CARDINALITY CONSTRAINT FOR TRUSS SIZING OPTIMIZATION OF PLANAR TRUSS STRUCTURES

Abstract

The optimization goal function of minimal weight generally produces results which use many different cross-sections. In practice, using such a large number of different bar stocks is irrational, impractical and expensive. This research explores the influence of implementing cardinality constraints in sizing optimization by limiting the number of possible different cross-sections that are used in any iteration of the optimization process. The example chosen to showcase the difference in using such an approach is the typical 47-bar planar problem. Results using the cardinality constraint are compared to results from an analytical solution where the entire structure was sized according to the structure’s most stressed bar, as well as to comparable results from literature, which do not limit the number of different cross-section dimensions.