计算机科学与探索

• 学术研究 •    

融合排序弹性碰撞的改进麻雀搜索算法

王子恺,黄学雨,朱东林,郭伟   

  1. 1.江西理工大学 信息工程学院,江西 赣州 341000
    2.江西理工大学 软件工程学院,南昌 330013
    3.南昌市虚拟数字工程与文化传播重点实验室,南昌 330013

Improved Sparrow Search Algorithm Combining Ranking-Based Elastic Collision

WANG Zikai, HUANG Xueyu, ZHU Donglin, GUO Wei   

  1. 1.School of Information Engineering, Jiangxi University of Science and Technology, Ganzhou, Jiangxi 341000, China
    2.School of Software Engineering, Jiangxi University of Science and Technology, Nanchang 330013, China
    3.Nanchang Key laboratory of Virtual Digital Factory and Cultural Communications, Nanchang 330013, China

摘要: 为了改善麻雀搜索算法(Sparrow Search Algorithm, SSA)种群初始化结果不充分导致多样性丧失,勘探和开采过程中易受到个别位置信息干扰影响寻优精度等缺点,提出了融合排序弹性碰撞的新型麻雀搜索算法,简称为XSSA。首先,采用改进的无限折叠迭代混沌映射(Iterative Chaotic Map with Infinite Collapses, ICMIC)初始化种群,提升了初始种群分布的分散程度;其次,使用高斯随机游走策略平衡算法的探勘和开发能力;此外,在发现者更新后对所有个体执行排序弹性碰撞策略,避免算法过早地收敛到局部极值;最后,根据不同阶段的寻优特点制定多策略边界处理机制,保留住种群数量,避免多样性的丧失。同时,结合重要的位置信息对超出边界的个体进行位置再更新,使得处理后的位置更加合理,为接下来的迭代搜索提供质量保证。通过对12个基准函数进行仿真实验,并画出收敛精度图直观展示算法性能。借助各策略的贡献测试、Wilcoxon秩和检验、Friedman检验的综合排名等证明了XSSA的有效性、独特性和具有较好的寻优性能。

关键词: 麻雀搜索算法, 无限折叠迭代混沌映射(ICMIC), 高斯随机游走, 排序弹性碰撞, 多策略边界处理

Abstract: In order to improve the Sparrow Search Algorithm (SSA) population initialization results are insufficient, resulting in loss of diversity, and is easy to be interfered by individual location information during the exploration and mining process, which affects the optimization accuracy. An improved sparrow search algorithm based on fusion ranking elastic collision, referred to as XSSA, is proposed to solve the problem easily disturbed by individual location information. Firstly, an improved Iterative Chaotic Map with Infinite Collapses (ICMIC) improves the dispersion degree of the initial population distribution. Then, the Gaussian random walk strategy is used to balance the exploration and development capabilities of the algorithm. In addition, all individuals are sorted elastically collided after the discoverer is updated, which avoids premature convergence of the algorithm to the local extreme value. Finally, a multi-strategy boundary processing mechanism is formulated according to the characteristics of optimization at different stages to retain the population and avoid the loss of diversity. At the same time, the position of individuals beyond the boundary is re-updated in combination with important position information, so that the processed position is more reasonable and provides quality guarantee for the next iterative search. Through the simulation experiment of 12 benchmark functions, and the convergence accuracy graph demonstrates the performance of the algorithm. By means of contribution tests for each strategy, Wilcoxon rank sum test, and the comprehensive ranking of Friedman test, the effectiveness, uniqueness and superior performance of XSSA are proved.

Key words: Sparrow search algorithm, ICMIC map, Gauss walk learning, Ranking-based elastic collision, Muti-strategy boundary processing mechanism