This work explores structural form finding through the implementation of behavorial based strategies and Multi Agent Systems. The objective is to facilitate the design generation, analysis, and construction of complex shell structures using large scale additive manufacturing (contour crafting). A computational design methodology is proposed where a given design brief (i.e. pavilion design) is decomposed into design requirements and a series of performance based criteria including: a) the structural logic of the reciprocal frame b) a shading strategy and c) building constraints and cost of robotic construction (additive manufacturing).
Through the implementation of a multi-agent system, design requirements are formulated into different agencies (i.e. generative, structural, materialization) and behavioral form finding is explored. Material and fabrication constraints, which relate to large scale additive manufacturing of concrete are modelled into the agent behavior. While the work intentionally privileges design intricacy and the investigation of non-linear design techniques it also aims to couple such generative approaches with numerical and analytical data for improving the design outcomes.
The use interactive optimisation and algorithms, such as mesh relaxation and ant colony optimizations, enable the designer to: a) formulate the objective functions, b) run a desired number of iterations, c) evaluate design outcomes both qualitatively and quantitatively, d) adjust the design parameter range according to her objectives and d) reinitiate the optimization routine. Thus, not only the capability of the MAS is investigated to not only manage design complexity but to also equally generate non-deterministic design outcomes.