Even $1 \times n$ Edge-Matching and Jigsaw Puzzles are Really Hard

TL;DR

This paper proves that even simple 1 by n edge-matching and jigsaw puzzles are computationally intractable to solve optimally or approximately, establishing their NP-hardness and inapproximability.

Contribution

It provides the first correct proof of inapproximability for 1 by n edge-matching and jigsaw puzzles, including a tight approximation hardness result.

Findings

NP-hardness of optimal puzzle arrangement

NP-hardness of approximate solutions within a factor of 0.9999999851

Existence of a simple 1/2-approximation algorithm

Abstract

We prove the computational intractability of rotating and placing $n$ square tiles into a $1 \times n$ array such that adjacent tiles are compatible--either equal edge colors, as in edge-matching puzzles, or matching tab/pocket shapes, as in jigsaw puzzles. Beyond basic NP-hardness, we prove that it is NP-hard even to approximately maximize the number of placed tiles (allowing blanks), while satisfying the compatibility constraint between nonblank tiles, within a factor of 0.9999999851. (On the other hand, there is an easy $1 \over 2$-approximation.) This is the first (correct) proof of inapproximability for edge-matching and jigsaw puzzles. Along the way, we prove NP-hardness of distinguishing, for a directed graph on $n$ nodes, between having a Hamiltonian path (length $n-1$) and having at most $0.999999284 (n-1)$ edges that form a vertex-disjoint union of paths. We use this gap hardness and gap-preserving reductions to establish similar gap hardness for $1 \times n$ jigsaw and edge-matching puzzles.