| paper | d4-conf-atal-2011-0188 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | yes |
The paper has no named result asserting an algorithm, computational formulation used to compute, or complexity classification. The proposed continuous blocking problem is a plausible research question built from the paper’s OCF model, but it cannot supply the missing computational result of this paper.
fails bit a — no named computational result to mirror
The proposed extension concerns arbitrated-core blocking for Threshold Task Games; it does not cover the nucleolus or the Shapley-value results.
There are no qualifying computational anchors in this paper. It has no numbered theorem, lemma, corollary, or proposition asserting that a problem is in P, NP-hard, FPT, or otherwise classifying its complexity. The conclusion’s informal claim that computing the OCF Shapley value is “generally hard” is not a numbered result. So I can’t honestly give the required anchor-based case.
The closest result is Theorem 4.1, proved here, but it is a structural characterization rather than a complexity result: when \(G\) has the ECS property, an outcome \((CS,x)\) is in the \(A\)-core exactly when, for every player subset \(J\) and every deviation \(CS'\), its current payoff \(p_J(CS,x)\) is at least the arbitration value \(v^*(CS,J,CS')+\sum_l\phi_l(CS,x,J,CS')\). That is a natural basis for a continuous research question, but it does not meet the requested anchor rule. Theorem 5.2 is also not a computational anchor: it proves the arbitrated nucleolus is nonempty when \(v\) has ECS and \(A^*\) is continuous.
The strongest *unanchored* positive lead is Continuous Arbitrated-Core Blocking for Threshold Task Games. Take a society with finite type set \(T\) and rational masses \(\mu_t\) summing to one. A type specifies the agent’s full OCF-relevant attributes: resource capacity, task-specific productivity, and any relevant costs or payoff parameters. This fits, for example, a large fleet of interchangeable compute or sensor units, where \(n\) is large and the number \(\tau=|T|\) of distinct hardware and capability profiles is small.
Use a finite task menu with thresholds \(d_\ell\) and payoffs \(p_\ell\). If coalition \(j\) receives type-resource masses \(c^j_t\), its value is
\[
v(c^j)=\max\left\{p_\ell:\sum_{t\in T}w_{t\ell}c^j_t\ge d_\ell\right\}.
\]
An outcome specifies how each type’s mass allocates its unit resource across at most \(U\) overlapping projects and how project payoffs are distributed. Each project’s total payout must not exceed its task value, and no agent receives payoff from a project they do not support. Represent the outcome by finite atomic measures over per-agent contribution and payoff plans, with rational masses.
The decision problem is: given that society, task menu, outcome, and an explicitly represented rational piecewise-linear arbitrator \(A\), does there exist a positive-mass subpopulation and a feasible deviation that strictly improves the total payoff of almost every member of that subpopulation? A witness is the deviating submeasure, its new resource allocations and coalition formation, and a feasible payoff distribution, including any payoff permitted by \(A\) from coalitions it changes. A “no” answer means the outcome is stable against all such deviations. This asks about the underlying core-blocking problem, not an encoding.
I would initially expect this bounded fragment to be tractable when \(U\) and the task menu are fixed: there are finitely many task choices, and the remaining mass and payoff constraints look linear. The key unresolved step is proving that a blocking deviation always has a polynomial-size representation. With variable \(U\) or task menus, the classification is open; the questions are whether task-choice combinatorics preserve discrete hardness, or whether blocking in the continuum creates new hardness. This proposal covers only the arbitrated-core part of the paper, especially the TTG setting, not its nucleolus or Shapley-value results.
The authors could plausibly recognize this as their problem: it retains overlapping resource contributions, task values, payoff allocation, and arbitration, while changing the population to a distribution over complete agent types. The soft spot is substantial: the paper studies stability and solution concepts, not algorithms, and the finite proof of Theorem 4.1 uses a graph argument on individual players. Extending that argument—or finding a computational replacement—for measurable subpopulations is a real open step. That makes this a credible mirror proposal, but not a case anchored in a named computational result of the paper.
The strongest case against is that this paper gives no named computational result to continuize. Theorem 4.1 characterizes stability by quantifying over every player subset and deviation; it gives no input representation, running time, or complexity classification. Theorem 5.2 proves non-emptiness of the nucleolus under continuity assumptions. Neither is a computational anchor. The paper’s closing suggestion to investigate algorithms for the OCF Shapley value is a research direction, not a result whose continuous mirror can be assessed.
The proposed continuous blocking problem therefore adds substantial choices of its own: a finite type space, a Threshold Task Game, fixed \(U\), an encoding for the arbitrator, and a representation of outcomes and deviations as atomic measures. Its “almost every member” condition also chooses a continuum blocking notion that the finite theorem does not specify. Theorem 5.2 is an especially poor computational anchor: carrying its compactness and continuity argument into a continuum model would be an existence question, outside ChoCo’s computational focus.
There is a real scaling choice behind the fleet example. If each unit of mass keeps an agent’s fixed resource endowment, task thresholds and payoffs must be interpreted at the scale of the normalized population. If instead they retain their finite-agent meaning, a deviation by a small finite group can disappear as its population share tends to zero. The proposed fixed-\(U\) model picks the first interpretation. That is coherent, but it is a newly specified macro-scale OCF problem, not a consequence of the paper’s characterization.
This is the best negative case, but it does not establish that no worthwhile mirror exists. A large fleet of interchangeable sensors or compute units is a plausible high-multiplicity regime, and the paper itself invites algorithmic work on its solution concepts. Theorem 4.1 gives a natural starting point for a computational study of mass blocking. So the case against is strongest as a scope objection—this paper lacks a computational result to mirror—and weak as a universal claim: the continuous TTG proposal remains plausible if the programme accepts a new computational question built from the paper’s model.
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.