Including problem-knowledge based modification into a Differential Evolution Algorithm for optimizing planar moment-resisting steel frames

The Differential Evolution Algorithm (DEA) has been demonstrated to be capable of effectively addressing engineering challenges, although its performance varies considerably when applied to different problems. Customizing the algorithm to the specific characteristics of a given problem has been identified as a valid strategy to enhance its effectiveness and reliability. In this study, a tailored version of the DEA is proposed for the optimization of planar Moment-Resisting Steel Frames (MRSFs) subjected to static loads . A diverse set of heuristics and techniques were incorporated, including advanced strategies for parameter control, initialization, mutation operators , crossover operators, diversity conservation, constraints handling, and dynamic population management. To evaluate the performance of the proposed heuristics and techniques, 7800 DEA configurations were applied to the optimization of seven representative MRSFs. Results indicate that through problem-specific modifications the DEA is highly likely to identify the optimal solutions. By emphasizing both computational efficiency and solution quality, this research provides valuable insights into enhanced applicability of the DEA to structural optimization problems. It is shown that a customized algorithm is a reliable, effective, and robust tool to optimize MRSFs. • Tailored DEA optimizes MRSFs using heuristics and adaptive strategies effectively. • Six quality filters ensure robust DEA evaluation, aligning efficiency and reliability. • Testing 7800 configurations on 7 structures ensures robustness across all dimensions.