Abstract:

In order to further improve the population diversity, global optimization ability and search efficiency of immune clonal algorithm in high-dimensional object optimization, a clonal selection algorithm combining cloud model and reverse learning（CRCSA） is proposed. The concept of cloud model is introduced, and the forward cloud generator is used to generate cloud mutation factors and mutate the cloned population. The reverse learning strategy is used to find the reverse solution of the population before and after mutation, and then realize the selection of population antibody. Finally, the convergence of the algorithm is proved by Markov chain theory. The test results of six groups of high-dimensional functions show that, compared with Differential Genetic Algorithm（DGA）, Immune Genetic Algorithm（IGA） and Adaptive Chaotic clonal Selection Algorithm（ACSA）, the proposed algorithm achieves 100% optimization, and the minimum iterative algebra, convergence algebra and iterative algebra standard deviation are reduced by 33.7%, 19.8% and 29.1%, respectively, which verifies its strong optimization ability, high search efficiency and good stability.

Key words: cloud model, reverse learning, clone selection, cloud mutation, high-dimensional function optimization

摘要：

为了进一步改善高维对象优化时免疫克隆算法的种群多样性，提高算法全局优化能力和搜索效率，提出了融合云模型和反向学习的克隆选择算法。引入云模型概念，使用正向云发生器产生云变异因子，进而对克隆后种群进行变异；利用反向学习策略，对变异前后的种群求反向解，进而实现种群抗体选择；通过马尔可夫链理论证明了算法收敛性。六组高维函数测试结果表明，与差分遗传算法、免疫遗传算法和自适应混沌克隆选择算法相比，该算法实现了100%的寻优，且最小收敛代数、平均收敛代数及迭代代数标准差分别平均减少33.7%、19.8%、29.1%，从而验证了其强优化能力、高搜索效率和好稳定性。

关键词: 云模型, 反向学习, 克隆选择, 云变异, 高维函数优化