MILP-based discrete sizing and topology optimization of truss structures: new formulation and benchmarking
Discrete sizing and topology optimization of truss structures subject to stress and displacement constraints has been formulated as a Mixed-Integer Linear Programming (MILP) problem. The computation time to solve a MILP problem to global optimality via a branch-and-cut solver highly depends on the problem size, the choice of design variables, and the quality of optimization constraint formulations. This paper presents a new formulation for discrete sizing and topology optimization of truss structures, which is benchmarked against two well-known existing formulations. Benchmarking is carried out through case studies to evaluate the influence of the number of structural members, candidate cross sections, load cases, and design constraints (e.g., stress and displacement limits) on computational performance. Results show that one of the existing formulations performs significantly worse than all other formulations. In most cases, the new formulation proposed in this work performs best to obtain near-optimal solutions and verify global optimality in the shortest computation time.
2022 Structural and Multidisciplinary Optimization (MILP-based discrete sizing and topology opti).pdf
publisher
openaccess
CC BY-NC-ND
3.51 MB
Adobe PDF
6ae3c5451c59f5285d27d43b8117f703