Weakly-Popular and Super-Popular Matchings with Ties and Their Connection to Stable Matchings

Gergely Csáji

2024-06-26

Abstract:In this paper, we study a slightly different definition of popularity in bipartite graphs $G=(U,W,E)$ with two-sided preferences, when ties are present in the preference lists. This is motivated by the observation that if an agent $u$ is indifferent between his original partner $w$ in matching $M$ and his new partner $w'\ne w$ in matching $N$, then he may probably still prefer to stay with his original partner, as change requires effort, so he votes for $M$ in this case, instead of being indifferent. We show that this alternative definition of popularity, which we call weak-popularity allows us to guarantee the existence of such a matching and also to find a weakly-popular matching in polynomial-time that has size at least $\frac{3}{4}$ the size of the maximum weakly popular matching. We also show that this matching is at least $\frac{4}{5}$ times the size of the maximum (weakly) stable matching, so may provide a more desirable solution than the current best (and tight under certain assumptions) $\frac{2}{3}$-approximation for such a stable matching. We also show that unfortunately, finding a maximum size weakly popular matching is NP-hard, even with one-sided ties and that assuming some complexity theoretic assumptions, the $\frac{3}{4}$-approximation bound is tight. Then, we study a more general model than weak-popularity, where for each edge, we can specify independently for both endpoints the size of improvement the endpoint needs to vote in favor of a new matching $N$. We show that even in this more general model, a so-called $\gamma$-popular matching always exists and that the same positive results still hold. Finally, we define an other, stronger variant of popularity, called super-popularity, where even a weak improvement is enough to vote in favor of a new matching. We show that for this case, even the existence problem is NP-hard.

Computer Science and Game Theory

What problem does this paper attempt to address?

This paper attempts to solve the problem of how to define and find popular matchings in the bipartite graph matching problem with ties. Specifically:

1. **Weakly - Popular Matchings**:
- The author proposes a different definition of popularity, called weak - popularity. Under this definition, if an agent does not get a better partner in the new matching and needs to change partners, he prefers to stay as he is.
- Research shows that this new definition guarantees the existence of weakly - popular matchings, and a matching with a size of at least $\frac{3}{4}$ of the maximum weakly - popular matching can be found in polynomial time.
- In addition, the size of this matching is at least $\frac{4}{5}$ of the size of the maximum stable matching, which is better than the current best $\frac{2}{3}$ - approximate solution.

2. **Super - Popular Matchings**:
- The author also defines a stronger variant of popularity, called super - popularity. In this case, even a tiny improvement is enough to vote for a new matching.
- However, research has found that even deciding whether such a matching exists is NP - hard.

3. **γ - Popular Matchings**:
- The author further studies a more general model, in which the amount of improvement required for each endpoint of each edge to vote for a new matching can be specified.
- The results show that in this more general model, γ - popular matchings always exist, and the same positive results still hold.

### Main Contributions
- **Weakly - Popular Matchings** and **γ - Popular Matchings** always exist, and matchings with approximately maximum size can be found in linear time.
- A $\frac{3}{4}$ - approximate algorithm is proposed, and the size of the output matching is at least $\frac{4}{5}$ of the size of the maximum stable matching.
- The edge - replication technique is used to process and analyze the problem, and this method is very effective in stable - matching - related problems.
- It is proved that in the case of ties, finding the maximum weakly - popular matching is NP - hard, even if the ties only appear on one side of the market.

### Mathematical Formula Summary
- The size of the weakly - popular matching is at least $\frac{3}{4}$ of the maximum weakly - popular matching:
\[
|M| \geq \frac{3}{4} |N|
\]
- The size of the weakly - popular matching is at least $\frac{4}{5}$ of the maximum stable matching:
\[
|M| \geq \frac{4}{5} |N|
\]
- For any γ - minimum stable matching $N$, we have:
\[
|M| \geq \frac{4}{5} |N|
\]

Through these studies, the author not only extends the definition of popular matchings but also provides effective algorithms and theoretical analysis, thus providing solutions to the matching problem with ties.