Akademik Veri Yönetim Sistemi
STRUCTURAL AND MULTIDISCIPLINARY OPTIMIZATION, cilt.68, sa.12, 2025 (SCI-Expanded, Scopus)
In this study, a robust high-dimensional discrete evolution strategy algorithm, termed HADES, is proposed for solving large-scale structural design optimization problems. While the individual components of HADES are well established in the evolutionary computation literature, the novelty of the algorithm lies in their purposeful integration and adaptation into a cohesive framework specifically tailored for discrete truss optimization with hundreds of design variables. Distinct from earlier discrete ES variants, HADES avoids parameter adaptation schemes that are prone to mis-adaptation in vast search spaces, instead employing a simplified yet effective strategy that balances exploration and exploitation. The performance of the proposed algorithm was evaluated using four high-dimensional structural optimization problems from discrete sizing design of steel trusses, with the number of design variables ranging from 96 to 336. To obtain statistically meaningful data, the algorithm was run independently 35 times for each problem. The results demonstrate that HADES consistently produces best-known solutions for the tested problems. Moreover, it exhibits remarkable consistency across multiple independent runs, indicating a high probability of yielding a good optimum solution in each execution. Collectively, these findings suggest that HADES is a practical, efficient, and reliable tool for high-dimensional structural optimization problems, particularly in discrete sizing design of steel truss systems.