Abstract
Multi-objective optimization is considerably increasing its importance in building design since the design goals are moving from the solely energy saving target to the whole building performance, comprehensive of energy, cost and comfort targets. Optimization algorithms coupled with building simulation codes are frequently used in academic researches. However, they are limitedly adopted in real building design due to the high number of expensive simulation runs required by optimization algorithms such as direct search methods, evolutionary algorithm, particle swarm optimization and hybrid algorithms. For this reason, an efficient optimization scheme is essential for the diffusion of the optimization tools in building performance design outside the academic world. The research focuses on the development of an Efficient Global Optimization (EGO) scheme based on a radial basis function network (RBFN) model to emulate the expensive evaluations of the building performance simulation (BPS). The test bed of the method is the optimal building refurbishment of three simplified module representative of existing buildings, for which the optimal solutions have been also calculated by using the brute force approach, i.e. evaluating the performance of all the possible combinations of the retrofit measures. Finally, the EGO performances were also compared with those offered by the popular Non Sorting Genetic Algorithm (NSGA-II).
The results show the extent to which the EGO algorithm is able to find optimal solutions with a reduced number of expensive simulation runs. This capability makes the EGO algorithm suitable for the optimization of expensive simulation codes such as lighting models, CFD codes or dynamic simulation of building and HVAC systems.