Optimization in a realistic structural engineering context: redesign of the Market Hall in Ghent

Abstract

<p>Numerical optimization has a large potential in the context of structural design, but practical applications remain scarce. Even metaheuristic algorithms, which are easy to use, are rarely adopted in practice. Possible explanations are the fact that for problems with many design variables, metaheuristic algorithms converge slowly, and that structural optimization often leads to very complex structures, resulting in a high construction cost. The aim of this paper is to illustrate the potential of numerical optimization in a realistic design context. The focus is on the steel structure of the Ghent Market Hall, which is redesigned using a genetic algorithm. The structural member groups from the original design are maintained, such that the number of design variables is sufficiently low, and that the complexity of the design remains limited. Using this approach, a design is obtained that consumes 15 % less material than the original design.</p>