| paper | d4-journals-jet-2003-0033 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | unclear |
Theorem 1 is the sole anchor, and its sequence may require infinitely many steps to reach a stable matching. The paper provides no finite input representation, terminating algorithm, or complexity guarantee, and the proponent concedes this gap. The proposed type-based problem is a plausible extension, but it cannot supply the missing computational result in the paper.
fails bit a — no named computational result to mirror
The proposed extension concerns stable-schedule existence only; it leaves polarity, optimality, comparative statics, lattice structure, and the examples untouched.
The strongest positive candidate is the paper’s Theorem 1 (Existence), proved here. It says that in a finite market of firms and workers, with bounded schedules and continuous, consistent, persistent choice functions, a stable matching exists. The proof constructs a generalized Gale–Shapley sequence of offers and counteroffers whose limit is stable; the sequence need not terminate after finitely many steps.
There is a real hiring-market regime behind a population mirror: large recurring cohorts of workers and employers with the same complete choice rules, quotas, and schedule bounds. A type is that full description, and its mass is its fraction of the relevant side of the market. The number of people can be very large while the number of types stays small. Here the society is continuous; divisible work schedules are already part of the paper’s model.
One precise candidate is Type-Homogeneous Continuous Schedule Stability. An instance gives finite worker and firm type sets \(I,J\), positive rational masses \(\alpha_i,\beta_j\) summing to \(1\) on each side, and rational piecewise-affine descriptions of bounded, continuous, consistent, persistent choice functions \(C_i^W\) and \(C_j^F\). For each pair of types, it gives a per-pair schedule bound \(b_{ij}\). The question is to find a rational flow \(x_{ij}\) of schedule-hours between types, or report that none exists, such that
\[ 0\le x_{ij}\le \alpha_i\beta_j b_{ij}. \]
The per-capita schedules are \(u_i(j)=x_{ij}/\alpha_i\) for workers and \(v_j(i)=x_{ij}/\beta_j\) for firms. They must be acceptable, meaning \(C_i^W(u_i)=u_i\) and \(C_j^F(v_j)=v_j\). For every \(i,j\), stability requires either \(u_i\) to be \(j\)-satiated or \(v_j\) to be \(i\)-satiated, using the paper’s definition of satiation. A solution is the flow \(x\) satisfying those conditions.
My best bet for this problem is Class A, conditional on the stated compact descriptions of choice functions: with finitely many types, the paper’s alternating-choice construction and continuous optimization look like plausible algorithmic tools. But Theorem 1 proves neither a finite-time algorithm nor a running-time bound, so tractability is a conjecture for the mirror, not a result inherited from the paper. Useful follow-up questions are whether the flow can be computed exactly in polynomial time, how it relates to finite high-multiplicity markets, and whether stable outcomes must be allowed to vary within a type.
The soft spot is substantial: Theorem 1 is an existence theorem, not a named complexity result. The paper has no named NP-hardness, tractability, or parameterized-complexity result to anchor a strict computational-complexity case. Also, my proposed flow assumes everyone in a type receives the same per-capita schedule. That is plausible for symmetric cohorts, but identical choice functions do not by themselves imply identical assignments. So this is a credible continuous research question suggested by the paper, but under the stated anchor rule it is only a near-miss, not a fully qualifying positive case.
The proposed anchor does not meet ChoCo’s computational-result criterion. Theorem 1 asserts existence for a finite \(F \times W\) market with continuous choice functions. Its proof uses an iteration that may continue indefinitely and takes a limit. The paper gives no finite input representation for those functions, no decision or search problem, and no runtime bound. The later theorems establish structural properties of stable matchings, not computational complexity. Calling Theorem 1 a generalized algorithm does not supply the missing complexity result.
The type-flow proposal is a coherent way to formulate a new computational market problem. Repeated hiring cohorts make a high-multiplicity regime plausible, and I would not reject the proposal on that ground. The difficulty is that the paper defines each agent’s choice over a vector indexed by individually named partners. The proposal instead defines choices over totals for partner types and uses representative schedules \(u_i\) and \(v_j\). That can be a sensible co-modification, but it is an added model and input language. If same-type agents may receive different schedules, their average flow does not record whether each person’s schedule is acceptable or satiated: the paper assumes no convexity that would make those properties survive averaging. Tracking schedule distributions and their matching across types would repair that issue, while making the computational problem depend on further modeling and representation choices.
Even the strongest repaired version therefore supplies a worthwhile question only on its own terms. It does not mirror a computational result in this paper, and Theorem 1 gives no complexity classification for it. The negative case is strongest as a reason to exclude this paper from a programme specifically seeking continuous mirrors of computational results. It is weak as the universal claim that no continuous mirror could be worthwhile: a symmetric, type-based stable-scheduling problem for hiring cohorts remains plausible, even though it would be a new computational study rather than a mirror of the paper’s contribution.
The adversarial triple: the proponent anchors on up to three named results; the opponent sees that case and must defeat every anchor; the judge decides which case convinced it. These are the pipeline’s own outputs, generated by tools/triple_run.py — no human edited them. The paper’s own text is not reproduced here beyond the quoted statement above.