Reducing the number of different members in truss layout optimization

Abstract Despite the long history of truss layout optimization approach, its practical applications have been limited, partly due to high manufacturing costs associated with complex optimized structures consisting members different cross-sectional areas and member lengths. To address this issue, study considers optimizing comprising limited types members. On topic, two distinct problems are considered, wherein first problem, same type share area (i.e., section-type problem); in second length member-type problem). A novel post-processing approach is proposed tackle target problems. In from traditional geometry approaches used as starting points, which then separated into groups by k-means clustering approach. Subsequently, clustered geometrically reduce deviations differences between area/length values corresponding cluster means). Several 2D 3D examples presented demonstrate capability approaches. For can be reduced near 0 for any given number. The problem relatively more complex, but providing clusters, maximum deviation below thresholds. Through number trusses substantially decreased, thereby significantly reducing costs.