| paper | d4-conf-aaai-1994-0217 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | no |
No result in the paper satisfies the required anchor test: the NP-hardness statement is attributed background, while CABINS’s computational claims are experimental findings without a named numbered computational result. The recurring job-family proposal is plausible, but it requires defining a new computational problem rather than mirroring a result from this paper.
fails bit a — no named computational result to mirror
No eligible computational result is covered; CABINS’s case-based repair method and experimental comparisons remain outside the mirror.
There is no eligible named computational result to anchor a continuous mirror on. The paper has no numbered theorem, lemma, corollary, or proposition asserting a complexity or algorithmic result. Its statement that job-shop scheduling is NP-hard is unnumbered background, attributed to French (1982); CABINS’s own results are experimental comparisons, not named computational results.
So there are zero anchors, and I can’t honestly write the requested anchor-by-anchor continuous problems. The most tempting candidate—the job-shop scheduling hardness claim—is not a result of this paper.
The proponent is right that this paper offers no eligible anchor: it proves no named complexity or algorithmic result. Its job-shop NP-hardness statement is background attributed to prior work. CABINS’s claims about capturing preferences, improving schedules from different starting points, and matching simulated annealing are experimental findings, not results about worst-case complexity or algorithms. So there is no paper result here for the programme to continuize.
The strongest candidate for a new question would be a factory processing recurring, standardized job families. One could describe the population by masses of job types—routes, due-date classes, and weights—and ask about continuous schedule repair. That is a plausible high-multiplicity setting. But it would require specifying the computational problem anew. CABINS’s central contribution is to use a case base of context-dependent user judgments in place of an explicit objective. Its cases depend on global schedule features, focal activities, repair histories, and effects that cascade across jobs. A distribution over job types alone would not specify those judgments or the changing schedule context. Including all that context in each type would make the “types” describe schedule states and histories, rather than a recurring population of interchangeable jobs.
One could instead add an explicit objective or formalize a distribution of user feedback and study a fluid scheduling model. Either may be worthwhile research, but each changes what is being studied: the former sidelines CABINS’s preference representation, while the latter supplies a computational problem the paper does not define. The paper’s experiments do not establish complexity claims for either model.
This is a narrow negative case, not proof that no useful continuous job-shop problem exists. Repeated production lots make a continuous population quite plausible. The claim I can support is that such a problem would be a new computational study, not a continuous mirror of a computational result in this paper.
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.