Applying desirability functions to preference modelling in low-energy building design optimization

Building performance optimization is a valuable aid to design decision-making. Most existing research takes an ‘a posteriori’ approach, where stakeholder preferences are considered after deriving optimised results. Whilst this approach yields technically optimal solutions, it overlooks sub-optimal solutions that still satisfy stakeholder preferences. This research develops a technique to incorporate preferences into optimization by applying a “desirability function” to each criterion for multiple stakeholders. The approach enables the tradeoffs between decision-makers to be visualised as a Pareto frontier and aids “democratic” decision-making. Hence, incorporating preferences in advance of optimization may increase the likelihood of finding a desirable solution.