Integrating expertise and parametric analysis for a data-driven decision-making practice

Abstract

This study explores the integration of expert design intuition and parametric data analysis. While traditional professional design expertise helps to rapidly frame relevant aspects of the design problem and produce viable solutions, it has limitations in addressing multi-criteria design problems with conflicting objectives. On the other hand, parametric analysis, in combination with data analysis methods, helps to construct and analyze large design spaces of potential design solutions and tradeoffs, within a given frame. We explore a process whereby expert design teams propose a design using their current intuitive and analytical methods. That design is then further optimized using parametric analysis. This study specifically explores the specification of geometric and material properties of building envelopes for two typically conflicting objectives: daylight quality and energy consumption. We compare performance of the design after initial professional design exploration, and after parametric analysis, showing consistently significant performance improvement after the second process. The study explores synergies between intuitive and systematic design approaches, demonstrating how alignment can help expert teams efficiently and significantly improve project performance.

"TAL-CAT" a Computer-Aided Tool Prototype to Quantify User Experience in Design Workflow: A Case Study of Teamwork Assessment in Primary Care Clinics

“…Although the evidence-based design has endeavored to evaluate healthcare settings' performance from different aspects effectively, applying this process requires a lot of time and labor. The progress of computational design has widely promoted the development of many Computational-Aided Design tools (Bernal et al, 2015(Bernal et al, , 2020. The paper shows how computational design can be expanded to design evidence-based analytical tools to evaluate designs' performance.…”

Design Spaces of Structurally Pre-evaluated Funicular Shapes

“…This implementation works as a three-step process: generation, selection, and optimization. The generation process outputs a design space, supported by a custom Design of Experiments for sampling and reducing the size of this design space (Bernal et al, 2020). The selection step is based on designers' preferences within the pool of admissible alternatives.…”

Constraint handling methods for a generative envelope design using genetic algorithms: The case of a highly constrained problem

“…Moreover, it has been shown that integrated approaches are more efficient than traditional empirical design methods. 26 However, these methods are still under-used by architecture agencies, 7 even if some notable exceptions can be found. 27,28 According to the literature, 1,[29][30][31] there could be many reasons for this underutilization: ideological obstacles, excessive computation time, and lack of training magnified by the multidisciplinary nature of the approach.…”