| paper | d4-journals-mss-2008-0033 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | yes |
The proposed continuous Owen allocation evaluation is statable and fits a plausible high-multiplicity regime. But none of the paper’s numbered results asserts an algorithm, complexity bound, or computational formulation used to compute; the explicit dual-price formula is not a computational result anchored by this paper. The required computational bit therefore fails.
fails bit a — no named computational result to mirror
The candidate evaluates the grand-coalition Owen charges; it does not cover Theorem 3.2’s game-class equivalence or Theorems 4.4 and 4.12’s axiomatic characterizations.
The strongest plausible positive case is a society of many small firms with recurring demand and cost profiles: for example, retailers with one of \(\tau\) standard demand vectors and production, inventory, and backlog cost schedules, where \(\tau\) is much smaller than the number of firms. A type is the full demand-and-cost profile; its mass is its fraction of firms. This is a credible high-multiplicity regime for the paper’s production-inventory setting.
One precise candidate is Continuous Owen Allocation Evaluation. The input is a finite support of types \(r\), rational masses \(\mu_r>0\) summing to \(1\), integer per-firm demand vectors \(d^r\), and nonnegative rational cost schedules \(p^r,h^r,b^r\) over \(T\) periods. Set \(\bar d=\sum_r\mu_r d^r\), and in each period take \(\bar p,\bar h,\bar b\) to be the minimum corresponding cost among types with positive mass. Compute the production-inventory dual prices \(y^*\) for those costs, then return each type’s per-firm charge \(a_r=\sum_{\ell=1}^T d^r_\ell y^*_\ell\). The charges aggregate to the grand-coalition cost: \(\sum_r\mu_r a_r=\sum_{\ell=1}^T\bar d_\ell y^*_\ell\). The paper’s displayed formula for \(y^*\) makes this look tractable, with a direct computation polynomial in \(T,\tau\), and the rational input length. The authors could plausibly recognize the output as their Owen allocation aggregated by type.
That is only an unanchored prospect, though. It covers evaluation of the grand-coalition allocation, not a named computational result of this paper. Extending it to coalition stability would raise real questions: how to define population monotonicity over measurable subpopulations, and what happens when a type’s mass approaches zero, since even a tiny positive mass makes its technology available to the whole coalition. The type-level allocation also co-modifies the model by allowing fractional aggregate demand and production.
The anchor requirement is decisive here: I find no named computational result in the paper asserting that a problem is in P, NP-hard, FPT, or similar. Theorem 3.2, proved here using an extreme-ray result cited from Sprumont (1990), says every nonnegative cost game with a population-monotonic allocation scheme is a production-inventory game. Theorem 4.4, proved here, says efficiency, nonemptiness, positivity, inessentiality, and additivity of players’ demands characterize the Owen point. Theorem 4.12, also proved here, says efficiency, nonemptiness, population monotonicity, and anonymity characterize it. These are structural and axiomatic results, not computational ones. The paper formulates an optimization problem and gives a dual-price formula, but does not make either a named complexity result. So the proposed continuous evaluation problem is the best honest foothold I see, but it cannot serve as a valid per-anchor mirror under the stated criteria.
The strongest case against a mirror is that the paper supplies no computational result to mirror. Theorem 3.2 identifies two classes of finite-player games. Theorems 4.4 and 4.12 characterize the Owen point by axioms. Making the population continuous would change the objects in those statements, but would not turn them into complexity questions.
Theorem 3.2’s best continuous counterpart would compare production-inventory games with games admitting a population-monotonic allocation scheme over measurable subpopulations. That could be an interesting mathematical characterization, but it remains a class-equality question. To make it computational, one would need to choose a finite encoding of the continuum games and ask a new membership or recognition question; the theorem itself does not point to one.
The same problem holds for Theorem 4.4’s efficiency, positivity, inessentiality, and demand-additivity axioms, and for Theorem 4.12’s population monotonicity and anonymity. Their natural continuous versions are still axiomatic uniqueness results. In particular, interpreting population monotonicity over every measurable subpopulation does not supply an algorithmic task.
The proponent’s firms-with-standardized-profiles scenario is plausible, and their continuous Owen allocation evaluation is a sensible model. But the paper gives the Owen point as \(D y^*(N)\) and supplies an explicit formula for \(y^*\). With finitely many types, the proposed calculation is to find the minimum costs among active types, compute those dual prices, and apply them to each type’s demand. That is a straightforward extension of the formula, not a computational question asserted or motivated by one of the paper’s results. Computing coalition costs has the same shape. More ambitious tasks would require adding a new objective or a new input representation.
So the negative case is strong on fit with ChoCo’s computational programme, but it is not a proof that no useful adjacent project exists. The paper’s axiomatic results could inspire one. They do not, on their own, provide the computational anchor needed to make this a strong continuous-mirror candidate.
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.