文献类型：期刊文献

英文题名：Research on Dragonfly Algorithm Based on Elite Reverse Learning and Cauchy Variation Candidate Positions

作者：Xie, Yuxuan[1]; Dai, Ming[1]

机构：[1] College of Mathematics and Computer Science, Guangdong Ocean University, Guangdong, Zhanjiang, China

年份：2026

起止页码：83

外文期刊名：BDAIA 2025 - 2025 2nd International Conference on Big Data Analytics and Artificial Intelligence Application

收录：EI(收录号：20261020204167)

语种：英文

外文关键词：Artificial intelligence - Dynamics - Iterative methods - Learning algorithms - Learning systems

外文摘要：To address the issues of low search accuracy, slow iteration speed, and weak population diversity in the Dragonfly Algorithm (DA), a novel Dragonfly Algorithm based on Elite Reverse Learning and Cauchy Mutation Candidate Positions (RNCDA) is proposed. First, an elite reverse learning strategy is utilized to initialize the population, enhancing population diversity to improve the quality of initial solutions. Second, the linear weight variation is replaced with nonlinear dynamic weight variation, balancing the algorithm's global search capability and local exploitation ability, thus improving the optimization accuracy and convergence speed of the algorithm. Finally, candidate positions are generated using Cauchy mutation based on the current solution's position, and better solutions are selected by comparing the fitness of the candidate positions, enhancing the search capability of the solutions. Experiments on 10 typical benchmark functions demonstrate the effectiveness of RNCDA, comparing it with the original Dragonfly Algorithm, classical Particle Swarm Optimization, Wolf Pack Algorithm, and Ant Lion Optimization. The experimental results indicate that both the Wolf Pack Algorithm and Ant Lion Algorithm outperform the Dragonfly Algorithm across all test functions, while RNCDA approaches the performance of the Wolf Pack Algorithm and Ant Lion Algorithm in some functions, and outperforms them in the others, overall being superior to both. Compared to the Dragonfly Algorithm, RNCDA exhibits higher search accuracy and faster convergence speed. ? 2025 Copyright held by the owner/author(s).