In many-objective optimization problems (MaOPs), the decomposition-based algorithms are widely used since they have promising performances in maintaining the diversity of solutions. However, few studies have been reported on how to utilize relationships between subproblems to promote global convergence. To fill this gap, we develop an automatic estimation mechanism based on the modified Ant Colony Algorithm to assist the co-evolution between subproblems, where two species of ants are designed. The working-ants execute local exploitation by recording the information of subproblems. The command-ants control global exploration by adjusting co-evolution between working-ants. Moreover, the automatic estimation mechanism is expanded into three modes to verify the more efficient one, and they are embedded separately in the decomposition-based algorithm to construct the combined algorithms. The proposed algorithms are compared with five state-of-the-art algorithms on multiple test suites. The experimental results show that the proposed algorithms perform comparably or better than all referenced algorithms. Given the better performance of the proposed algorithms, it is evident that the hybrid mechanism may be a potential manner to handle MaOPs.

在多目标优化问题（MaOP）中，基于分解的算法在保持解决方案多样性方面具有令人鼓舞的性能，因此被广泛使用。但是，关于如何利用子问题之间的关系来促进全局收敛的研究很少。为了填补这一空白，我们基于改进的蚁群算法开发了一种自动估计机制，以协助设计两个蚂蚁种类的子问题之间的协同进化。工作人员通过记录子问题的信息来执行本地开发。指挥蚂蚁通过调整工作蚂蚁之间的共同进化来控制全局探索。此外，自动估算机制被扩展为三种模式以验证更有效的一种：然后将它们分别嵌入基于分解的算法中，以构造组合算法。将所提出的算法与多个测试套件上的五个最新算法进行了比较。实验结果表明，所提出的算法性能优于或优于所有参考算法。考虑到所提出算法的更好性能，显然混合机制可能是处理MaOP的潜在方式。