Chaotic dung beetle optimization algorithm based on adaptive t-Distribution
Shutong Li, Jinhong Li
Abstract
The Dung Beetle optimizer (DBO) is a metaheuristic algorithm that creates a set of random individuals within a given search space and combines, moves, and evolves these individuals during the iterative process. Although the DBO improves the convergence speed and accuracy in both global and local search compared to other intelligent optimization algorithms, the use of a random distribution in the initialization process leads to an uneven distribution of individuals in the initial population, and the rigid structure of the dung beetle cannot be adaptively adjusted, resulting in invalid iterations in the later stages. To address these issues, this study proposes an Improved Dung Beetle optimizer (IDBO) based on an adaptive t-distribution strategy. First, the population is initialized using a logistic map to achieve a chaotic state, and then an adaptive t-distribution is introduced to update the positions of the dung beetles. The effectiveness of the IDBO is validated by conducting experiments on 23 standard benchmark functions, and the results show a significant improvement in optimization performance and search accuracy compared to traditional DBO and other improved algorithms. This demonstrates the effectiveness of the IDBO algorithm.