一种基于差分策略的群搜索优化算法

发布时间：2019-05-06 08:14

【摘要】：针对群搜索优化(Group Search Optimizer,GSO)算法易陷入局部最优、收敛速度较慢、收敛精度较低等问题,提出一种基于差分策略的群搜索优化(Differential Ranking-based Group Search Optimizer,DRGSO)算法。主要进行两方面改进:1)按照适应度值的大小对种群进行排序,适当增加发现者的数目,使种群能够获得更好的启发式信息,加快了算法的收敛速度,有效地避免了算法陷入局部最优;2)在发现者搜索过程中,引入4种不同的差分变异策略,提高了算法的收敛精度,增强了算法的群体多样性在。11组国际标准测试函数上的实验测试结果显示,与GA,GSO,PSO算法相比,DRGSO算法具有较强的全局搜索能力以及局部资源勘探能力,算法整体收敛性能明显提高。
[Abstract]:A group search optimization (Differential Ranking-based Group Search Optimizer,DRGSO) algorithm based on differential strategy is proposed to solve the problems of group search optimization (Group Search Optimizer,GSO) algorithm which is easy to fall into local optimization, slow convergence rate and low convergence precision. The main improvements are as follows: 1) according to the size of the adaptive value, the population is ordered, and the number of discoverers is increased appropriately, so that the population can obtain better heuristic information, and the convergence speed of the algorithm is speeded up. The algorithm is effectively avoided from falling into local optimization. 2) in the process of discovery search, four different differential mutation strategies are introduced to improve the convergence accuracy of the algorithm, and the experimental results of the diversity of the algorithm on .11 international standard test functions show that it is in agreement with GA,GSO,. Compared with PSO algorithm, DRGSO algorithm has strong global search ability and local resource exploration ability, and the global convergence performance of the algorithm is obviously improved.
【作者单位】：天津科技大学计算机科学与信息工程学院;
【基金】：天津市高等学校科技发展基金计划项目(20140803) 天津科技大学青年教师创新基金(2014CXLG30) 国家自然基金面上项目(61272509) 国家自然科学基金青年基金(61402332)资助
【分类号】：TP18

【相似文献】