An outer approximation bi-level framework for mixed categorical structural optimization problems

In this paper, mixed categorical structural optimization problems are investigated. The aim is to minimize the weight of a truss structure with respect cross-section areas, materials, and type. proposed methodology consists using bi-level decomposition involving two problems: master slave. problem formulated as mixed-integer linear where constraints incrementally augmented outer approximations slave solution. addresses continuous variables problem. tested on three different test cases increasing complexity. comparison state-of-the-art algorithms emphasizes efficiency in terms optimum quality, computation cost, well its scalability dimension. A challenging 120-bar dome 90 choices per bar also tested. obtained results showed that our method able solve efficiently large-scale problems.