| paper | d4-conf-ijcai-1999-0103 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | unclear |
The paper gives no numbered Theorem, Lemma, Corollary, or Proposition asserting a computational result, so bit (a) fails. Its unnumbered hardness claim, solution procedures, and simulations cannot serve as anchors under the stated rule; the proposed population model also does not repair that gap.
fails bit a — no named computational result to mirror
No named computational result is covered. The unnumbered single-agent hardness claim, three solution procedures, and simulations do not meet the anchor rule.
This paper has no qualifying anchor: it states no numbered Theorem, Lemma, Corollary, or Proposition asserting a computational result. The nearest candidate is the unnumbered claim in §1 that the single-agent optimization problem is NP-hard. The paper also describes algorithms and reports simulations, but those are not named complexity results under your rules. So there are zero anchors, and I can’t honestly give a result-anchored continuous problem.
The best unanchored lead would be a continuous-population version of the satellite-sharing problem. A type would specify a co-funder’s quota, picture requests, and picture weights; \(\mu_t\) would be the fraction of the consortium with that complete profile. Over a fixed horizon, the decision would remain a daily selection of physically admissible pictures. The question could be to compute the Pareto frontier between mass-weighted normalized satisfaction and fairness of cost shares. Picture selection stays discrete, so the paper’s single-agent hardness claim suggests hardness could persist even with one type—though that claim is not a numbered anchor.
The main weakness is the regime: the paper describes only 3–8 agents with disjoint, agent-specific requests. A large population of co-funders sharing a small number of complete profiles is conceivable, but is a substantial step away from its stated setting.
The proponent is right that there is no qualifying named result to anchor. The paper’s only explicit complexity claim is the unnumbered NP-hardness statement for the single-agent problem; it does not analyze the complexity of the multi-agent sharing problem. Its contributions are three solution procedures and simulations. That leaves no result-anchored mirror to defend.
The proposed Pareto-frontier question is the strongest unanchored lead. Give it its best setting: a large consortium whose members fall into a small number of complete profiles, including quotas and request histories, and ask for the frontier between mass-weighted satisfaction and fairness. But in this paper, a profile must include the actual pictures requested and their weights. Requests are pairwise disjoint, and each picture is valuable only to its requester. So nonempty profiles cannot simply repeat: exact types proliferate with the agents, leaving no high-multiplicity compression. With a fixed satellite and horizon, each selected picture serves at most one agent; as individual masses shrink, aggregate normalized satisfaction tends to vanish. That makes the continuum degenerate in the very objective the proposed mirror would study.
One could avoid that degeneration by treating requests as interchangeable patterns or by letting one image serve a positive-mass cohort. But then the model must abstract away the image identities and their compatibility relations, or allow shared requests contrary to the paper’s assumptions. Scaling satellite capacity with the population changes the fixed-resource setting. These may be worthwhile scheduling models, but they do not supply a continuous mirror of a computational result in this paper.
The negative case is strongest against the paper as a ChoCo anchor, not against every imaginable satellite-sharing variant. A large consortium with common monitoring needs might support a useful mass model. That possibility keeps the universal claim weak; it does not repair the lack of a named result here, or the mismatch between the paper’s agent-specific demands and a finite-support continuous population.
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.