Robust building design optimization using the robust optimality criterion

This thesis introduces a robust building design optimization framework for handling building design problems that are exposed to uncertainties. The framework comprises five main steps, which can assist stakeholders in exploring the design space in the light of different sources of uncertainty, and obtaining a solution that performs satisfactorily with respect to the considered objectives and constraints: a) uncertainty quantification; b) exhaustive search; c) building performance simulation (BPS); d) robust optimization; and e) solution analysis. The application of these steps to the design optimization of a real-world building indicates that the framework can assist practitioners in designing buildings that are optimal in the defined objectives and robust to the inherent uncertainties.

Robust optimization is performed with help of a novel robustness indicator, called robust optimality criterion, which ensures the Pareto optimality of solutions under all uncertain conditions. In contrast to existing robustness indicators, the robust optimality criterion offers a control of the degree of robustness, which supports the articulation of the (risk) preferences of the decision-makers during solution analysis.