Improved Doppler effect metaheuristic with adaptive learning and opposition-based initialization for large-scale frequency-constrained dome-truss optimization
This paper introduces the improved Doppler effect (IDE) algorithm, a hybrid metaheuristic with adaptive learning and opposition-based initialization, specifically designed for frequency-constrained structural optimization problems. Although the original Doppler effect–mean Euclidean distance threshold (DE-MEDT) algorithm performs well on benchmark functions and structural optimization problems with static constraints, it suffers from slow convergence and reduced accuracy in high-dimensional applications. IDE addresses these limitations through five key enhancements: (1) an opposition-based learning (OBL) initialization strategy to promote population diversity and broaden search coverage; (2) an adaptive bounce-back boundary-handling mechanism that preserves natural search trajectories while enforcing boundary constraints; (3) an adaptive hybrid elite-guided position-update scheme to effectively balance exploration and exploitation; (4) a memory-driven parameter adaptation mechanism to generate new parameter values based on historical success; and (5) a periodic population diversification mechanism to improve global exploration and prevent premature convergence. The proposed IDE algorithm is validated on three challenging large-scale dome-truss sizing optimization problems with frequency constraints. Comparative results demonstrate that IDE achieves higher accuracy, faster convergence, and improved robustness relative to DE-MEDT and state-of-the-art metaheuristics in this problem domain, providing an efficient and reliable framework for frequency-constrained structural optimization.
Kaveh, A., Biabani Hamedani, K., and Hosseini, S.M. (2025). Improved Doppler effect metaheuristic with adaptive learning and opposition-based initialization for large-scale frequency-constrained dome-truss optimization. Computers & Structures, 327, 108246. https://doi.org/10.1016/j.compstruc.2026.108246
This work is based upon research funded by Iran National Science Foundation (INSF) under project No. 4024911. The first and second authors gratefully acknowledge this support.