This research project focuses on the development, applicability, and evaluation of a multi-agent systems (MAS) approach for facade design. A design methodology is proposed and tested through the implementation of a computational workflow which combines bespoke generative agent based modelling techniques with numerical (environmental) analysis. The methodology is applied for the generation of facade panels on office buildings. Through the use of the MAS, a large solution space of façade designs are generated and ranked in an automated fashion for their improvement in their affect on lighting distribution, intensity, and efficiency.
The designer describes the component, generation rules and deisgn objectives and the system generates deisgn alternatives and provides feedback for the design performance of each alternative. Through an case study of an office building in Los Angeles, facade design are generated and analyzed environmentally and different heuristic and optimization functions are investigated in order to drive the generative tool towards meainingfull design solutions
The aim of the research is to enable designers explore larger solution spaces and articulate the benefits of highly intricate geometry as not only a feasible but as well potentially more optimal approach within contemporary architectural discourse and production.