Designing a generalised reward for Building Energy Management Reinforcement Learning agents

Martinez, Ruben MuleroGoikolea, Benat ArregiBeitia, Inigo MendialduaMartinez, Roberto GarayMulero, RubénArregi, BeñatMendialdua, IñigoGaray, RobertoSolic, PetarNizetic, SandroRodrigues, Joel J. P. C.Rodrigues, Joel J.P.C.Gonzalez-de-Artaza, Diego Lopez-de-IpinaPerkovic, ToniCatarinucci, LucaPatrono, Luigi2021-09-08Martinez , R M , Goikolea , B A , Beitia , I M , Martinez , R G , Mulero , R , Arregi , B , Mendialdua , I & Garay , R 2021 , Designing a generalised reward for Building Energy Management Reinforcement Learning agents . in P Solic , S Nizetic , J J P C Rodrigues , J J P C Rodrigues , D L-I Gonzalez-de-Artaza , T Perkovic , L Catarinucci & L Patrono (eds) , unknown . 2021 6th International Conference on Smart and Sustainable Technologies, SpliTech 2021 , IEEE , pp. 1-6 , 6th International Conference on Smart and Sustainable Technologies, SpliTech 2021 , Bol and Split , Croatia , 8/09/21 . https://doi.org/10.23919/SpliTech52315.2021.9566345conference978-1-6654-4202-2978-953-290-112-2978-9-5329-0112-29789532901122researchoutputwizard: 11556/1256researchoutputwizard: 11556/1255Publisher Copyright: © 2021 University of Split, FESB.The reduction of the carbon footprint of buildings is a challenging task, partly due to the conflicting goals of maximising occupant comfort and minimising energy consumption. An intelligent management of Heating, Ventilation and Air Conditioning (HVAC) systems is creating a promising research line in which the creation of suitable algorithms could reduce energy consumption maintaining occupants' comfort. In this regard, Reinforcement Learning (RL) approaches are giving a good balance between data requirements and intelligent operations to control building systems. However, there is a gap concerning how to create a generalised reward signal that can train RL agents without delimiting the problem to a specific or controlled scenario. To tackle it, an analysis and discussion is presented about the necessary requirements for the creation of generalist rewards, with the objective of laying the foundations that allow the creation of generalist intelligent agents for building energy management.6203716enginfo:eu-repo/semantics/openAccessDesigning a generalised reward for Building Energy Management Reinforcement Learning agentsconference output10.23919/SpliTech52315.2021.9566345Reinforcement learningRewardGeneralisedBuildingEnergy efficiencyHVACReinforcement learningRewardGeneralisedBuildingEnergy efficiencyHVACArtificial IntelligenceComputer Science ApplicationsDecision Sciences (miscellaneous)Renewable Energy, Sustainability and the EnvironmentSafety, Risk, Reliability and QualityHealth InformaticsSDG 7 - Affordable and Clean EnergySDG 12 - Responsible Consumption and ProductionSDG 13 - Climate ActionFunding InfoThe work described in this paper was partially supported by the Basque Government under ELKARTEK project (LANTEGI4.0 KK-2020/00072).The work described in this paper was partially supported by the Basque Government under ELKARTEK project (LANTEGI4.0 KK-2020/00072).Project IDLANTEGI4.0 KK-2020/00072LANTEGI4.0 KK-2020/00072http://www.scopus.com/inward/record.url?scp=85118449323&partnerID=8YFLogxK