Table of Contents
Fetching ...

A poset representation for stable contracts in a two-sided market generated by integer choice functions

Alexander V. Karzanov

TL;DR

This paper analyzes SGMM, an integer-valued version of the Alkan–Gale stable two-sided market model. It develops a poset-based representation of the lattice of stable g-matchings by extracting rotations from an active graph and organizing them into a labeled-rotation poset, enabling an explicit isomorphism between stable g-matchings and closed functions on the poset. The authors show the poset size is at most $O(b^{ m max}|E|)$ and provide procedures to construct it in pseudo-polynomial time; imposing a gapless condition on choice functions further yields polynomial-size representations and weakly polynomial-time construction. They also outline a min-cut reduction for computing minimum-cost stable g-matchings and demonstrate that, without the gapless condition, there exist graphs with large posets, highlighting intrinsic complexity in SGMM. Overall, the work provides a concrete, constructive framework to represent and compute stable g-matchings via a distributive-lattice via rotations, with clear implications for algorithm design in two-sided markets with integer data.

Abstract

Generalizing a variety of earlier problems on stable contracts in two-sided markets, Alkan and Gale introduced in 2003 a general stability model on a bipartite graph $G=(V,E)$ in which the vertices are interpreted as ``agents'', and the edges as possible ``contract'' between pairs of ``agents''. The edges are endowed with nonnegative capacities $b$ giving upper bounds on ``contract intensities'', and the preferencies of each ``agent'' $v\in V$ depend on a \emph{choice function} (CFs) that acts on the set of ``contracts'' involving $v$, obeying three well motivated axioms of \it{consistence}, \it{substitutability} and \it{cardinal monotonicity}. In their model, the capacities and choice functions can take reals or discrete values and, extending well-known earlier results on particular cases, they proved that systems of \it{stable} contracts always exist and, moreover, their set $\cal S$ constitutes a distributive lattice under a natural comparison relation $\prec$. In this paper, we study Alkan--Gale's model when all capacities and choice functions take integer values. We characterize the set of rotations -- augmenting cycles linking neighboring stable assignments in the lattice $(\cal S,\prec)$, and construct a weighted poset in which the lattice the closed functions is isomorphic to $(\cal S,\prec)$, thus obtaining an explicit representation for the latter. We show that in general the size of the poset is at most $b^{\rm max}|E|$, where $b^{\rm max}$ is the maximal capacity, and the poset can be constructed in pseudo polynomial time. Then we explain that by imposing an additional condition on CFs, the size of the poset becomes polynomial in $|V|$, and the total time reduces to a polynomial in $|V|,\log b^{\rm max}$.

A poset representation for stable contracts in a two-sided market generated by integer choice functions

TL;DR

This paper analyzes SGMM, an integer-valued version of the Alkan–Gale stable two-sided market model. It develops a poset-based representation of the lattice of stable g-matchings by extracting rotations from an active graph and organizing them into a labeled-rotation poset, enabling an explicit isomorphism between stable g-matchings and closed functions on the poset. The authors show the poset size is at most and provide procedures to construct it in pseudo-polynomial time; imposing a gapless condition on choice functions further yields polynomial-size representations and weakly polynomial-time construction. They also outline a min-cut reduction for computing minimum-cost stable g-matchings and demonstrate that, without the gapless condition, there exist graphs with large posets, highlighting intrinsic complexity in SGMM. Overall, the work provides a concrete, constructive framework to represent and compute stable g-matchings via a distributive-lattice via rotations, with clear implications for algorithm design in two-sided markets with integer data.

Abstract

Generalizing a variety of earlier problems on stable contracts in two-sided markets, Alkan and Gale introduced in 2003 a general stability model on a bipartite graph in which the vertices are interpreted as ``agents'', and the edges as possible ``contract'' between pairs of ``agents''. The edges are endowed with nonnegative capacities giving upper bounds on ``contract intensities'', and the preferencies of each ``agent'' depend on a \emph{choice function} (CFs) that acts on the set of ``contracts'' involving , obeying three well motivated axioms of \it{consistence}, \it{substitutability} and \it{cardinal monotonicity}. In their model, the capacities and choice functions can take reals or discrete values and, extending well-known earlier results on particular cases, they proved that systems of \it{stable} contracts always exist and, moreover, their set constitutes a distributive lattice under a natural comparison relation . In this paper, we study Alkan--Gale's model when all capacities and choice functions take integer values. We characterize the set of rotations -- augmenting cycles linking neighboring stable assignments in the lattice , and construct a weighted poset in which the lattice the closed functions is isomorphic to , thus obtaining an explicit representation for the latter. We show that in general the size of the poset is at most , where is the maximal capacity, and the poset can be constructed in pseudo polynomial time. Then we explain that by imposing an additional condition on CFs, the size of the poset becomes polynomial in , and the total time reduces to a polynomial in .

Paper Structure

This paper contains 15 sections, 30 theorems, 25 equations.

Key Result

Lemma 3.1

Let $v\in V$, $e\in E_v$, $z\in{\cal A}_v$, and $z(e)<b(e)$. The following are equivalent: (i) $e$ is interesting for $v$ under $z$; (ii) there is $\widetilde{z}\in{\cal A}_v$ such that $\widetilde{z}\succ_v z$ and $\widetilde{z}(e)>z(e)$; (iii) $C_v(z+{\bf 1}^e)\ne z$ (where ${\bf 1}^\bullet$ stand

Theorems & Definitions (30)

  • Lemma 3.1
  • Lemma 3.2
  • Lemma 3.3
  • Proposition 3.4
  • Proposition 3.5
  • Lemma 3.6
  • Lemma 3.7
  • Corollary 3.8
  • Corollary 3.9
  • Corollary 4.1
  • ...and 20 more