{"id":"09d51b6e-300e-43e8-89fb-252ec2f5e5fc","arxiv_id":"1908.06866","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"The authors derive MBLP formulations for V2V multihop scheduling and power control under CCI and ACI, including latency and AoI constraints, and propose clustering and distributed algorithms to make them scalable.","lead":"This paper formulates joint scheduling and power control for vehicle-to-vehicle multicast with multihop relaying as mixed-integer optimization problems that account for both co-channel and adjacent channel interference. It also proposes a clustering method and a distributed scheduling algorithm that works without channel state information.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The connectivity objective (35) is mis-indexed: (35b) sums over all intended receivers, so a source is credited as connected to every receiver once any single receiver succeeds. This must be fixed before the claimed MBLP-optimal connectivity results can be accepted.","rationale":"The strongest claim is that every problem is an MBLP whose solution is optimal for the stated objective. The throughput and AoI/latency formulations are largely coherent: the SINR Big-M constraint, the W first-reception logic, and the probabilistic error-probability conversion are plausible. But the connectivity problem (35) is the paper's headline evaluation, and as printed its objective is not pairwise connectivity. The extra inner summation over j in R_i makes Z_i,j independent of the specific receiver and lets one successful link certify connectivity to all intended receivers. This is an internal inconsistency, not a disagreement with consensus or a missing practical condition. It would be demonstrated by the minimal instance above. The reader flagged the same line as an 'apparent indexing typo' but made the slowly varying CSI assumption the weakest point; I see the typo as the more load-bearing issue because it invalidates the published optimality claim even under perfect CSI. No code or artifacts are provided, so the equations are the only specification. Since the defect is localized and plausibly fixable, the paper should remain conditional, but the condition must include correcting (34)/(35b) and re-running the connectivity simulations.","tokens_in":24327,"tokens_out":13508,"duration_ms":153981,"concrete_test":"Build a minimal instance with N=3, R_0={1,2}, and choose H so that VUE 2 can receive a message from VUE 0 but VUE 1 cannot (e.g., a blocking vehicle makes all links into VUE 1 have SINR below gamma_T and no relay path can reach VUE 1). Solve (35) exactly as printed with Gurobi. If the optimum returns Z_0,1=1 because W_2,m,t=1, the printed constraint is overcounting. Then replace (35b) with Z_0,1 <= sum_m sum_t W_1,m,t and re-solve; if the two optima differ, the published MBLP is not the intended pairwise-connectivity problem. Re-run Figure 4 with the corrected constraint to quantify the effect on the reported connectivity curves.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The connectivity formulation in Section III-C3 is internally inconsistent. Constraint (35b) reads Z_i,j <= sum_{t=0}^{T-1} sum_{j in R_i} sum_{m in M_i} W_{j,m,t} for all i,j in R_i. The inner sum runs over every intended receiver of VUE i, so the right-hand side is the same for every destination j and becomes positive as soon as any single receiver in R_i decodes any message from i. Since Z_i,j <= 1 and the objective maximizes sum Z_i,j, the solver sets Z_i,j = 1 for all j in R_i whenever one link succeeds, even if that particular j never received the message. Equation (34) contains the same over-summing. Problem (35) therefore does not maximize pairwise connectivity; it maximizes, per transmitter, an indicator that at least one intended receiver was reached. The claimed MBLP-optimal solutions are optimal for a different objective. This holds even with perfect CSI, so it is more fundamental than the reader's CSI concern. The error is plausibly a typo (the inner sum should be over m and t only), but it must be corrected before the benchmark-optimality claim for connectivity can be accepted.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies joint scheduling and power control for V2V multicast with multihop relaying, incorporating both co-channel and adjacent-channel interference. It introduces binary scheduling variables X, link-success indicators Y, first-reception indicators W, and derives SINR constraints, latency variables, and age-of-information variables. The authors formulate several optimization problems as mixed Boolean linear programs: sum-throughput maximization, worst-case throughput, connectivity maximization, and connectivity under latency or AoI requirements. They also provide sufficient conditions that translate probabilistic latency/AoI constraints into deterministic ones via an end-to-end error-probability bound. For scalability, a clustering/grouping algorithm is proposed, along with a clustering-based distributed scheduler (CDS) that needs no channel state information. Simulations on a one-dimensional convoy topology compare the MBLP solutions (obtained with Gurobi) against the CDS heuristic and earlier scheduling schemes.","tokens_in":24599,"tokens_out":16266,"duration_ms":159342,"significance":"Modulo the issues below, the paper provides a valuable reference framework: the SINR transformation in Eq. (19) is coherent, the probabilistic latency/AoI arguments in Appendices B and C are genuine sufficient-condition derivations rather than fitted curves, and the clustering/CDS proposals are concrete answers to the scalability and no-CSI problems. If the connectivity formulation is repaired, the paper would supply a useful benchmark for V2V multihop multicast RRM with ACI. However, the connectivity objective as written does not maximize pairwise connectivity, and since the numerical section is built on that objective, the claimed benchmark-optimality of the connectivity results is not currently supported. The latency-constrained formulation also has an unresolved quantifier over messages. These are local, fixable issues, but they affect central claims.","major_comments":[{"comment":"The inner sum over j in R_i inside the definition of Z_{i,j} makes the right-hand side of (35b) independent of the destination j. As soon as any single receiver in R_i decodes any message from VUE i in any timeslot, the right-hand side is at least 1, so the constraint permits Z_{i,j}=1 for every j in R_i, even for receivers that never received a message. The objective (35a) then counts |R_i| connections whenever at least one link from i succeeds, so Problem (35) optimizes a per-source all-or-nothing objective rather than pairwise connectivity. This also invalidates the equivalence with the throughput problem (32) claimed in Section VII-B. The fix is to replace the inner sum over j with a sum over the specific receiver: Z_{i,j} <= sum_{m in M_i} sum_{t in S} W_{j,m,t} for each fixed j, in both (34) and (35b). Since the simulations in Section VII-B use this objective, the numerical results must be re-examined after the correction.","section":"Section III-C3, Eqs. (34) and (35b)"},{"comment":"The constraint tau_{j,m} <= tau_T + zeta (1 - Z^tau_{i,j}) is written without a quantifier over the message index m. If |M_i| > 1, it is undefined whether a pair (i,j) is counted as satisfying the latency requirement only when every m in M_i meets the deadline, or when at least one m does. The objective (37a) maximizes over pairs, so the meaning changes depending on the intended quantifier. The formulation should specify the quantifier and, if the requirement applies to all messages, impose the constraint for each m in M_i; if it applies to at least one message, introduce an auxiliary binary variable per message to represent the OR condition. As written, Problem (37) is not well-posed for the general multi-message setting described in Section II.","section":"Section III-C5, Eq. (37b)"}],"minor_comments":[{"comment":"The claim that zeta = gamma_bar (sigma^2 + N P^max) is sufficient assumes H_{k,j} <= 1 for all k,j and lambda_r <= 1 for all r; the manuscript should state this normalization explicitly or define zeta using max_{k,j} H_{k,j}.","section":"Section III-A4, after Eq. (18)"},{"comment":"The phrase 'if and only if' is too strong: the constraint only enforces that a VUE cannot transmit a message it does not yet hold. A VUE that holds a message remains free not to transmit, which is consistent with the optimization but should be described as a necessary condition.","section":"Section III-A5, Eq. (22)"},{"comment":"The text says a VUE is allocated a contiguous RB-group of 10 RBs, giving F=5, whereas the system model in Section II assumes one message per RB; the relationship between RBs and RB-groups should be defined in Section II to avoid confusion.","section":"Section VII-A, RB-Group definition"},{"comment":"The exponent notation in the complexity bounds is hard to parse; for example, the exponent 2^{|X|+|Y|+|W|} is rendered ambiguously as 2|X|+|Y|+|W| in places and should be typeset clearly.","section":"Section VI, complexity expressions"},{"comment":"There are several grammatical slips, including 'This paper investigate' in the abstract, 'probablistic' in Section III-B, and 'To the best of out knowledge' in Section VII-B.","section":"Abstract and Section VII-B"}],"recommendation":"major_revision","confidential_remarks":"The indexing error in the connectivity formulation appears to be a typo, but because the entire simulation study is built on (35), the authors will need to rerun the experiments after the fix and confirm the qualitative conclusions. The paper would also benefit from stating the quantifier over the message index in (37b)."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper extends the authors' earlier single-hop V2V scheduling and power control work in four directions: multihop relaying, AoI/latency constraints, clustering for scalability, and a CSI-free distributed scheduler. That combination — multihop plus ACI plus AoI in one MBLP framework — is genuinely new. The SINR, latency, and AoI derivations are coherent, and the probabilistic guarantees in Appendices B and C are valid sufficient conditions. The CDS scheduler is a sensible greedy heuristic with polynomial complexity, and the paper tests it against baselines and the MBLP optimum.\n\nThe main soft spot is a mis-indexing in the connectivity objective. Equations (34) and (35b) sum over j in R_i inside the definition of Z_{i,j}, so the right-hand side is the same for every destination j and becomes positive as soon as any one intended receiver decodes any message from i. The objective then maximizes, per transmitter, the number of receivers times an indicator that at least one receiver succeeded — not pairwise connectivity. The intended constraint is almost certainly Z_{i,j} ≤ sum_t sum_m W_{j,m,t} with the inner sum over j removed; under that correction, the paper's claim that (35) equals (32) in the one-message-per-VUE simulation scenario holds. As printed, the benchmark-optimality claim for connectivity is for the wrong objective. This is probably a typo rather than a fundamental error, but it has to be fixed before the connectivity results can be accepted.\n\nMinor issues: the simulation figures have no error bars, so it is hard to tell the signal from the noise in the comparison curves, and no code or data is released. The centralized MBLP needs slowly varying large-scale CSI for all VUE pairs, which is a real practical constraint, but the paper acknowledges it and CDS explicitly avoids the need for channel knowledge.\n\nThe citation pattern is fair. The self-citations are to the authors' own earlier framework, which is exactly what this paper extends.\n\nBottom line: this deserves a serious referee. With a corrected (35) and ideally error bars it is publishable. I would not desk-reject it.","headline":"Useful paper, but the connectivity objective as written is mis-indexed and solves a different problem; fix is trivial but mandatory.","tokens_in":25126,"tokens_out":4326,"would_cite":true,"duration_ms":41477,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Joint multihop V2V scheduling and power control with adjacent-channel interference can be solved optimally as mixed Boolean linear programs.","keywords":["V2V multicast communication","multihop scheduling","power control","adjacent channel interference","co-channel interference","mixed Boolean linear programming","age of information","latency constraints"],"falsifier":"Take a small convoy, say $N=4$ vehicles, $T=3$ timeslots, and $F=2$ frequency slots, fix all parameters, and solve the mixed Boolean linear programming formulation for maximum connectivity; then exhaustively enumerate every feasible Boolean schedule and, for each, solve the linear power-control subproblem. If enumeration produces a larger objective value than the mixed-integer solution, the formulation is not exact. Separately, in simulation with random link errors, check whether the fraction of end-to-end latency or age-of-information violations exceeds the chosen probability requirement; if it does despite the stated threshold rule, the error-probability conversion is too optimistic.","tokens_in":24157,"feed_emoji":"🚗","tokens_out":7124,"duration_ms":70432,"temperature":0.7,"pith_summary":"This paper claims that the hard radio-resource problem in vehicle-to-vehicle (V2V) multihop multicast—deciding who transmits when, on which frequency, and at what power, with both co-channel and adjacent-channel interference—can be recast as mixed Boolean linear programming for a whole family of objectives. Those objectives include total or worst-case throughput, connectivity, latency, and average age of information, each with a controllable probability of meeting a target. If this is right, then for any sufficiently small network a standard mixed-integer solver returns the provably optimal schedule and power allocation, giving a benchmark that heuristics can be measured against. The paper also supplies a clustering scheme to make the approach scale and a distributed, channel-free fallback for when no central scheduler has channel data.","feed_headline":"V2V multihop scheduling and power control solved as Boolean programs","feed_subtitle":"One framework optimizes throughput, connectivity, latency, and information age under adjacent-channel interference.","key_machinery":"The load-bearing mechanism is the conversion of a mixed Boolean and continuous radio-resource problem into mixed Boolean linear programming. Its central object is the triplet of Boolean matrices $X$ (who transmits which message in which resource block), $Y$ (which links meet the SINR threshold), and $W$ (which messages are first received when), connected by linearized Boolean OR, AND, and min constraints, together with the continuous power matrix $P$ inside a linearized SINR inequality. A second piece is the end-to-end error-probability bound: choosing $\\gamma_T$ so that $\\epsilon(\\gamma_T) \\leq \\epsilon_{\\text{req}}/N_{\\text{Tx}}$ guarantees any scheduled path's error probability stays below $\\epsilon_{\\text{req}}$, which lets probabilistic latency and age-of-information requirements be enforced by deterministic deadline and age constraints. The third piece is clustering: fixing a reuse distance from large-scale channel gains confines inter-cluster interference to $\\delta\\sigma^2$ and lets each group schedule independently, reducing the size of the exponential mixed-integer problem.","core_discovery":"The central discovery is that multihop relaying can be folded into an exact optimization model without dropping adjacent-channel interference. The paper treats each possible transmission as a Boolean variable $X_{i,m,f,t}$, each potentially successful link as $Y_{i,j,f,t}$, and each first reception as $W_{j,m,t}$, and writes all physical and logical constraints—power limits, half-duplex, SINR thresholds, message generation, relaying, latency, and age of information—as linear equalities and inequalities over mixed Boolean and continuous variables. The SINR condition is linearized through $P_{i,f,t}H_{i,j} \\geq \\bar{\\gamma}_T(\\sigma^2 + \\sum_{k,f'} P_{k,f',t} H_{k,j} \\lambda_{|f'-f|}) - \\zeta(1-Y_{i,j,f,t})$, with $\\bar{\\gamma}_T = \\gamma_T/(1+\\gamma_T)$ and $\\lambda_0 = 1$ for co-channel interference. Probabilistic latency and age-of-information requirements are converted into deterministic constraints by raising the SINR threshold so the end-to-end error probability is bounded appropriately. In this model, the optimal schedule and power allocation for throughput, connectivity, latency, information age, or max-min fairness is a computable mixed Boolean linear program, not just an aspiration.","pith_inferences":["A natural use of the formulation is as an exact benchmark oracle: for small networks, one could compare any fast heuristic's schedule against the optimum to measure suboptimality as a function of network size, timeslots, and frequency slots.","Because the framework only needs slowly varying average gains over a short scheduling horizon, a rolling-horizon implementation that re-estimates channel gains every scheduling interval is a direct extension, preserving optimality within each window.","The adjacent-channel interference model is parameterized by a generic mask $\\lambda_r$, so replacing the standardized mask with a measured power-amplifier emission spectrum should require no reformulation—only a new table of $\\lambda_r$ values.","The age-of-information probability guarantee uses a bound over the number of generated messages; empirical simulations may show the achieved violation probability is far below the target, meaning the SINR threshold could be relaxed in practice."],"forward_implications":["For any instance that fits the model, a solver returns a schedule and power allocation that is optimal among all multihop schedules under the chosen objective; no polynomial heuristic can beat it within the model.","Multihop relaying pays off: the simulations show joint multihop scheduling markedly improves average connectivity over the no-relay version, especially when there are more resource blocks than transmitters.","Probabilistic requirements on latency and age of information are guaranteed by construction once the SINR threshold is raised according to the end-to-end error bound, so the framework covers safety-message deadlines without needing retransmissions.","Clustering reduces the exponential complexity to per-group problems while bounding inter-cluster interference, and the distributed CDS algorithm removes the need for any channel knowledge while retaining multihop gains.","Worst-case (max-min) throughput and connectivity objectives sit in the same mixed Boolean linear programming framework, so fairness can be enforced exactly rather than by heuristic weighting."],"supporting_citations":[{"why":"Supplies the lemma that a link error probability below $\\epsilon$ is guaranteed when the SINR exceeds a threshold $\\gamma_T$, underpinning the success-condition constraints.","marker":"[35]"},{"why":"Defines the adjacent-channel interference ratio $\\lambda_r$ used to model leakage between neighboring frequency slots.","marker":"[36]"},{"why":"The authors' previous ACI-aware scheduling and power control study, which this paper generalizes to multihop, clustering, and age-of-information metrics.","marker":"[14]"},{"why":"A QoS-based D2D multicast scheduling algorithm adapted as a comparison baseline for connectivity maximization.","marker":"[51]"},{"why":"Provides the standardized ACIR emission mask and system parameters used to set $\\lambda_r$ and simulation values.","marker":"[38]"},{"why":"Supplies the measurement-based V2V pathloss model used to generate channel gains $H_{i,j}$ in the simulations.","marker":"[46]"},{"why":"Provides measured penetration loss from blocking vehicles, motivating both adjacent-channel interference awareness and multihop relaying.","marker":"[10]"},{"why":"Defines the age-of-information metric and its statistics, forming the basis for the AoI objective and requirements.","marker":"[3]"}],"fun_headline_variants":["Exact MILP for V2V multihop scheduling and power control","V2V multihop: Boolean programming optimizes throughput and AoI","Adjacent-channel interference handled in V2V multihop MILP","Joint scheduling and power control for V2V multihop, now exact"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central controller must know, in advance and for the whole scheduling window, the average radio signal strength between every relevant pair of vehicles; if that information is stale or missing, the optimal schedules promised by the model cannot be realized in a real vehicle network.","fun_headline_variants_meta":{"raw":{"variants":["Exact MILP for V2V multihop scheduling and power control","V2V multihop: Boolean programming optimizes throughput and AoI","Adjacent-channel interference handled in V2V multihop MILP","Joint scheduling and power control for V2V multihop, now exact"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000782,"raw_usage":{"total_tokens":3488,"prompt_tokens":1012,"completion_tokens":2476,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":628,"completion_tokens_details":{"reasoning_tokens":2394}},"tokens_in":628,"tokens_out":2476,"duration_ms":19678,"temperature":1.0,"reasoning_tokens":2394,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:51:54.683133+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a small convoy, say $N=4$ vehicles, $T=3$ timeslots, and $F=2$ frequency slots, fix all parameters, and solve the mixed Boolean linear programming formulation for maximum connectivity; then exhaustively enumerate every feasible Boolean schedule and, for each, solve the linear power-control subproblem. If enumeration produces a larger objective value than the mixed-integer solution, the formulation is not exact. Separately, in simulation with random link errors, check whether the fraction of end-to-end latency or age-of-information violations exceeds the chosen probability requirement; if it does despite the stated threshold rule, the error-probability conversion is too optimistic.","supporting_citations":[{"cited_title":"Ra- dio resource management for D2D-based V2V communication,","cited_arxiv_id":null,"evidence_quote":"Supplies the lemma that a link error probability below $\\epsilon$ is guaranteed when the SINR exceeds a threshold $\\gamma_T$, underpinning the success-condition constraints."},{"cited_title":"Dahlman, S","cited_arxiv_id":null,"evidence_quote":"Defines the adjacent-channel interference ratio $\\lambda_r$ used to model leakage between neighboring frequency slots."},{"cited_title":"Scheduling and power control for V2V broadcast communications with co-channel and adjacent channel interference,","cited_arxiv_id":null,"evidence_quote":"The authors' previous ACI-aware scheduling and power control study, which this paper generalizes to multihop, clustering, and age-of-information metrics."},{"cited_title":"A resource allocation scheme for D2D multicast with QoS protection in OFDMA-based sys- tems,","cited_arxiv_id":null,"evidence_quote":"A QoS-based D2D multicast scheduling algorithm adapted as a comparison baseline for connectivity maximization."},{"cited_title":"Evolved Universal Terrestrial Radio Access (E-UTRA); Radio Frequency (RF) system scenarios,","cited_arxiv_id":null,"evidence_quote":"Provides the standardized ACIR emission mask and system parameters used to set $\\lambda_r$ and simulation values."},{"cited_title":"Path loss modeling for vehicle-to-vehicle communications,","cited_arxiv_id":null,"evidence_quote":"Supplies the measurement-based V2V pathloss model used to generate channel gains $H_{i,j}$ in the simulations."},{"cited_title":"A measurement based shadow fading model for vehicle-to-vehicle network simulations,","cited_arxiv_id":null,"evidence_quote":"Provides measured penetration loss from blocking vehicles, motivating both adjacent-channel interference awareness and multihop relaying."},{"cited_title":"Reli- able ransmission of short packets through queues and noisy channels under latency and peak-age violation guarantees,","cited_arxiv_id":null,"evidence_quote":"Defines the age-of-information metric and its statistics, forming the basis for the AoI objective and requirements."}],"review_version":1}