Hanan Tanasra

Hanan Tanasra is an Architect and an MSc student at the Faculty of Architecture and Town Planning at the Technion.

E mail Linkedin

Research Topic: Automation in Design
Combining Machine Learning and Rule-Based Layout Checking to Generate Interior Design

Automation in architecture has developed rapidly through the use of parametric design, generative and computational design tools focusing on efficiency, reducing manual steps, work speed, and accuracy. Recently, ML and AI have given the world more technological tools, but few of them have been used in architecture. (Sönmez, 2018)
The lifecycle of a project includes detailing and evaluating, architects detail 2D plans with walls, furniture, doors, and windows. They change and redraw the design many times in order to achieve a more efficient design, especially when collaborating with counselors for reviewing the design performance. False information or design errors transfer makes the process time-consuming. (Bloch & Sacks, 2018)
The research intermediates the early stages of residential interior design by suggesting an alternative design evaluation methodology. Applying ML algorithms to furnish empty 2D floor plan drawings with the combination of Rule-Based Layout checking for evaluating the rooms’ design, connectivity logic, furniture filling, and inserting room quality rating.
Design decision-making questions the quality of the design alternatives and declines irrelative ones relying on certain criteria considerations that are derived nowadays from the intuitive capabilities of a designer. The research introduces a different method for testing interior design quality through technological tools, proposing an automated process for choosing the most suitable design.
“to what extent can designers rely on intuition based on experience to make those decisions? How can we verify such an effectiveness?” (Haymaker et al., 2018) Not all designers know the design rules considerations, leading to intuitive or less accurate choices. Moreover, design evaluation by humans can take more time if done manually, leading to one or two thorough discussions, canceling the diversity of design choices.
Therefore, the proposal output investigates the performance of two workflows, questioning the influence on the interior design work, comparing the human efforts to systematic computed tools results, and demonstrating the high quality of this approach in helping designers.