Artemisinin slime mold algorithm for large-scale truss

Artemisinin slime mold algorithm for large-scale truss optimization under frequency constraints

A B S T R A C T

This study presents a novel bio-inspired metaheuristic, the artemisinin slime mold algorithm (ASMA), which integrates the adaptive foraging behavior of slime molds with the chemotactic characteristics of artemisinin, a compound known for its antimalarial properties. ASMA enhances global optimization performance through a hybrid mechanism that dynamically balances exploration and exploitation. Agents traverse the search space via chemotaxis like behavior, while the artemisinin inspired component directs them toward promising solutions, thereby accelerating convergence and improving solution accuracy. The algorithm was evaluated on the CEC 2017 benchmark set and demonstrated superior performance over several leading algorithms. Additionally, ASMA is utilized for sizing optimization of truss structures under dynamic frequency constraints, which require intensive eigenvalue computations. To address this challenge, the largest eigenvalues of sparse matrices (LESM) technique is employed to extract only the dominant eigenvalues relevant to dynamic constraints, avoiding fullspectrum eigenvalue analysis. This significantly reduces computational costs while maintaining accuracy, particularly for large-scale truss systems with inherently sparse stiffness matrices. The ASMA-LESM framework is validated on a range of dome structures, demonstrating high efficiency, scalability, and robustness for structural optimization under dynamic constraints. The source code is available at: https:/github.com/nut123456/ASMA.