基于均匀设计的免疫克隆多峰函数优化

摘要

In order to get all the best solutions of multi-model function, an immune clonal algorithm with uniform design for multi-modal function optimization problem is proposed. Uniform design is used to initialize the population so that the initialization population is uniform and with diversity. In addition, Lamarck learning is used as a local search strategy to enhance the search ability for the optimal solution. Uniform design is used to convert parameter establishment problem into the experimental design of multi-factor and multi-level, it can reduce the test times of simulation experiments. The experi-mental results show that the algorithm has better optimization ability.%为了尽可能多地求得多峰函数的全部最优解，提出了基于均匀设计的免疫克隆多峰函数优化。算法采用均匀设计初始化种群，保证初始抗体群体分布的均匀性和多样性。采用Larmark学习策略对群体进行局部搜索，以增强算法的收敛速度和搜索精度。在免疫克隆参数设置上，将参数设定问题描述成多因素多水平的均匀设计问题，减少设置参数所需的实验次数。实验结果表明，该算法寻优能力较强。