Core Concepts
The article establishes tight bounds on the parameterized complexity of the Metric Dimension and Geodetic Set problems with respect to the vertex cover number of the input graph. It provides FPT algorithms and matching lower bounds based on the Exponential Time Hypothesis.
Abstract
The paper studies the parameterized complexity of two metric-based graph problems, Metric Dimension and Geodetic Set, with respect to the vertex cover number (vc) of the input graph.
Key highlights:
- The authors provide FPT algorithms for both problems running in time 2^O(vc^2) * n^O(1), and kernelization algorithms that output kernels with 2^O(vc) vertices.
- They prove that, unless the Exponential Time Hypothesis (ETH) fails, these problems do not admit FPT algorithms running in time 2^o(vc^2) * n^O(1), nor kernelization algorithms that reduce the solution size and output kernels with 2^o(vc) vertices, even on graphs of bounded diameter.
- These results constitute a rare set of tight lower bounds in parameterized complexity, with only a few other problems known to exhibit such tight bounds.
- The authors develop novel techniques involving set identifying gadgets and vertex selector gadgets to establish the lower bounds.
- The results provide a clear boundary between parameters yielding single-exponential and double-exponential running times for these two problems.