Multi-objective scientific approach to problem solving-inspired optimization integrated with the finite element method for automated structural design

This paper presents a simulation-driven framework integrating the Multi-Objective Scientific Approach to Problem Solving-inspired Optimization (MOSAPSO) algorithm with the finite element method (FEM) for automated structural design in construction. The proposed MOSAPSO integrates chaotic initialization, Lévy flight dynamics, elite population control, and sparsity-biased Pareto archiving to enhance convergence and diversity, while embedding the scientific research process, including review and problem definition, hypothesis formulation, data collection, and analysis and interpretation, into a unified optimization strategy. A temporal control strategy balances exploration and exploitation during the optimization process. Benchmarking on 24 CEC-2020 test functions reveals that MOSAPSO outperforms 11 established multi-objective algorithms across hypervolume (HV), generational distance (GD), and spacing (SP) metrics. Integrated with FEM, MOSAPSO–FEM automatically generates Pareto-optimal designs for five large-scale structural systems, balancing weight, displacement, and stability constraints. The framework provides a robust foundation for intelligent, simulation-driven decision-making in construction design, offering significant opportunities for integration with BIM, digital twins, and automated design tools.