一种正交对角化的磷虾群算法

A Novel KH Algorithm Based on Orthogonal Diagonalization

针对磷虾群算法在处理高维复杂问题时,容易陷入局部最优、收敛速度慢、求解精度低等不足,提出了一种正交对角化的磷虾群算法。该算法根据适应度值排序分组磷虾群,通过正交对角化策略,吸收优质磷虾的经验,引导磷虾群寻优,同时兼顾“平凡”磷虾按常规方法寻优,从而提升了全体磷虾寻优能力,并保持了种群多样性。对标准测试函数的实验表明了该算法在全局寻优的精度和时间效率上都具有较明显的优势。

To overcome the shortcomings of krill herd algorithm,such as local optimization,slow convergence speed and low solution accuracy,an orthogonal diagonalization krill herd algorithm was proposed.This algorithm sorted and grouped krill according to the fitness value,and absorbed the experience of high quality krill through orthogonal diagonalization strategy to guide the herd to the best position.Meanwhile,the“ordinary”krill were optimized by the conventional method,which improved the optimization ability of all krill and maintains the population diversity.Experiments on standard test functions showed that the algorithm had obvious advantages in global optimization accuracy and time efficiency.