Comprehensive learning phasor particle swarm optimization of structures under limited natural frequency conditions

• Published in: Acta mechanica Sinica Vol. 39; no. 4
• Main Authors: Pyone, Ei Cho, Tangaramvong, Sawekchai, Van, Thu Huynh, Bui, Linh Van Hong, Gao, Wei
• Format: Journal Article
• Language: English
• Published: Beijing The Chinese Society of Theoretical and Applied Mechanics; Institute of Mechanics, Chinese Academy of Sciences 01.04.2023; Springer Nature B.V
• Edition: English ed.
• Subjects: Algorithms; Classical and Continuum Physics; Computational Intelligence; Engineering; Engineering Fluid Dynamics; Heuristic methods; Learning; Optimization; Particle swarm optimization; Phasors; Research Paper; Resonant frequencies; Robustness (mathematics); Structural members; Theoretical and Applied Mechanics; Trigonometric functions; Trussed structures; Natural frequency; Comprehensive learning; Particle swarm optimization; Phasor theory; Space truss
• ISSN: 0567-7718; 1614-3116
• DOI: 10.1007/s10409-023-22386-x

The paper presents the combined phasor particle swarm optimization and comprehensive learning method for the optimal design of space truss structures under the constraints addressing limited natural frequency. The proposed approach enhances the performance of the standard particle swarm optimization approach by incorporating the efficient phasor theory in mathematics with comprehensive learning strategy. Within the optimization process, a so-called phase angle adopts the periodic sine and cosine functions to model the key parameters defining velocities that learn from a sole medium of exemplar’s velocity selected among previous best positions of all particles. The scheme not only enables the fast learning of swarm particles, but also computes the safe and optimal size distribution of structural members under applied forces and natural frequency conditions at modest computing resources. Various design benchmarks on practical-scale (three-dimensional space) engineering applications have been successfully solved by the proposed design method. These illustrate the accuracy and robustness of the algorithm as compared with various state-of-the-art metaheuristic approaches recently developed.