ABSTRACT

The objective of this contribution is to give a framework for the general optimization of 3D trusses using evolution strategies. These stochastic methods have proven their efficiency in the past for several problems. They are very robust and converge in most cases to the true optimum. An important part of the optimization process is the verification of the constraints. This will be explained briefly for the case of truss structures which are often used in civil structures such as bridges and roof structures. The whole optimization technique is implemented in the Space Truss Evolution Strategy (STES) program written in JAVA and in which the cross-sectional areas of the truss members, the coordinates of the nodes and the topology of trusses can be optimized simultaneously. However, in this paper each part of the general optimization is tested seperately and the results are compared with those found in the literature. Finally an example is presented where a one-layered dome is submitted to the evolutionary optimization process. For this particular structural engineering problem it is imperative to verify the global buckling of the dome, since this failure phenomenon tends to be dominant for this kind of structures.