Managing variable-dimension structural optimization problems using generative algorithms

摘要: This article introduces a novel design abstraction concept for efficient truss topology and geometry optimization. The core advancement introduced here is to represent using rules of generative algorithms, operate on algorithm genetic rather than directly the description. indirect representation supports exploration variable high-dimension topologies. Generative strategies are also independent any kind ground structure, thus avoiding inherent limitations structure approaches that may hinder innovative solutions by defining priori what topologies be considered. We present new automatically satisfy structural stability constraints, can produce with diversity patterns within an individual design. Truss optimized in outer-loop operates rules, size optimization performed inner-loop each candidate sequential linear programming. proposed methodology concurrent topology, geometry, size. layer variable-dimension structures, which generated from same fixed-dimension rule set. Finally, we demonstrate effectiveness examining archetypal two- three-dimensional problems.