论文部分内容阅读
群搜索优化(Group Search Optimizer,GSO)算法是一种新的群集智能优化算法,适宜于解决多模态高维问题。对GSO算法进行了一些改进,简化了计算过程,提高了优化性能。主要在两个方面进行改进,一是在迭代过程中,控制允许变异的维的数量,使之从多到少变化,以提高收敛速度。二是用随机数来确定生成个体新位置所用的一组随机值的正负数比例,避免正负数比例趋于固定,增加随机性。经过6个常用测试函数测试及与其他文献结果对比后可知,在低维情况下,此算法与GA、EP、ES、PSO、GSO算法相比有较好的整体收敛性能,高维时,此算法与GA、PSO、GSO比较,收敛性能有明显优势。
Group Search Optimizer (GSO) algorithm is a new cluster intelligent optimization algorithm, which is suitable for solving multi-modal high-dimensional problems. Some improvements have been made to the GSO algorithm, which simplifies the calculation process and improves the optimization performance. Mainly in two aspects to be improved, first, in the iterative process, the number of control allows variation in the number of dimensions, so that from more to less change in order to improve the convergence rate. Second, use random numbers to determine the positive and negative proportion of a random set of values used to generate a new position of an individual to avoid positive and negative proportion tends to be fixed, increasing randomness. After testing six commonly used test functions and comparing with other literature results, we can see that this algorithm has better overall convergence performance than GA, EP, ES, PSO and GSO algorithms in the low dimension case. In the high dimension, Compared with GA, PSO and GSO, the convergence performance has obvious advantages.