Abstract:

Aiming at the shortcomings of backtracking search optimization algorithm with slow convergence speed and easy to fall into local optimum, an improved backtracking search optimization algorithm based on combined mutation strategy is proposed. In order to improve the diversity of historical populations and expand the search space of the algorithm, the Cauchy population generation strategy is used in the iterative process of the algorithm to generate historical populations using Cauchy distribution scale coefficients. A combination based on chaotic map and Gamma distribution is introduced. The mutation strategy mutates the poor individuals to generate better quality individuals under certain probability. The out-of-bounds processing strategy is adopted for the cross-border individuals in the new population to ensure that the algorithm searches within the predetermined search space. In this paper, eleven standard test functions are selected, and numerical simulations are carried out in low-dimensional and high-dimensional states, and compared with three well-performing algorithms. The results show that the improved algorithm has great advantages in convergence speed and convergence precision.

Key words: improved backtracking search optimization algorithm, Cauchy population generation strategy, combined mutation strategy, out-of-bounds processing strategy, function optimization

摘要：

针对回溯搜索优化算法收敛速度慢和易陷入局部最优的缺陷，提出了一种基于组合变异策略的改进回溯搜索优化算法。为了提高历史种群的多样性并扩大算法的搜索空间，在算法迭代过程中采用柯西种群生成策略，利用柯西分布尺度系数生成历史种群；引入基于混沌映射和伽玛分布的组合变异策略，在一定概率下对较差个体进行变异生成质量较好的个体；对新种群中越界个体采用越界处理策略，确保算法在预定的搜索空间内搜索。选取了11个标准测试函数，在低维和高维状态下进行数值仿真，并与3种表现良好的算法进行比较，结果表明该改进算法在收敛速度和收敛精度上有很大优势。

关键词: 改进回溯搜索优化算法, 柯西种群生成策略, 组合变异策略, 越界处理策略, 函数优化