An evolutionary approach for layout optimization of a three-dimensional truss

Layout optimization is defined as the simultaneous optimization of three aspects, i.e. topology, size and geometry. In this paper, a new approach based on evolutionary genetic search is employed to find the optimal design. The proposed approach departs from standard Genetic Algorithms (GA) in that it does not use binary strings in the representation of the possible solutions. The approach is applied to a 2-D benchmark example and a well-known space truss example. This 25-bar-truss is solved by using the proposed approach, with good results. The effect of removing the restriction of area grouping and the free location of joints and supports in the 25-bar truss are also investigated, with and without symmetry. Major drawbacks in finding the optimal solution by using the ground structure approach are overcome in the proposed method.