最小Manhattan网络的算法复杂性研究.docVIP

最小Manhattan网络的算法复杂性研究 计算机学院 郭泽宇 指导教师 孙贺 摘要：本文研究了最小Manhattan网络的复杂性和近似算法。通过给出3-可满足问题的基于部件的归约，本文证明了最小Manhattan网络是强NP-完全的，从而解决了这一长达十年的未知问题。此外，本文描述了一个时间复杂度为O(nlogn)的2-近似算法。该算法在运行时间和近似度上都达到目前最优。 关键词：NP完全性 近似算法 计算几何 Abstract: In this paper, we have studied the complexity of Minimum Manhattan Network and its approximations. Using a gadget-based reduction from 3-SAT, we showed that Minimum Manhattan Network is strongly NP-complete, which was open for ten years. Moreover, we have described a 2-approximation algorithm with time complexity O(nlogn), which has the best time complexity and approximation ra