Optimization of a user-involved floor layout recommendation system at the operation stage

The parametric modelling method, an algorithmic thinking process based on visual programming, is recently established as a supportive tool in the decisionmaking practice for architecture. For the time being, the method helps to inform, control and optimize the architectural design by expressing the input parameters and conditional rules that are generated according to the design objectives. This paper describes how a parametric-model-based recommendation system is developed for an interior floor layout optimization problem which supports user involvement in different aspects of the process. The system connects the design objectives and the user preferences to propose customized space layouts in the operation stage of buildings, and one of the feasible relaxations of this design study is to substantially generalize it as a multi-objective optimization problem. As such, the algorithmpart of the system contains three different functionalities: (a) a screening scheme to select the available spaces concerning given requirements, (b) a generation process to figure out different possibilities of interior plans, and followed by (c) an evaluation system to compare and recommend the best-matching solutions. Simultaneously, the user interface of the system builds up various interactions between the users and the parametric models throughout the process of design, so as to collect the criteria, preferences, and priorities of the design objectives. The recommendation system is also implemented in a real case study, the floor planning of the IAK-1 building of the European Investment Bank in Kirchberg, Luxembourg, to assess the algorithm performance and user experience. The results illustrate the applicability of this approach in real-life design, and the pros and cons of the generated plans are also analysed by comparing to traditional designs given by expert architects.