Truss layout design and optimization using a generative synthesis approach

• The approach uses load flow principal for automating the design of cable trusses.
• Grammar rules use this domain knowledge to quickly find optimal solutions.
• Approach and results handle problems with multiple loads and supports.
• Approach and results address 2D and 3D test problems without increase in the time.
• Tree-search algorithm explores design space of many topologies.

This paper describes a generative methodology for optimum design of truss structures. The novelty of the proposed method is its combination of generative design synthesis methods with conventional gradient-based optimization and simulation models. This combination is able to achieve optimal topologies and shapes for cable trusses under various constraints: stress, displacement, stability. Furthermore, manufacturing issues properly limit the generated solutions. A graph grammar is used to define different topologies for a given load problem. The effectiveness of the method is checked by solving a variety of available test problems found in the literature as well as several new three-dimensional solutions.