基于Sobol序列和随机差分变异的改进人工兔优化算法

An Improved Artificial Rabbit Optimization Algorithm Based on Sobol Sequences and Stochastic Difference Variants

摘要: 针对人工兔优化算法在收敛精度不足和易陷入局部最优的不足，提出了一种基于Sobol序列和随机差分变异的改进人工兔优化算法(SDARO)。算法用Sobol序列替代简单的随机初始化，使初始种群能够更均匀地分布在搜索空间中，从而增强群体的全局搜索能力，以提高算法的收敛速度和精度。在算法的位置更新过程中，引入改进的随机差分变异策略，对部分个体进行变异操作，以帮助群体跳出局部最优。23个基准函数的数值结果表明，与其他优化算法相比，所提出的算法SDARO在收敛精度、收敛速度和伸缩性方面均表现较好。三个工程设计问题的实验结果也进一步验证了算法SDARO能够获得较对比算法更优的设计方案。

Abstract: An improved artificial rabbit optimization algorithm (SDARO) based on Sobol sequences and stochastic differential variants is proposed to overcome the shortcomings of the artificial rabbit optimization algorithm such as low convergence accuracy and easy falling into local optimum. The algorithm replaces simple random initialization with Sobol sequences to make the initial population distribute more evenly in the search space, which enhances the global search capability of the population and improves the convergence speed and accuracy of the algorithm. In the position updating process of the algorithm, a stochastic differential mutation strategy is improved and used to enable some individuals to perform the mutation operation in order to help the population to escape from the local optimum. Numerical results on 23 benchmark functions show that, compared to other optimization algorithms, the proposed algorithm SDARO performs better in convergence accuracy, convergence speed and scalability. The experimental results on three engineering design problems also show that the proposed algorithm SDARO can obtain better design solutions than the comparison algorithms.