Pattern Matching in Doubling Spaces

TL;DR

This paper studies the computational complexity of pattern matching in metric spaces with bounded doubling dimension, providing hardness results and near-linear time approximation algorithms for fixed pattern sizes.

Contribution

It establishes NP-hardness and W[1]-hardness for the problem, and introduces near-linear time approximation algorithms for fixed pattern sizes in doubling spaces.

Findings

NP-hardness and W[1]-hardness results for the problem

Near-linear time approximation algorithms for fixed pattern size

Extension of results to minimum distortion problem

Abstract

We consider the problem of matching a metric space $(X,d_X)$ of size $k$ with a subspace of a metric space $(Y,d_Y)$ of size $n \geq k$, assuming that these two spaces have constant doubling dimension $\delta$. More precisely, given an input parameter $\rho \geq 1$, the $\rho$-distortion problem is to find a one-to-one mapping from $X$ to $Y$ that distorts distances by a factor at most $\rho$. We first show by a reduction from $k$-clique that, in doubling dimension $\log_2 3$, this problem is NP-hard and W[1]-hard. Then we provide a near-linear time approximation algorithm for fixed $k$: Given an approximation ratio $0<\varepsilon\leq 1$, and a positive instance of the $\rho$-distortion problem, our algorithm returns a solution to the $(1+\varepsilon)\rho$-distortion problem in time $(\rho/\varepsilon)^{O(1)}n \log n$. We also show how to extend these results to the minimum distortion problem in doubling spaces: We prove the same hardness results, and for fixed $k$, we give a $(1+\varepsilon)$-approximation algorithm running in time $($dist$(X,Y)/\varepsilon)^{O(1)}n^2\log n$, where dist$(X,Y)$ denotes the minimum distortion between $X$ and $Y$.