{"id":"214e039a-1aca-4f47-a693-31a8cdb29ab2","arxiv_id":"2505.02530","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"The authors report that a zebra optimization algorithm for user pairing and power allocation improves simulated energy efficiency of a CR-NOMA smart grid network by up to roughly 50% over the compared baselines.","lead":"This paper proposes a zebra-optimization-based algorithm for user pairing and power allocation in a cognitive radio NOMA smart grid network, reporting large energy efficiency gains in simulations. A smart generalist would read it as an incremental optimization paper with no code or error bars, showing how metaheuristics are applied to smart grid spectrum sharing.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equation (4) omits the strong user's interference in the weak user's SINR, inflating rates and EE; corrected rates would change the reported gains.","rationale":"The reader's weakest assumption identifies the same load-bearing flaw: Eq. (4) assigns the weak user a rate that ignores the strong user's interference, despite the described SIC order. This is not a stylistic disagreement or a matter of convention; it is an internal inconsistency in the rate model used to compute every numerical result. Because the optimization objective and constraints are built from these rates, the claimed EE values and comparative gains are not reliable as reported. The appropriate response is to require a correction and re-evaluation, which matches the reader's CONDITIONAL verdict; my review therefore does not change that verdict. I would not escalate to REJECT because the error appears fixable: the algorithmic framework (BCD decomposition, ZOA-based pairing and power allocation) could still be valid with corrected SINR expressions, and the qualitative comparisons might survive. A simple re-derivation and simulation rerun would settle the matter, so a conditional acceptance with a concrete request is proportionate.","tokens_in":17492,"tokens_out":3051,"duration_ms":36596,"concrete_test":"Re-derive the two-user uplink NOMA rates with the stated SIC order (weak user ν decoded first, strong user µ decoded after cancellation): γ_ν = log2(1 + δν p |gν|^2 / (δµ p |gµ|^2 + σ^2)) and γ_µ = log2(1 + δµ p |gµ|^2 / σ^2). Re-run the ZOUP and ZOUPPA simulations with these corrected rates under the same parameters as Table II and compare the resulting EE values and percentage improvements in Table III. If the 15 dB improvement over UPWO drops substantially or ZOUPPA no longer dominates, the central claim fails; if the ranking is preserved, the qualitative conclusion survives.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central quantitative claim is that ZOUPPA achieves 53.25% higher EE than UPWO and 25.39% higher than ZOUP at 15 dB SNR (Table III). Every EE value and improvement percentage is computed from the rate expressions in Eqs. (4) and (5). Equation (4) gives the weak user's uplink NOMA rate as log2(1 + δν p |gν|^2 / σ^2), with no interference term from the strong user. But the text states that the DC decodes the weak user first, before SIC removes the strong user's signal, so the strong user's transmission is present as interference during that decoding. The correct SINR for the weak user is δν p |gν|^2 / (δµ p |gµ|^2 + σ^2), not δν p |gν|^2 / σ^2. Equation (5) is also inconsistent with the stated SIC order: after the weak user is cancelled, the strong user should be decoded with only noise, so its SINR should be δµ p |gµ|^2 / σ^2, not δµ p |gµ|^2 / (δν p |gµ|^2 + σ^2). These errors are not cosmetic: the missing interference term in Eq. (4) typically reduces the weak user's rate, and because the objective (6) and constraint C1 both depend on these rates, the reported EE values and all percentage improvements in Section IV would change under a correct rate model. The central claim therefore rests on an internally inconsistent rate derivation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper considers an uplink CR-NOMA system for a smart-grid neighborhood area network in which smart meters transmit to a data collector. It formulates a joint user pairing, channel assignment, and power allocation problem to maximize energy efficiency, decomposes it via block coordinate descent into pairing and power-allocation subproblems, and solves each subproblem with the zebra optimization algorithm. The claimed contribution is a heuristic (ZOUPPA) that, according to Monte Carlo simulations, improves energy efficiency by 53.25% over UPWO and 25.39% over ZOUP at 15 dB SNR, with similar gains across SNR, path loss, user density, channel availability, and coverage radius.","tokens_in":17892,"tokens_out":8131,"duration_ms":81640,"significance":"If the reported results were valid, the proposed heuristic would be a practically relevant contribution to resource allocation in smart-grid NANs, combining CR spectrum access with NOMA user pairing and power control. The paper covers an extensive parameter sweep and compares with several benchmarks. However, the current manuscript contains two load-bearing quantitative errors: the NOMA rate expressions are inconsistent with the stated SIC order, and the improvement percentages in Table III are computed with the wrong denominator. In addition, the NP-hardness and convergence claims are asserted rather than proved, and no code or data are provided. These issues prevent the central claim from being accepted at face value.","major_comments":[{"comment":"The rate expressions do not match the stated successive interference cancellation order. The text states that the DC decodes the weak user ν first and then the strong user μ; under that order, the weak user's decoded signal suffers interference from the strong user, so Eq. (4) should contain δμ p |gν|^2 in the denominator, and after cancellation the strong user's Eq. (5) should have no δν term. As written, Eq. (4) omits the strong-user interference and Eq. (5) includes a term that the stated SIC order would have removed. Because the objective in (6) and the QoS constraint C1 in (7b) depend directly on these rates, every reported EE value and percentage improvement in Section IV is computed under an optimistic and internally inconsistent model. The authors should correct the rate model and rerun the simulations.","section":"Section II-A, Eqs. (4)-(5) and objective (6)"},{"comment":"The percentage improvements are not consistent with the reported EE values. For the 15 dB SNR row, the table lists % improvement to UPWO = 53.25% and to ZOUP = 25.39%; these are obtained as (ZOUPPA−UPWO)/ZOUPPA and (ZOUPPA−ZOUP)/ZOUPPA, respectively, not as (ZOUPPA−benchmark)/benchmark. Using the standard convention, the correct values for that row would be 113.9% vs UPWO and 34.0% vs ZOUP. The same pattern appears in every row of Table III. Since the abstract and conclusion quote these percentages, the headline quantitative claims are not supported by the table. The authors should either recompute the percentages with the benchmark as denominator or explicitly define the different convention, and then update the abstract and conclusion.","section":"Section IV-B, Table III"},{"comment":"The statement 'F(Si+1) ≥ F(Si)' and the convergence criterion are asserted without proof. ZOA is a stochastic metaheuristic that uses random foraging and defense updates, and the BCD decomposition solves each subproblem heuristically, so monotonic improvement of the objective is not guaranteed. The cited reference [52] does not establish convergence for this algorithm. Either provide a proof under the specific update rules and penalty/feasibility handling, or reword the section to describe empirical convergence and remove the claim that ZOUPPA converges when changes in the energy efficiency metric fall below a threshold in the sense of a rigorous optimization guarantee.","section":"Section III-C, Convergence Analysis"},{"comment":"The claim that (P1) is NP-hard is made without a proof or a reference. Because the paper proposes a heuristic, this claim is not central to the algorithmic contribution, but if it is kept it should be justified by a reduction from a known NP-hard problem, or it should be replaced by a statement about computational intractability due to the mixed-integer structure.","section":"Section II-B and Section III"}],"minor_comments":[{"comment":"The abstract quotes 53.2% and 25.4% at 15 dB, while Table III gives 53.25% and 25.39%; the values should be consistent after the percentage calculation is corrected.","section":"Abstract and Section IV-B"},{"comment":"The user-pairing improvement percentages, such as the '9.5% improvement in UPWO' at 40 dB, are reported without stating the denominator used; the same denominator ambiguity as in Table III may affect these numbers, and they should be verified.","section":"Section IV-A"},{"comment":"The power allocation coefficients δu and δv are not clearly defined; δu is expressed as a sum involving both gu and gv, and then δv = 1 − δu, which makes it unclear which term corresponds to which user. Please define the notation and state the feasible range of δu and δv explicitly.","section":"Section II-A, Eq. (9)"},{"comment":"There are numerous typographical errors and inconsistent notations, e.g., 'sub-probelm', 'V arying', 'channle', 'MA X', 'δu and δv' versus 'δµ and δν', and the use of 'bpj/Hz' as the EE unit. A careful proofreading pass is needed.","section":"Throughout"},{"comment":"No simulation code, data, or error bars are provided, which makes it impossible to assess the statistical significance of the reported average EE gains; please consider making the code available or at least reporting confidence intervals.","section":"Section IV"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's central quantitative claims are not currently reproducible from the data given; the rate-model error and the denominator error in Table III are both fixable but require rerunning the simulations. I recommend major revision with an explicit request for corrected tables and code/data availability. The heavy reliance on the authors' own prior work (refs. [44]-[46], [51], [54]) for both the system model and one of the benchmarks is not improper, but the novelty claims should be checked against those references."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nWhat you should know: the headline EE improvements in this paper rest on a NOMA rate model that contradicts the stated SIC order. Equation (4) gives the weak user's rate without the strong user's interference, and Eq. (5) gives the strong user's rate with weak-user interference in the denominator, exactly backwards from what the text describes. Since the objective and constraint C1 are built on these rates, every EE value and percentage improvement in Section IV changes when the equations are corrected. The stress-test note is right about this, and it is the load-bearing flaw.\n\nThe paper isn't a waste. The specific combination—zebra optimization for both user pairing and power allocation in a CR-NOMA smart grid NAN—is new as far as I know. The system model, BCD decomposition, and benchmark comparisons are clearly laid out. The authors also did a thorough literature review, even if it's heavier on citations than on comparative insight. If the rate equations are fixed and the simulations rerun, the method might still show a reasonable heuristic gain. That's a modest but real contribution.\n\nSoft spots beyond the equation error: the convergence analysis is two sentences asserting monotonic improvement without proof, which isn't adequate for a so-called convergence claim. There's no code and no error bars, so the Monte Carlo numbers are not independently checkable. The abstract's 15 dB ZOUP numbers aren't in the tables, but I don't see a direct inconsistency with Table III—the 13.6% and 78.8% refer to a different comparison that is only plotted, not tabulated. So I'd focus on the equation issue, not that.\n\nThe reader's report called it conditional. I'm a little more negative: the central quantitative claim is invalid as written. But the flaw is mechanical, not conceptual, and a revision could turn this into a passable engineering paper. I would send it to peer review with a clear note about the rate expressions, and require a corrected model before anything is accepted.\n\nRegards.","headline":"The central EE gains are built on swapped NOMA SINR denominators, but the flaw is mechanical and the paper is a modest, fixable engineering application.","tokens_in":18353,"tokens_out":6481,"would_cite":false,"duration_ms":71229,"reading_group":"maybe","serious_thinker":"no","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that a zebra-optimization pairing and power allocation scheme maximizes energy efficiency in CR-NOMA smart-grid neighborhood networks, with simulated gains of 53.25% over unoptimized pairing at 15 dB SNR.","keywords":["smart grid communications","cognitive radio","NOMA","energy efficiency","user pairing","power allocation","zebra optimization algorithm","neighborhood area network"],"falsifier":"Recompute the Monte Carlo runs with the weak-user rate written as log2(1 + δν p |gν|^2 / (δμ p |gν|^2 + $σ^{2}$)) instead of the interference-free form in Eq. (4), keeping the same random channel draws, and check whether ZOUPPA still beats UPWO and ZOUP by the reported percentages.","tokens_in":17335,"feed_emoji":"⚡","tokens_out":7246,"duration_ms":81569,"temperature":0.7,"pith_summary":"This paper tries to establish that a joint user pairing and power allocation method, built on the zebra optimization algorithm, maximizes the energy efficiency of a cognitive-radio non-orthogonal multiple access (CR-NOMA) smart grid neighborhood area network. The authors cast the pairing-and-power problem as a single NP-hard, nonconvex optimization and then split it into two subproblems using block coordinate descent, solving each with the zebra heuristic. In Monte Carlo simulations they report that the joint method lifts efficiency by 53.25% over unoptimized pairing and 25.39% over pairing-only optimization at 15 dB signal-to-noise ratio, with comparable gains as SNR, path loss, user density, channel availability, and coverage radius vary. If this holds, smart meter data collection could use less energy and less spectrum while still meeting minimum rate requirements.","feed_headline":"Zebra algorithm lifts smart-grid network efficiency 53%","feed_subtitle":"Pairing and power heuristic beats simpler schemes for cognitive-radio neighborhood networks, per simulation.","key_machinery":"The load-bearing machinery is block coordinate descent (BCD) applied to the mixed-integer nonlinear fractional objective (P1), splitting it into a channel and user-pairing subproblem (P1-A) and a power-allocation subproblem (P1-B). Each subproblem is solved by the zebra optimization algorithm (ZOA), a bio-inspired metaheuristic that updates candidate solutions through a foraging phase and a predator-defense phase, keeping the better of the initial and updated candidates. The pairing stage uses a beta-based power formula from the literature to set initial power coefficients, and the energy efficiency objective is evaluated as the sum of NOMA rates of each pair divided by the pair's power. This decomposition is what turns the intractable joint problem into a sequence of tractable, locally convergent updates.","core_discovery":"On the paper's own terms, the central discovery is that optimizing the pairing before optimizing the power allocation, both through the zebra optimization algorithm, produces a near-optimal energy-efficient schedule for a CR-NOMA smart grid network. The paper states that the resulting ZOUPPA procedure achieves its largest reported gain at 15 dB SNR (53.25% over UPWO and 25.39% over ZOUP) and that this advantage persists across all parameter sweeps reported in Table III. It frames this as the first joint treatment of user pairing, power allocation, and channel allocation with CR-NOMA in the neighborhood-area-network layer of a smart grid.","pith_inferences":["An editorial next step would be to benchmark ZOUPPA against an exact or convexified fractional-programming solver on the same instances; the paper compares only against heuristics, so the remaining optimality gap is unknown.","The objective enters the algorithm only through the fitness function, so the same BCD-plus-ZOA structure could be reused for other NAN goals such as latency, fairness, or throughput without changing the pairing or power-update rules.","Because the channel list is assumed fixed for at least 48 hours, a natural extension is to re-run ZOUPPA at each channel-database update and quantify how much efficiency is lost if the pairing is frozen in between."],"forward_implications":["Using ZOUPPA in a neighborhood area network would pair smart meters on the same channel and assign their transmit powers to deliver more data per unit energy than fixed power allocation, constant-beta power allocation, unoptimized pairing, adjacent pairing, or random pairing, under the simulated conditions.","Jointly optimizing pairing and power adds roughly 13% to 52% to efficiency over pairing-only optimization across the parameter sweeps, so power allocation is a substantial part of the gain and not an afterthought.","The reported advantage persists across suburban to dense-urban path-loss exponents, SNR from 10 to 40 dB, 100 to 300 smart meters, 60 to 120 channels, and coverage radii of 100 to 500 m, suggesting the heuristic is not tuned to one operating point.","The stated convergence rule (stop when the change in energy efficiency drops below a small threshold) gives an implementable online stopping criterion for a data collector that re-runs the algorithm when the channel database updates."],"supporting_citations":[{"why":"Defines the zebra optimization algorithm whose foraging and defense updates generate the candidate pairings and power vectors for ZOUP and ZOUPPA.","marker":"[53]"},{"why":"Supplies the beta-based power allocation formula (Eq. 9) used as the initial power allocation when the pairing subproblem is solved.","marker":"[50]"},{"why":"Establishes the block coordinate descent convergence rationale that justifies splitting the joint problem into P1-A and P1-B.","marker":"[52]"},{"why":"Defines the UPWO benchmark (strong and weak user grouping) against which ZOUP and ZOUPPA report percentage improvements.","marker":"[54]"},{"why":"Provides the clustering-based channel allocation model and the assumption that channel availability is stable during assignment in the NAN.","marker":"[44]"},{"why":"Sets the IEEE 802.11af cognitive-radio smart grid communication model that frames the secondary-user scenario for smart meter transmissions.","marker":"[46]"}],"fun_headline_variants":["Zebra algorithm boosts smart-grid energy efficiency 53%","Smart-grid CR-NOMA gains 53% energy efficiency via zebra pairing","NOMA smart grid: zebra pairing lifts energy efficiency 53%","Energy-efficient smart grids via zebra-optimized NOMA pairing","ZOUPPA: zebra-powered pairing boosts smart-grid efficiency 53%"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The efficiency result assumes that when the weaker user's signal is decoded, the stronger user's signal adds no interference, even though both signals occupy the same channel and the strong signal is still present.","fun_headline_variants_meta":{"raw":{"variants":["Zebra algorithm boosts smart-grid energy efficiency 53%","Smart-grid CR-NOMA gains 53% energy efficiency via zebra pairing","NOMA smart grid: zebra pairing lifts energy efficiency 53%","Energy-efficient smart grids via zebra-optimized NOMA pairing","ZOUPPA: zebra-powered pairing boosts smart-grid efficiency 53%"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000316,"raw_usage":{"total_tokens":1828,"prompt_tokens":1024,"completion_tokens":804,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":640,"completion_tokens_details":{"reasoning_tokens":709}},"tokens_in":640,"tokens_out":804,"duration_ms":8714,"temperature":1.0,"reasoning_tokens":709,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T00:48:33.861725+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the Monte Carlo runs with the weak-user rate written as log2(1 + δν p |gν|^2 / (δμ p |gν|^2 + $σ^{2}$)) instead of the interference-free form in Eq. (4), keeping the same random channel draws, and check whether ZOUPPA still beats UPWO and ZOUP by the reported percentages.","supporting_citations":[{"cited_title":"Zebra o ptimization algorithm: A new bio-inspired optimization algorithm for s olving op- timization algorithm,","cited_arxiv_id":null,"evidence_quote":"Defines the zebra optimization algorithm whose foraging and defense updates generate the candidate pairings and power vectors for ZOUP and ZOUPPA."},{"cited_title":"A general pow er allocation scheme to guarantee quality of service in downlink and uplin k NOMA systems,","cited_arxiv_id":null,"evidence_quote":"Supplies the beta-based power allocation formula (Eq. 9) used as the initial power allocation when the pairing subproblem is solved."},{"cited_title":"Capacity optimiz ation of next-generation UA V communication involving non-orthogonal multiple access,","cited_arxiv_id":null,"evidence_quote":"Defines the UPWO benchmark (strong and weak user grouping) against which ZOUP and ZOUPPA report percentage improvements."},{"cited_title":"Clustering-based channel allocation scheme for neighbor hood area network in a cognitive radio based smart grid communication ,","cited_arxiv_id":null,"evidence_quote":"Provides the clustering-based channel allocation model and the assumption that channel availability is stable during assignment in the NAN."},{"cited_title":"Joint power and channel allocation scheme for IEEE 802.11 af based smart grid communication network,","cited_arxiv_id":null,"evidence_quote":"Sets the IEEE 802.11af cognitive-radio smart grid communication model that frames the secondary-user scenario for smart meter transmissions."}],"review_version":1}