基于动态反向学习和黄金正弦策略的麻雀算法研究

Research on Sparrow Search Algorithm Based on Dynamic Opposite Learning and Golden Sine Strategy

摘要: 针对麻雀算法存在的收敛速度慢，寻优精度不高，易陷入局部最优的问题，提出一种基于动态反向学习和黄金正弦策略改进的麻雀算法。引入动态反向学习策略生成初始种群，从而扩大搜索空间，增加多样性。此外，引入黄金正弦因子来提高种群个体质量，以改善算法性能，并采用多项式变异扰动与双面镜反射理论边界优化策略，以增强跳出局部最优解的能力。为了验证新算法优化性能和普适性，选用10个标准测试函数与其他优化算法进行比较。实验结果表明，相对于传统的麻雀算法、粒子群算法以及灰狼算法，本文提出的改进算法具有更好的收敛精度和更快的收敛速度，在性能上表现出一定的优越性。

Abstract: In this paper, an improved Sparrow Search Algorithm (DGSSA) based on Dynamic Opposite Learning and Golden Sine strategy is presented to address the limitations of the standard Spar-row Search Algorithm (SSA), with the aim of improving convergence speed, optimization precision, and overcoming local optimization issues. The improved method includes the introduction of dynamic opposite learning strategy to generate the initial population, thereby expanding the search space and increasing the diversity. In addition, the Golden-Sine factor is introduced to enhance individual mass within the population, thereby improving the algorithm’s performance. Moreover, the polynomial variation and double-faced mirror reflection theory boundary optimization strategy are introduced to improve the capability of escaping local optimal solutions. To verify that the DGSSA optimizes performance and universality, this paper chooses 10 standard test function comparing with other optimization algorithms. The experimental results show that compared with SSA, particle swarm algorithm and gray wolf algorithm, the DGSSA algorithm has better convergence accuracy and faster convergence speed, and has presented in this paper exhibits notable performance advantages.