of suitable alternative solutions and their initial assessment; - development of concept and practical solutions for effective control and reduction of thermal energy consumption, collection and analysis of necessary data in multi-apartment buildings, as well as improvement of thermal comfort; - the installation of additional measurement equipment for the practical testing and improvement of artificial intelligence models developed in experimental buildings and heat distribution units; - an analysis and assessment of possible strategies for minimising energy losses in the heating pipeline network; - the compilation and publication of results, ensuring the protection of intellectual property. Detailed project management activities and systematic risk analysis are planned to ensure successful project implementation. The dissemination of knowledge will be carried out through a variety of information and educational activities, including presentations at conferences and scientific publications aimed at professionals in the field. Two project partners will provide infrastructure and equipment for data collection from existing multi-apartment buildings, including support for data connection and server infrastructure. They will also provide the necessary infrastructure solutions for the management of heating systems, such as sensors, meters and programmable remotely controlled radiator thermostats. Access to historical monitoring data from existing buildings will be essential for optimising both existing and newly developed algorithms. In addition, partners will practically install additional data storage sensors to meet the research needs of the project. The main expected results are the demonstration and implementation of the Model's Predictive Management Approach for optimizing the management of heating systems by integrating it into the management systems of experimental buildings. This approach will use data analytics-based models and new neural network architectures, enabling the development of easily scalable solutions with significant potential to improve indoor climate and energy efficiency without significant resource investments. In addition, the analysis of the regimes in the heating pipeline network (from the heat source to the building) and the associated heat losses, as well as the inclusion of weather forecasts, will help to optimise and plan more efficient heat carrier temperature regimes. This will also contribute to a more efficient use of energy and the reduction of heat losses. The project corresponds to RIS3 Smart Specialisation Area ‘Smart Energy and Mobility’ and RIS3 Growth Priority 3 ‘Improving energy efficiency, including the creation of new materials, optimisation of production processes, introduction of technological innovations, use of alternative energy sources and other solutions’. The results of the project will contribute to the development of the economic sector "Electricity, Gas Supply, Heating and Air Conditioning". According to the OECD classification, this project falls under field 2.2. "Electrical, electronic and information engineering". The scientific director of the project is Dr. Phys. Andris Jakovičs. The total cost of this non-commercial project, which is planned to be implemented over a 36-month period (1 September 20