一种双精英引导的群体智能优化算法OA
A Dual-Elite-Guided Swarm Intelligence Optimization Algorithm
针对现有群体智能算法在连续优化中搜索不稳定且易早熟的问题,以及单精英引导在复杂多峰环境下鲁棒性不足的局限,本文提出双精英引导的群体智能优化算法(DEGSIO).该算法同步维护历史最优与次优个体,基于二者协同设计更新规则,引入线性递减随机控制参数以平衡全局探索与局部开发.在 8 个基准函数上与 HHO、PSO、GWO、WOA 及 DE 进行对比,并采用 Friedman 检验.实验表明,DEGSIO 在多种函数上有效避免早熟,在精度、速度和稳定性上具有竞争力,整体性能与HHO统计相当,验证了双精英引导策略的有效性.
To address the unstable search and premature convergence of existing swarm intelligence algorithms in continuous optimization,as well as the insufficient robustness of single-elite guidance in complex multimodal environments,this paper proposes a Dual-Elite-Guided Swarm Intelligence Optimization(DEGSIO)algorithm.The algorithm simultaneously maintains the historical best and second-best individuals,designs an individual update rule based on their collaborative guidance,and introduces a linearly decreasing random control parameter to balance global exploration and local exploitation.The algorithm is compared with HHO,PSO,GWO,WOA,and DE on eight benchmark functions,and the Friedman test is employed.The experimental results show that DEGSIO effectively avoids premature convergence on multiple functions,is competitive in terms of accuracy,convergence speed,and stability,and achieves overall performance statistically comparable to that of HHO,which verifies the effectiveness of the dual-elite guidance strategy.
郑水飞;聂斌;章祉瑶;李想
江西中医药大学,江西 南昌 330004江西中医药大学,江西 南昌 330004江西中医药大学,江西 南昌 330004江西中医药大学,江西 南昌 330004
信息技术与安全科学
群体智能优化双精英策略个体更新规则连续优化
swarm intelligence optimizationdual-elite strategyindividual update rulecontinuous optimization
《现代信息科技》 2026 (15)
71-76,6
国家自然科学基金(82260849)江西中医药大学校级科技创新团队发展计划(CXTD22015)
评论