An Improved Slime Mould Optimization Algorithm Based on Integrating Multiple Strategies and Its Applications

In this paper, a multi-strategy enhanced slime mould algorithm (IMSMA) is proposed to address issues such as population initialization, convergence speed, and the tendency to fall into local optima in the slime mould optimization algorithm (SMA). Firstly, an optimal point set is introduced for population initialization to enhance initial diversity and improve algorithm stability. Secondly, the fixed parameter controlling population randomness is adjusted to adaptively change with the iteration number, balancing the exploration capability at the beginning phases and the optimization capability in the final phases. Additionally, a convergence stagnation judgment mechanism is incorporated to prevent premature convergence. Comparative experiments with similar intelligent optimization algorithms and the original SMA on functions from CEC2014 demonstrate that the enhanced algorithm achieves better stability and accuracy. Finally, the improved algorithm is applied to search for SVM kernel parameters, and experiments on eight UCI datasets further validate the feasibility and practicality of the enhanced algorithm.