Hybrid many-objective cuckoo search algorithm (HMaOCS) is a newly proposed method for Many-objective optimization problems (MaOPs), and has achieved promising performance. However, Lévy and Gaussian distributions used in global search manner of HMaOCS is originally proposed for optimization problems with one objective, and they are not suitable for MaOPs as illustrated in this paper. To further exploit the potential of HMaOCS, this paper investigates four different probability distributions and their six corresponding combinations. Comparison results illustrate that the combination of Lévy and Exponential distributions is able to greatly improve HMaOCS. On the basis of comparison results and analysis on both DTLZ and WFG test suites with 2, 3, 4, 6, 8 and 10 objectives, it can be concluded that HMaOCS with Lévy and Exponential distributions exhibits better performance compared with most advanced algorithms.

中文翻译：

---

具有Lévy和指数分布的混合多目标布谷鸟搜索算法

混合多目标布谷鸟搜索算法（HMaOCS）是针对多目标优化问题（MaOPs）提出的一种新方法，并取得了令人鼓舞的性能。但是，最初提出将HMaOCS的全局搜索方式中使用的Lévy和Gaussian分布用于具有一个目标的优化问题，并且它们不适用于本文中所描述的MaOP。为了进一步挖掘HMaOCS的潜力，本文研究了四种不同的概率分布及其六个对应的组合。比较结果说明，Lévy的组合指数分布可以大大改善HMaOCS。根据比较结果和对具有2、3、4、6、8和10个目标的DTLZ和WFG测试套件的分析，可以得出结论，与大多数高级算法相比，具有Lévy和指数分布的HMaOCS表现出更好的性能。