Fully Constrained Design: A general and scalable method for discrete member sizing optimization of steel truss structures

• We present the Fully Constrained Design (FCD) method for discrete sizing optimization.
• We compare FCD to leading existing methods using benchmarking examples.
• FCD produces superior quality solutions (>4%) compared to optimality criteria.
• FCD is roughly 100× more computationally efficient than leading heuristic methods.
• We present a successful industry application of FCD that yields cost savings of 19%.

Fully Constrained Design (FCD) is a new method for discrete sizing optimization of steel structures that balances computational efficiency with solution quality for application to large-scale problems. The proposed method is based on optimality criteria, but does not require gradient information and handles discrete variables. Based on benchmarking studies, FCD produces superior quality solutions to optimality criteria (>4%), but inferior to heuristic methods (<2%). FCD is approximately 10× less computationally efficient than optimality criteria and 100× more efficient than heuristic methods. We present a successful industry application of FCD that yields cost savings of 19% compared to conventional design methods.