Flipping Matchings is Hard
Carla Binucci, Fabrizio Montecchiani, Daniel Perz, Alessandra Tappini
TL;DR
The paper addresses the problem of transforming one plane perfect matching into another on a fixed point set using a sequence of edge flips, formalized as FlippingBetweenMatchings. It introduces a gadget-based NP-hardness reduction from Planar Vertex Cover, establishing that a yes instance with budget $k$ exists exactly when a vertex cover of size at most $c$ exists, via a construction with vertex gadgets, edge gadgets, blockers, separators, and a budget $k=2c+5|E|$. The core insight is that activating a subset of vertex gadgets corresponding to a vertex cover suffices to reconfigure $ ext{M}_1$ into $ ext{M}_2$ in the specified number of flips, and any feasible sequence encodes a vertex cover of size $oxed{c}$. The result holds for integer-coordinate point sets with polynomial-area bounding box, highlighting a fundamental hardness barrier for reconfiguring plane perfect matchings and prompting further work on general-position instances and flip-graph connectivity.
Abstract
Given a point set $\mathcal{P}$ and a plane perfect matching $\mathcal{M}$ on $\mathcal{P}$, a flip is an operation that replaces two edges of $\mathcal{M}$ such that another plane perfect matching on $\mathcal{P}$ is obtained. Given two plane perfect matchings on $\mathcal{P}$, we show that it is NP-hard to minimize the number of flips that are needed to transform one matching into the other.
