English

Approximating MIS over equilateral $B_1$-VPG graphs

Data Structures and Algorithms 2019-12-18 v1

Abstract

We present an approximation algorithm for the maximum independent set (MIS) problem over the class of equilateral B1B_1-VPG graphs. These are intersection graphs of LL-shaped planar objects % (and their rotations by multiples of 90o90^o) with both arms of each object being equal. We obtain a 36(log2d)36(\log 2d)-approximate algorithm running in O(n(logn)2)O(n(\log n)^2) time for this problem, where dd is the ratio dmax/dmind_{max}/d_{min} and dmaxd_{max} and dmind_{min} denote respectively the maximum and minimum length of any arm in the input equilateral LL-representation of the graph. In particular, we obtain O(1)O(1)-factor approximation of MIS for B1B_1-VPG -graphs for which the ratio dd is bounded by a constant. % formed by unit length LL-shapes. In fact, algorithm can be generalized to an O(n(logn)2)O(n(\log n)^2) time and a 36(log2dx)(log2dy)36(\log 2d_x)(\log 2d_y)-approximate MIS algorithm over arbitrary B1B_1-VPG graphs. Here, dxd_x and dyd_y denote respectively the analogues of dd when restricted to only horizontal and vertical arms of members of the input. This is an improvement over the previously best nϵn^\epsilon-approximate algorithm \cite{FoxP} (for some fixed ϵ>0\epsilon>0), unless the ratio dd is exponentially large in nn. In particular, O(1)O(1)-approximation of MIS is achieved for graphs with max{dx,dy}=O(1)\max\{d_x,d_y\}=O(1).

Keywords

Cite

@article{arxiv.1912.07957,
  title  = {Approximating MIS over equilateral $B_1$-VPG graphs},
  author = {Abhiruk Lahiri and Joydeep Mukherjee and C. R. Subramanian},
  journal= {arXiv preprint arXiv:1912.07957},
  year   = {2019}
}
R2 v1 2026-06-23T12:48:20.407Z