# arXiv cs.AI: FPTR for Deadline-Constrained Beam and Resource Allocation [Canonical reading](https://escapemeridian.com/readings/arxiv-cs-ai-fptr-for-deadline-constrained-beam-and-resource-1b2f4c93) Minghao Kong and coauthors propose FPTR, a method for short-interval wireless schedulers that must produce a complete, feasible beam-and-resource allocation before an execution deadline. Their central idea is to keep a validated allocation—the incumbent—safe while testing improvements separately. A candidate replaces it only if it is complete, passes validation before its stage cutoff, and improves the objective. Four refinement stages adjust resource selection, user sharing, beam masks and the reassignment of two resources. The authors argue this lets spare computation time improve an allocation without risking the usable result already in hand. In tests on 150 fixed synthetic instances, the full pipeline raised traffic by 3.83% over a lightweight beam-first reference and 1.79% over its Base constructor. In a separate matched-input suite, it exceeded six search baselines by 6.68–8.35%, with outputs later than 100 ms assigned zero utility. All 750 full-pipeline runs in the internal suite returned feasible allocations within 100 ms. As an Escape Meridian interpretation, these results describe the reported test suites, not demonstrated performance in deployed wireless networks. The authors say the implementation is open source for independent reproduction. Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims. ## Original source arXiv cs.AI · Article Original source date: Nov 4, 2026 (2026-11-04). Date basis: Date basis not provided. Added to Escape Meridian: 2026-11-09T06:38:46.059Z. First observed: 2026-11-05T05:09:33.452Z. Retrieved: 2026-11-09T06:32:20.227Z. - [FPTR: Feasibility-Preserving Transactional Refinement for Deadline-Constrained Joint Beam and Resource Allocation](https://arxiv.org/abs/2611.06491) ## Limits The reported evaluation includes 150 fixed synthetic instances and a separate internal suite; the abstract does not report results from deployed wireless networks. The abstract gives aggregate results but does not describe the test instances or evaluation methodology in enough detail to assess their representativeness. ## Open question How would FPTR perform on wireless scheduling workloads beyond the paper’s fixed synthetic instances and internal test suite? ## Source-stated The authors propose FPTR to preserve a validated feasible allocation while evaluating refinements separately; a candidate replaces it only after meeting completeness, feasibility, timing and objective-improvement conditions. > A candidate replaces the incumbent only if it is complete, passes the feasibility validator before its stage cutoff, and strictly improves the objective - [Supporting source](https://arxiv.org/abs/2611.06491) ## Source-stated The authors report that, on 150 fixed synthetic instances, Full raised traffic by 3.83% over a lightweight beam-first reference and 1.79% over its Base constructor. > On 150 fixed synthetic instances, the full pipeline (Full) raises traffic by 3.83% on average over a lightweight beam-first reference and by 1.79% over its Base constructor. - [Supporting source](https://arxiv.org/abs/2611.06491) ## Source-stated In a separate matched-input suite, the authors report that Full exceeded six search baselines by 6.68–8.35%; outputs later than 100 ms received zero utility. > In a separate suite with matched inputs, Full exceeds six independent search baselines (ALNS, tabu search, GA, simulated annealing, iterated local search, and GRASP) by 6.68-8.35%, where outputs later than 100 ms receive zero utility. - [Supporting source](https://arxiv.org/abs/2611.06491) ## Source-stated The authors report that all 750 Full runs in their internal suite returned feasible allocations and none exceeded 100 ms. > All 750 Full runs in the internal suite returned feasible allocations, and none exceeded 100 ms. - [Supporting source](https://arxiv.org/abs/2611.06491) ## Escape Meridian interpretation The reported results come from fixed synthetic instances and a separate internal suite, so the abstract does not establish performance in deployed wireless networks. > On 150 fixed synthetic instances - [Supporting source](https://arxiv.org/abs/2611.06491) ## Source and contributor context Kong, Minghao, Liu, Zirong, Wu, Jianhao, Chen, Muyao, Wu, Jianxin, Wang, Rongjie · Article author These are credits stated by the source; they do not establish a profile match or authorship of every passage. ## Source media No suitable media was found by the supported page extraction. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## arXiv arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed. - [Explore this context](/entities/arxiv) - [Reference](https://arxiv.org/) - [Reference](https://info.arxiv.org/help/api/index.html) ## TIME TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services. - [Explore this context](/entities/time) - [Reference](https://time.com/) ## Editorial provenance Version 1. Publication ID: d5e28d88-a11c-4d50-8c32-c123af1548fb. Prepared by gpt-6-luna; grounding review by gpt-6-luna. Source check: passed by anthropic/claude-opus-5.5 on 2026-11-09T06:38:46.059Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: material\_development. Observation: fd5dc073aef7445c9953ccaa947062f6d4287f41d4ddb19ddda5b0d2ae8733ce. Source text hash: e9840f92d30a8a465dbc864d1c5b0ce5587eb993866025298fd4af507d27b1fa. ## Revision and source-check history ### Version 1 — 2026-11-09T06:38:46.059Z Minghao Kong and coauthors propose FPTR, a method for short-interval wireless schedulers that must produce a complete, feasible beam-and-resource allocation before an execution deadline. Their central idea is to keep a validated allocation—the incumbent—safe while testing improvements separately. A candidate replaces it only if it is complete, passes validation before its stage cutoff, and improves the objective. Four refinement stages adjust resource selection, user sharing, beam masks and the reassignment of two resources. The authors argue this lets spare computation time improve an allocation without risking the usable result already in hand. In tests on 150 fixed synthetic instances, the full pipeline raised traffic by 3.83% over a lightweight beam-first reference and 1.79% over its Base constructor. In a separate matched-input suite, it exceeded six search baselines by 6.68–8.35%, with outputs later than 100 ms assigned zero utility. All 750 full-pipeline runs in the internal suite returned feasible allocations within 100 ms. As an Escape Meridian interpretation, these results describe the reported test suites, not demonstrated performance in deployed wireless networks. The authors say the implementation is open source for independent reproduction. ## Original source arXiv cs.AI · Article Original source date: Nov 4, 2026 (2026-11-04). Date basis: Date basis not provided. Added to Escape Meridian: 2026-11-09T06:38:46.059Z. First observed: 2026-11-05T05:09:33.452Z. Retrieved: 2026-11-09T06:32:20.227Z. - [FPTR: Feasibility-Preserving Transactional Refinement for Deadline-Constrained Joint Beam and Resource Allocation](https://arxiv.org/abs/2611.06491) ## Limits The reported evaluation includes 150 fixed synthetic instances and a separate internal suite; the abstract does not report results from deployed wireless networks. The abstract gives aggregate results but does not describe the test instances or evaluation methodology in enough detail to assess their representativeness. ## Open question How would FPTR perform on wireless scheduling workloads beyond the paper’s fixed synthetic instances and internal test suite? ## Source-stated The authors propose FPTR to preserve a validated feasible allocation while evaluating refinements separately; a candidate replaces it only after meeting completeness, feasibility, timing and objective-improvement conditions. > A candidate replaces the incumbent only if it is complete, passes the feasibility validator before its stage cutoff, and strictly improves the objective - [Supporting source](https://arxiv.org/abs/2611.06491) ## Source-stated The authors report that, on 150 fixed synthetic instances, Full raised traffic by 3.83% over a lightweight beam-first reference and 1.79% over its Base constructor. > On 150 fixed synthetic instances, the full pipeline (Full) raises traffic by 3.83% on average over a lightweight beam-first reference and by 1.79% over its Base constructor. - [Supporting source](https://arxiv.org/abs/2611.06491) ## Source-stated In a separate matched-input suite, the authors report that Full exceeded six search baselines by 6.68–8.35%; outputs later than 100 ms received zero utility. > In a separate suite with matched inputs, Full exceeds six independent search baselines (ALNS, tabu search, GA, simulated annealing, iterated local search, and GRASP) by 6.68-8.35%, where outputs later than 100 ms receive zero utility. - [Supporting source](https://arxiv.org/abs/2611.06491) ## Source-stated The authors report that all 750 Full runs in their internal suite returned feasible allocations and none exceeded 100 ms. > All 750 Full runs in the internal suite returned feasible allocations, and none exceeded 100 ms. - [Supporting source](https://arxiv.org/abs/2611.06491) ## Escape Meridian interpretation The reported results come from fixed synthetic instances and a separate internal suite, so the abstract does not establish performance in deployed wireless networks. > On 150 fixed synthetic instances - [Supporting source](https://arxiv.org/abs/2611.06491) ## Source and contributor context Kong, Minghao, Liu, Zirong, Wu, Jianhao, Chen, Muyao, Wu, Jianxin, Wang, Rongjie · Article author These are credits stated by the source; they do not establish a profile match or authorship of every passage. ## Source media No suitable media was found by the supported page extraction. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## arXiv arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed. - [Explore this context](/entities/arxiv) - [Reference](https://arxiv.org/) - [Reference](https://info.arxiv.org/help/api/index.html) ## TIME TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services. - [Explore this context](/entities/time) - [Reference](https://time.com/) ## Editorial provenance Version 1. Publication ID: d5e28d88-a11c-4d50-8c32-c123af1548fb. Prepared by gpt-6-luna; grounding review by gpt-6-luna. Source check: passed by anthropic/claude-opus-5.5 on 2026-11-09T06:38:46.059Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: material\_development. Observation: fd5dc073aef7445c9953ccaa947062f6d4287f41d4ddb19ddda5b0d2ae8733ce. Source text hash: e9840f92d30a8a465dbc864d1c5b0ce5587eb993866025298fd4af507d27b1fa. ### passed — 2026-11-09T06:38:46.059Z — anthropic/claude-opus-5.5 Publication: d5e28d88-a11c-4d50-8c32-c123af1548fb attribution — pass: The reading correctly credits Minghao Kong and coauthors, matching the listed authors. It frames the method and results as the authors' proposal and reports. The limitation about deployed networks is explicitly labeled as an Escape Meridian interpretation. > Authors:Minghao Kong, Zirong Liu, Jianhao Wu, Muyao Chen, Jianxin Wu, Rongjie Wang summary — pass: The reading faithfully conveys the core idea: always keep a validated feasible allocation (the incumbent) and test improvements separately. It also covers the replacement conditions, the four stages, and the reported numbers. Terms such as beam-and-resource allocation, Base constructor and beam masks are kept exactly, consistent with the tooltip approach. The deadline stake is conveyed. The question follows from the evidence and is not leading. Optional minor repair: restore the source's 'strictly improves' and its point that a late or invalid output gives no utility in the current control cycle. > a late or invalid output gives no utility in the current control cycle date — pass: originalPublished 2026-11-04 matches the v1 submission date on arXiv. The feed date of 2026-11-05 was correctly not used. > \[Submitted on 4 Nov 2026\] support — pass: All numeric claims match the abstract: 150 instances, 3.83%, 1.79%, six baselines, 6.68-8.35%, 100 ms, and 750 runs. Saying runs finished 'within 100 ms' matches 'none exceeded 100 ms'. The interpretation about deployed networks is properly labeled and conditional. The limitations do not overreach, and no benefits are invented. > All 750 Full runs in the internal suite returned feasible allocations, and none exceeded 100 ms. media — pass: The selected media is the verified direct link to the arXiv abstract page for this paper. It is the most direct and relevant way to access the source. ## Recorded references Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter. No public references recorded here yet. [Reference data (JSON)](/readings/1b2f4c9325c9f11ea01f2292e3db0abf3b0cc1b4b8bc67b9fac26ed211550915/references.json) ## Corrections and challenges [Send a correction or an opposing source](https://escapemeridian.com/corrections/) [JSON record](https://escapemeridian.com/readings/arxiv-cs-ai-fptr-for-deadline-constrained-beam-and-resource-1b2f4c93.json)