Answers are grounded in this paper and cite their sources.
This study presents Scalarized Pareto Optimization (SPO), a workflow for computationally expensive multi-criteria design optimization using high-fidelity simulations. SPO integrates established scalarization and single-objective search concepts into structured sequential searches that reuse previously evaluated designs through a cumulative database for initialization, normalization, and Pareto extraction. The framework is applied to conventional reinforced concrete (RC) frames and composite systems combining RC columns with transversely hybrid variable-section (THVS) steel girders, optimizing life-cycle environmental impact and constructability, with THVS girder deflection as a third objective for the composite typology. Performance is assessed against PESA-II and SPEA2 under fixed budgets. Based on three independent high-fidelity optimization runs per method and case study, the adapted SPO strategy reduced the number of structural evaluations by about 50% in the two-objective RC problems while achieving higher hypervolume (HV) values than the best benchmark under the adopted reference points. In the three-objective RC–THVS problems, computational savings were less consistent, but SPO achieved higher HV values, reaching over twice the benchmark values in several cases. These savings apply only to the final high-fidelity optimization stage and exclude the offline surrogate development cost. Structurally, RC–THVS solutions in the environmentally driven Pareto region reduced life-cycle environmental impact by about 20–28% relative to RC solutions for the investigated cases and assumptions, highlighting the composite system’s environmental potential. Minimizing THVS deflection as an objective also strongly penalized environmental and constructability performance. Overall, SPO provides a competitive workflow for targeted high-fidelity structural optimization within the evaluated algorithms, budgets, and case studies.
No takes yet. Share an insight, caveat, or question.
Negrin et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: