{"id":"7f137893-0b74-402e-aedc-4eba03e38e4b","arxiv_id":"2501.10242","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":3.0,"correctness_risk":"high","formal_verification":"none","parameter_count":7,"one_line_summary":"A hybrid clustering-plus-quantum method is claimed to cut sensor-network routing energy by 83.16% in one simulation, but the evidence is incomplete and internally inconsistent.","lead":"This paper uses a quantum optimization recipe, split across smaller clusters, to plan routing in wireless sensor networks, and reports one simulated case where it saves 83 percent of energy. The result is not backed by open code, a clear comparison method, or a fully valid math model.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The QUBO in Eq. (10) does not encode the stated routing problem: flow conservation is infeasible in subgraphs without a sink, and the squared energy penalty forces energy equality, not inequality; reported savings are therefore not from a valid feasible routing.","rationale":"Most load-bearing: the validity of the central numerical claim depends entirely on Eq. (10) encoding the WSN routing problem. It does not. The flow-conservation term is infeasible for clusters without a sink, and the energy penalty converts a 'less than or equal' budget into an 'equal to' target. Even if the QAOA execution were perfect, the returned bitstrings cannot be interpreted as energy-minimizing routes. The paper does include a sensible high-level idea (partitioning + QAOA on subgraphs) and a discussion of near-term hardware constraints, but the evidence for the headline improvement is not tied to a valid model. The reader's REJECT is therefore appropriate; the arithmetic slip further reduces confidence, but the model flaw is decisive.","tokens_in":6930,"tokens_out":5734,"duration_ms":57177,"concrete_test":"Analytically verify infeasibility: for a reported subgraph that does not contain the base station, sum Eq. (4) over all nodes in that subgraph. The left-hand side telescopes to zero for every edge set, while the right-hand side equals the number of sensor nodes in the subgraph (positive). This contradiction proves the QUBO has no feasible solution, so any QAOA output for that subgraph violates flow conservation. For a computational cross-check, exhaustively enumerate all 2^|Es| assignments for the smallest reported cluster and confirm zero assignments satisfy Eq. (4).","verdict_should_be":"UNCHANGED","load_bearing_attack":"Eq. (10) cannot represent the routing problem defined in §II-D. For any subgraph Gs, summing the flow-conservation constraint (4) over i∈Vs gives Σ_i (out_i − in_i) = 0, but Σ_i b_i = (# sensors in Vs) when no base station is in Vs, and the BS term −Σ_all_sensors only appears in the single cluster containing the BS. Thus in the reported 5-cluster network (Fig. 3), at least four subgraphs have no feasible binary assignment, so QAOA minimizers for them are not valid routes. Additionally, Eq. (5) is the inequality Σ c_ij x_ij ≤ E_i, yet the penalty in Eq. (10) squares (Σ c_ij x_ij − E_i), which is minimized at exact equality and penalizes energy savings; this actively distorts the objective rather than enforcing the bound. A consequence visible in the text: C_initial − C_total,quantum = 94,593.5 − 15,982.9 = 78,610.6, not the claimed 78,901.6, so the reported 83.16% reduction is not derived from the paper's own numbers.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper proposes a hybrid classical-quantum framework for wireless sensor network (WSN) routing. Spectral clustering partitions the network into subgraphs; each subgraph's routing problem is formulated as a QUBO and solved with QAOA, followed by classical postprocessing (base-station/CH connection and BFS repair) to restore connectivity. The paper reports a simulation with 109 nodes and 5 clusters in which the hybrid QAOA approach yields 83.16% energy reduction versus 68.32% for an underspecified classical greedy search, and discusses resource-efficient compilation strategies for distributed QPUs.","tokens_in":7264,"tokens_out":3982,"duration_ms":37491,"significance":"If the modeling and numerical claims were valid, the paper would provide an interesting near-term application of QAOA to network routing with a concrete scalability strategy, and the compilation discussion addresses a real hardware constraint. The framework is a standard composition of spectral clustering, QUBO, and QAOA, and it is not circular: the objective and constraints are defined independently of the reported outcome. However, the core optimization model in Eq. (10) does not enforce the stated inequality constraint and is infeasible for most subgraphs in the reported experiment, and the central energy-saving number is arithmetically inconsistent. The paper also provides no code, data, or detailed baseline specification, so the main performance claim is not reproducible. The manuscript is therefore not publishable in its current form.","major_comments":[{"comment":"The quadratic penalty λ_energy Σ_i (Σ_j c_ij x_ij − E_i)^2 does not enforce the inequality constraint in Eq. (5). A squared penalty is minimized when Σ_j c_ij x_ij = E_i, so solutions using less energy than the budget are penalized rather than accepted; the term actively pushes each node toward its full energy budget and does not express Σ_j c_ij x_ij ≤ E_i. Since the validity of the optimized route is central to the energy-savings claim, the QUBO must be reformulated (for example with slack variables or a one-sided penalty) or the energy constraint must be shown to be nonbinding.","section":"II-D, Eq. (10)"},{"comment":"Summing the flow-conservation constraints over all nodes of a subgraph G_s gives Σ_i (Σ_j x_ij − Σ_j x_ji) = Σ_i b_i. For any subgraph that does not contain the base station, Σ_i b_i equals the number of sensor nodes in that subgraph (positive), while the left-hand side is zero when the QUBO is restricted to internal edges E_s as defined. In the reported 5-cluster network, at least four subgraphs have no sink, so no binary assignment can satisfy all node-wise constraints; the QAOA minimizers for those subgraphs are therefore not valid routes of the stated flow-conservation problem. This is a load-bearing infeasibility in the core model.","section":"II-D, Eq. (4); III, 5-cluster network"},{"comment":"The reported quantum saving does not follow from the paper's own numbers. C_initial − C_total,quantum = 94,593.5 − 15,982.9 = 78,610.6, not 78,901.6, and the corresponding reduction is 83.10%, not 83.16%. The percentage is computed from an inconsistent ΔC_quantum value; the paper should report one consistent set of numbers throughout Section III.","section":"III"},{"comment":"The classical baseline is described only as “a greedy search in subgroups,” with no algorithm specification, pseudocode, penalty or parameter values, or per-subgraph results. Since the paper's central claim is that hybrid QAOA beats classical methods, the comparison needs a clearly defined baseline, ideally with multiple random trials and dispersion measures; without this, the 68.32% versus 83.16% comparison is not reproducible.","section":"III"}],"minor_comments":[{"comment":"The definition of b_i for the base station, “− Σ_{i∈S} b_i”, reuses the summation index i and is ambiguous; use a separate index, e.g., b_BS = −Σ_{s∈S} 1.","section":"II-D"},{"comment":"Equation (3) refers to FlowConstraints(x) and EnergyConstraints(x) before their explicit forms are given in Eq. (10); reorder or label the equations for readability.","section":"II-D"},{"comment":"The caption contains a duplicated period: “Fig. 2. . Compilation strategies.”","section":"Fig. 2 caption"},{"comment":"The text says Figure 4 illustrates the optimized topology for both methods, but the caption only mentions the hybrid QAOA approach; reconcile the caption or include the classical routing result.","section":"III / Fig. 4"},{"comment":"No values are reported for λ_flow, λ_energy, the number of QAOA layers p, or the subgraph limit n_max, so the simulation cannot be reproduced from the text.","section":"III"},{"comment":"The terms “subgroups” and “subgraphs” are used interchangeably; choose one term throughout.","section":"II-F, III"}],"recommendation":"reject","confidential_remarks":"The manuscript is not in a publishable state: the core QUBO is internally inconsistent with the stated constraints, the reported headline numbers do not add up, and the classical baseline is not described. Correcting these issues requires a reformulated model, rerun experiments, and a reproducible comparison, which go beyond ordinary revision. The citation list is broad and no disclosure concern arises beyond the absence of code/data."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is a cleanly written but routine application of known tools: spectral clustering, QUBO formulation, and QAOA per subgraph, with classical stitching. The exposition is clear, the references are appropriate (including the authors' own prior distributed-QAOA work, which is relevant), and the complexity analysis is a nice touch. But the central claim—that the hybrid quantum approach beats a classical greedy search by 83%—does not survive a close look at the QUBO encoding.\n\nTwo load-bearing problems. First, the flow-conservation constraint in Eq. (10) is infeasible for any subgraph that does not contain the base station. Summing Eq. (4) over a subgraph gives zero on the left, but the right-hand side equals the number of sensor nodes in that subgraph, which is positive. In the reported 5-cluster network, at least four subgraphs have no feasible binary assignment, so QAOA minimizers for them are not valid routes. Second, the energy penalty squares (sum c_ij x_ij - E_i), which penalizes energy consumption below the cap as heavily as above it. That enforces equality, not the stated inequality, and actively distorts the objective.\n\nThe numbers also do not add up: C_initial - C_total,quantum = 94,593.5 - 15,982.9 = 78,610.6, not the claimed 78,901.6. So the reported 83.16% reduction is not derived from the paper's own values. The greedy baseline is never described, there are no error bars or repeated trials, and only one random instance is tested. The resource-efficient compilation framing is legitimate, and the paper may work as a proof-of-concept after major revision, but in its current form the optimized circuit is not solving the stated routing problem.\n\nI would not send this to peer review as is. The authors need to fix the QUBO encoding (slack variables or a proper inequality penalty), benchmark against a standard baseline, run multiple instances with statistics, and release code and data. If they do that, a revised version could be a modest contribution worth a second look.","headline":"Routine QAOA-for-WSN paper whose headline energy savings are invalidated by an infeasible QUBO encoding and inconsistent arithmetic.","tokens_in":7748,"tokens_out":2340,"would_cite":false,"duration_ms":23830,"reading_group":"no","serious_thinker":"no","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A hybrid spectral-clustering/QAOA routing method for wireless sensor networks reports an 83.16% energy reduction in a 109-node simulation, beating the 68.32% achieved by a greedy search baseline.","keywords":["wireless sensor networks","QAOA","spectral clustering","QUBO","distributed quantum computing","routing optimization","energy efficiency","hybrid classical-quantum algorithms"],"falsifier":"Inspect the reported optimized network: if any sensor node's outgoing-edge costs exceed its initial energy $E_i = 100$, or if any node's out-flow minus in-flow differs from $b_i$, then the penalty encoding failed and the claimed energy savings does not describe a feasible routing. Alternatively, solve the same small subgraphs ($n_s \\leq 25$) by exhaustive search and check whether the QAOA-selected edge set remains cheaper than the greedy baseline; if the greedy solution is cheaper, the central comparison fails.","tokens_in":6737,"feed_emoji":"📡","tokens_out":8728,"duration_ms":80051,"temperature":0.7,"pith_summary":"This paper claims that the routing problem in large wireless sensor networks can be made tractable for near-term quantum hardware by splitting the network with spectral clustering and solving each cluster's routing subproblem with the Quantum Approximate Optimization Algorithm (QAOA). The routing task is encoded as a Quadratic Unconstrained Binary Optimization (QUBO) problem whose binary variables select directed edges, with penalty terms for flow conservation and per-node energy limits. In a simulated 109-node network with 100 sensors, 8 cluster heads, and one base station, the authors report total energy consumption of 15,982.9 units after the hybrid QAOA approach, an 83.16% reduction from the initial 94,593.5 units, compared with 29,969.7 units (68.32% reduction) for a greedy search in subgroups. The significance would be that a practical hybrid classical-quantum workflow, not a full-scale fault-tolerant machine, could give large energy savings for network optimization on current hardware.","feed_headline":"Hybrid quantum routing cuts sensor-network energy 83%","feed_subtitle":"Spectral clustering plus QAOA beats greedy search on a 109-node network in simulation.","key_machinery":"The load-bearing mechanism is the QUBO formulation in Eq. (10) combined with a spectral-clustering partition. Binary variables $x_{ij}$ mark whether directed edge $(i,j)$ is used; the objective sums transmission costs $c_{ij} = \\varepsilon d_{ij}^2$ and adds squared penalties for flow conservation (out-flow minus in-flow equals $b_i$) and for exceeding each node's initial energy $E_i$. Spectral clustering partitions the network into $k$ subgraphs ($k$ equal to the number of cluster heads), keeping each subproblem small enough for a QAOA circuit with alternating problem and mixing Hamiltonians. A resource-efficient compilation strategy, including a heavy-hexagonal qubit layout and modular distributed QPUs, is what makes the per-subgraph QAOA execution plausible on current hardware.","core_discovery":"The central discovery is that a resource-efficient, distributed QAOA workflow can outperform a classical greedy search on the same wireless sensor network routing problem. The authors formulate routing over each spectral cluster as a QUBO, constrain subgraphs to at most 25 binary variables, simulate the QAOA circuit classically, and then stitch cluster-level routes together with breadth-first search so the whole network remains connected. On their 109-node test network the quantum-enhanced optimization yields total energy consumption of 15,982.9 units and an 83.16% energy reduction, exceeding the greedy search's 68.32% reduction. The paper presents this as evidence that partitioning plus QAOA is a scalable route to large network-optimization problems on near-term devices, while acknowledging that parameter selection and current hardware quality still limit solution optimality.","pith_inferences":["A test the paper does not run: execute the same per-subgraph QUBO on real heavy-hexagonal hardware with error mitigation, since the reported numbers come from classical simulation of the QAOA circuit.","The reported 83.16% figure is a comparison against the greedy-search baseline; solving the same small subgraphs exactly would show how close the QAOA solution sits to the true QUBO optimum.","If the penalty encoding and stitching step are sound, the same recipe should transfer to other graph-constrained combinatorial problems, where subgraphs small enough for QAOA can be carved out and solved independently.","The flow-conservation condition assumes the base station can absorb the total sensor flow even when it lies outside a cluster; if that assumption fails, extra stitching edges are needed and the savings could shift."],"forward_implications":["Large WSN routing problems can be attacked without a fault-tolerant quantum computer by capping each QAOA subgraph at a hardware-tolerable size and using classical post-processing for inter-cluster connectivity.","The tested 109-node instance is a proof of concept: the hybrid method reports 15,982.9 units of total energy versus 29,969.7 for greedy search, a difference the paper attributes to quantum-enhanced optimization.","The complexity analysis places spectral clustering at $O(N^3)$ and QAOA circuit depth at $O(p \\times \\log(n_s))$, so the practical bottleneck can shift between classical partitioning and quantum execution as the network grows.","Because each subgraph is solved independently and then stitched with breadth-first search, the same QUBO formulation works for a single QPU, a clustered QPU, or a modular multi-QPU setup, which is what makes the approach resource-efficient."],"supporting_citations":[{"why":"Supplies the spectral-clustering method used to partition the network into subgraphs.","marker":"[24]"},{"why":"Defines the QAOA ansatz whose alternating unitaries solve each subgraph's QUBO.","marker":"[9]"},{"why":"Provides the QUBO formulation conventions used to encode routing constraints as penalty terms.","marker":"[25]"},{"why":"Establishes the distributed QAOA approach on near-term hardware that the compilation strategy extends.","marker":"[13]"},{"why":"Supplies the role-based initial energy values (100 for sensors, 200 for cluster heads) used in the energy constraints.","marker":"[22]"},{"why":"Supplies the energy-efficiency modeling and metrics that define total consumption and savings.","marker":"[27]"},{"why":"Supplies the heavy-hexagonal device architecture that the resource-efficient compilation strategy targets.","marker":"[21]"},{"why":"Supplies the QAOA circuit-depth scaling $O(p \\times \\log(n_s))$ cited in the complexity analysis.","marker":"[16]"}],"fun_headline_variants":["Quantum QAOA routing beats greedy by 83% energy cut","Distributed quantum circuits slash sensor network energy 83%","Spectral clustering + QAOA wins WSN routing with 83% savings","Hybrid quantum-classical method cuts WSN energy 83%","QAOA on partitioned WSN cuts energy use 83% vs greedy"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the squared penalty terms in the encoded optimization problem (Eq. 10) force the solution to respect both flow conservation and each node's energy limit, and that a subgraph whose sink lies outside it still produces routes that can be stitched into a valid whole-network routing.","fun_headline_variants_meta":{"raw":{"variants":["Quantum QAOA routing beats greedy by 83% energy cut","Distributed quantum circuits slash sensor network energy 83%","Spectral clustering + QAOA wins WSN routing with 83% savings","Hybrid quantum-classical method cuts WSN energy 83%","QAOA on partitioned WSN cuts energy use 83% vs greedy"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000175,"raw_usage":{"total_tokens":1227,"prompt_tokens":826,"completion_tokens":401,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":442,"completion_tokens_details":{"reasoning_tokens":310}},"tokens_in":442,"tokens_out":401,"duration_ms":3918,"temperature":1.0,"reasoning_tokens":310,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T19:18:01.415297+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Inspect the reported optimized network: if any sensor node's outgoing-edge costs exceed its initial energy $E_i = 100$, or if any node's out-flow minus in-flow differs from $b_i$, then the penalty encoding failed and the claimed energy savings does not describe a feasible routing. Alternatively, solve the same small subgraphs ($n_s \\leq 25$) by exhaustive search and check whether the QAOA-selected edge set remains cheaper than the greedy baseline; if the greedy solution is cheaper, the central comparison fails.","supporting_citations":[{"cited_title":"A tutorial on spectral clustering,","cited_arxiv_id":null,"evidence_quote":"Supplies the spectral-clustering method used to partition the network into subgraphs."},{"cited_title":"Quadratic unconstrained binary optimiza- tion problem preprocessing: Theory and empirical analysis,","cited_arxiv_id":null,"evidence_quote":"Provides the QUBO formulation conventions used to encode routing constraints as penalty terms."},{"cited_title":"Energy- efficient communication protocol for wireless microsensor networks,","cited_arxiv_id":null,"evidence_quote":"Supplies the role-based initial energy values (100 for sensors, 200 for cluster heads) used in the energy constraints."},{"cited_title":"Modeling and optimization of energy efficient routing in wireless sensor networks,","cited_arxiv_id":null,"evidence_quote":"Supplies the energy-efficiency modeling and metrics that define total consumption and savings."},{"cited_title":"Ibm’s roadmap for scaling quantum technology,","cited_arxiv_id":null,"evidence_quote":"Supplies the heavy-hexagonal device architecture that the resource-efficient compilation strategy targets."},{"cited_title":"A review on quantum approximate optimization algorithm and its variants,","cited_arxiv_id":null,"evidence_quote":"Supplies the QAOA circuit-depth scaling $O(p \\times \\log(n_s))$ cited in the complexity analysis."}],"review_version":1}