Personalized thermal comfort modeling through genetic algorithm

No energy-saving actions are implied without maintaining comfortable levels for the residents, as lack of comfort results in stress while threatening the occupants health and well-being. In this paper, a novel algorithm for the estimation of individual metabolic rate and comfort level is introduced. A Genetic Algorithm is utilized for the metabolic rate computation of thermal comfort by eradicating all speculative factors, while creating a personal thermal comfort evaluator. Based on the occupants feedback, the subjective personal factors of thermal comfort (clothing insulation, metabolic rate) are estimated, generating a personal thermal comfort profile. Therefore, the proposed approach can be adapted to create the resident’s personal preferences to achieve accurate comfort level estimation. Ultimately, the proposed algorithm is evaluated against real-life indoor sensor data and users’ feedback, while the experimental results illustrate the efficiency of the proposed system. The Genetic algorithm succeeds 100% in finding the optimal metabolic rate solution while improving the thermal comfort estimation error. The thermal comfort profile is 98% accurate compared to a solution based on ASHRAE tables that has 73% of accuracy.