Optimal design of building environment with hybrid genetic algorithm, artificial neural network, multivariate regression analysis and fuzzy logic controller

Computational cost poses a major obstacle to the design of indoor environments with the current optimal method and computational fluid dynamics (CFD). A novel optimization method integrating a genetic algorithm (GA), an artificial neural network (ANN), multivariate regression analysis (MRA), and a fuzzy logic controller (FLC) was proposed in this paper to optimize the indoor environment and energy consumption based on simulation results. Thermal comfort (predicted mean vote) was set as the restrictive design objective. Indoor air quality (air age) and energy consumption were set as the optimal design objectives. Air supply parameters, such as ventilation rate, inlet temperature, and angle, were used as the design variables. The GA process was used to search for the optimal solution (individual), while the ANN and CFD tool were used to obtain the values of the objectives for each individual. MRA was used to reduce the variable space, and FLC was used to control the execution routine of the CFD process to reduce the computational cost. The results indicated that the ventilation rate has a lower impact on the design result compared with the other two design variables. When the MRA and FLC were included in the design process, the variable space and computational cost were reduced by 50% and 35.7%, respectively. The design efficiency was improved while the best found solution was maintained.