{"id":"e4e88e7b-d464-4df2-a092-024a5806e169","arxiv_id":"2501.05046","paper_version":1,"verdict":"REJECT","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":4,"one_line_summary":"A 7-node space logistics routing problem was encoded as a QUBO-style Hamiltonian and solved on QCi's Dirac-3 entropy quantum computer, producing a feasible but admitted-suboptimal plan.","lead":"This paper ran a small space supply-routing problem on a photonic quantum optimization machine called Dirac-3. The machine returned a workable but not optimal delivery plan, and the authors use this to argue quantum computing could help plan complex interplanetary missions.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central claim rests on a single manually post-processed sample from a 7-node, 2-commodity instance, with no verification that the repaired solution satisfies the penalty Hamiltonian and no classical baseline; the feasibility and intractability conclusions are unsupported.","rationale":"The paper's mathematical formulation is standard and clearly written, and the experimental section is transparent about the raw output's flaws. However, the central claim in the Conclusion goes beyond what the evidence can establish. The only quantitative support is a best-of-40 hardware sample that required manual post-processing and is admitted to be non-optimal. The reader's weakest-assumption analysis correctly identifies the unvalidated penalty separation and Hamiltonian construction as the place where the feasibility argument is least secure. I would add that the absence of any classical baseline makes the 'intractable for classical methods' statement untestable as written. The proposed check settles the matter directly: verify the repaired solution's feasibility against the stated constraints, and compare against an exact classical solver. If classical optimization solves this 7-node instance instantly, the paper's stronger conclusion fails regardless of whether the quantum hardware produced a feasible schedule. No ad hominem is intended; the concern is about the logical gap between one repaired sample and a general feasibility/advantage claim. The verdict REJECT remains appropriate because the paper does not supply the evidence needed to support its central assertion.","tokens_in":7375,"tokens_out":6355,"duration_ms":58872,"concrete_test":"Reconstruct the exact MILP from Tables 1-2 and Eqs. (1)-(4) with Delta_t_ij=1, W=100, L1=10, L2=20. First evaluate the reported repaired schedule (Tables 4-6) against constraints (2)-(4) and compute P2 for the Hamiltonian (7) using the paper's alpha; if P2 is nonzero or any conservation or capacity constraint is violated, the solution is infeasible. Second, solve the same instance with a classical MILP solver (e.g., PuLP/CBC or Gurobi) and record the optimal objective and solve time. If the classical solver finds an objective below 62.08 Km/s or solves in well under a second, the reported non-optimal quantum result cannot support the paper's 'intractable for classical methods' conclusion.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section IV.B reports that the best of 40 Dirac-3 samples (energy -1989.2251) contained spurious commodity flows that were removed by post-processing ('set all commodity flows to 0 when no vehicle was scheduled'), and the resulting schedule is explicitly admitted to be non-optimal ('one more vehicle than is required at time step 4 from N3 to N4'). The paper gives no value of P2(x,z) for the repaired solution, no check that constraints (2)-(4) hold on the reported Tables 4-6, and no comparison to a classical solver. The conclusion that such networks are 'intractable for classical methods' is therefore not supported by the experiment: the instance has 7 nodes and 2 commodities, and the authors do not even report a classical run time. The Hamiltonian separation in Eq. (7) with a manually chosen alpha is asserted, not demonstrated; because the raw device samples were infeasible, there is no evidence that low-energy samples correspond to feasible flows. Without a reproducibility artifact or an independent validation step, the central feasibility claim is a single anecdote rather than a demonstration.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a quantum-assisted framework for space logistics mission planning, formulating the problem as a time-dependent multicommodity network flow (MCNF) on a small seven-node interplanetary network. The authors construct a polynomial Hamiltonian from the MCNF objective and constraints, submit the resulting model to QCi's Dirac-3 entropy quantum computing hardware, and report a single best-of-40-sample solution that is post-processed to remove unscheduled commodity flows. The paper claims this experiment validates the feasibility of quantum-assisted frameworks for logistical networks that are intractable for classical methods.","tokens_in":7648,"tokens_out":4451,"duration_ms":45411,"significance":"If the central claim were established, the paper would provide a useful early demonstration of a quantum hardware platform applied to space logistics optimization. The authors deserve credit for a concrete end-to-end effort: they define a time-expanded MCNF instance, translate it into a polynomial Hamiltonian, execute it on real hardware, and report sample times and energy distributions. However, the evidence as presented is anecdotal: the reported solution is admitted to be non-optimal, the raw hardware output required manual repair, no classical solver comparison is provided, and the instance is tiny. The manuscript therefore does not currently support its stated conclusions about feasibility or classical intractability.","major_comments":[{"comment":"The central feasibility claim is not supported because the reported solution is the result of manual repair rather than the direct output of the quantum optimization. The text states that the raw best-energy sample contained spurious commodity flows with no scheduled vehicle, and that the authors therefore \"set all commodity flows to 0 when no vehicle was scheduled.\" The same paragraph admits that the resulting schedule is not optimal, specifically that there is \"one more vehicle than is required at time step 4 from N3 to N4.\" Since Tables 4-6 report only the repaired solution, the paper provides no evidence that low-energy samples of H(x,z) correspond to feasible flows. The authors should report the penalty value P2(x,z) for the repaired solution, verify constraints (2)-(4) explicitly, and state exactly how the post-processing changes the objective and feasibility status.","section":"Section IV.B (post-processing)"},{"comment":"No classical baseline is given. The instance has 7 nodes, 2 commodities, 8 arcs, and 6 time steps, yet the conclusion states that the results validate quantum-assisted frameworks for \"logistical networks that are intractable for classical methods.\" This claim is unsupported because the paper reports no classical run time, no exact solver comparison, and no scaling study. At minimum, the authors should solve the same instance with an exact MCNF or MILP solver, report the optimal objective and runtime, and either remove the intractability claim or substantiate it with problem-size scaling experiments.","section":"Section IV.B and Section V (comparison to classical methods)"},{"comment":"The Hamiltonian separation needed for the method is asserted rather than demonstrated. The penalty multiplier alpha is chosen manually, and because the raw device samples were infeasible before post-processing, the paper gives no evidence that the chosen alpha produces an energy gap separating feasible from infeasible states. The authors should provide either a proof that all infeasible states have energy greater than all feasible states for the chosen alpha, or an empirical sweep of alpha showing feasibility rates and constraint violations as functions of alpha.","section":"Eq. (7) and Section III.B (penalty construction)"},{"comment":"The mapping from the MCNF model to the hardware form is underspecified. The hardware variables xi are described as nonnegative real variables subject to a sum constraint, while the model contains integer decision variables x^{k,t}_{ij} and z^t_{ij}. The paper does not state how equality constraints (2) and inequality constraints (3) are converted into the polynomial P2, nor how integer variables are encoded using photon time bins. Without this information, the experiment is not reproducible and the reported 116 variables, 331 levels, and dynamic range cannot be independently checked.","section":"Section III.B (encoding of the MCNF model)"},{"comment":"The reported tables do not by themselves demonstrate a feasible flow. Table 6 is described with demand remaining in inventory at destination nodes, which conflicts with the standard flow-conservation form in Eq. (2) unless supplies and demands are treated asymmetrically; the relationship between the cargo movements in Table 5 and the commodity flows in Table 6 is not made explicit. A constraint-by-constraint verification of Eqs. (2)-(4) should be included for the final repaired solution, with any inventory conventions clearly defined.","section":"Tables 4-6 (feasibility verification)"}],"minor_comments":[{"comment":"There are several typographical errors, including \"issuies\" in the Introduction, \"vechicle\" and \"coeffients\" in Section IV.B, and \"auxilary\" in the same section; a careful proofreading pass is needed.","section":"Throughout"},{"comment":"The parameter table refers to costs as \"Listed in the arcs table,\" but no separate arcs table appears; the Delta-V values are only labeled in Fig. 1, which makes the exact c_ij values and the resulting objective value 62.08 Km/s difficult to verify.","section":"Table 2"},{"comment":"The paper reports 116 variables, 89 decision variables, 29 slack variables, and 331 levels, but does not explain how these counts are derived from the model or how the slack variables enter P2; a variable table or formula would improve reproducibility.","section":"Section IV.B (variable counts)"},{"comment":"The conclusion states that the approach \"simultaneously evaluates multiple solutions,\" but the experiment reports only 40 samples; this should be phrased as a hardware capability rather than a demonstrated advantage of the reported results.","section":"Section V (conclusion)"}],"recommendation":"reject","confidential_remarks":"The manuscript is closer to a hardware demonstration than a scientific validation: the only reported solution is post-processed, non-optimal, and unaudited against a classical solver. The load-bearing claims in Section V are not supported by the experiments, and I do not see a local revision that would fix this without a substantially new experimental and benchmarking effort. If the authors resubmit, the paper would need an exact classical baseline, a feasibility audit, a penalty-parameter sensitivity study, and a scaling argument before the intractability claim could be considered."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this one. It's a clean write-up of a standard time-expanded MCNF formulation with a QUBO penalty encoding, run on QCi's Dirac-3 for a 7-node, 2-commodity space logistics network. The authors are transparent that the best of 40 samples was infeasible and had to be repaired by zeroing unsupported cargo flows, and they admit the resulting schedule is suboptimal. That honesty is real credit.\n\nWhat's new: the specific demonstration of EQC hardware on a space logistics MCNF instance. The math itself—time-expanded MCNF and penalty-based QUBO—is textbook, following the cited tutorial. No new algorithm, no new bound, no reproducibility artifacts. The literature review is adequate, and the citation to their own prior entropy-computing work is appropriate.\n\nWhere it falls apart: the conclusion says the results 'validate the feasibility of quantum-assisted frameworks in optimizing logistical networks that are intractable for classical methods.' The instance has 7 nodes and 2 commodities. There is no classical solver comparison, no runtime comparison, and no check that the repaired solution actually satisfies constraints (2)–(4). The raw device samples were infeasible; the post-processed solution is hand-patched. The penalty weight alpha is manually chosen with no sensitivity analysis. The dynamic range discussion is plausible but doesn't rescue the claim. In short, the evidence is a single anecdote, not a demonstration.\n\nThe paper is honest about its own limitations in Section IV.B, which makes the conclusion's overreach puzzling. The 'intractable for classical methods' phrase is not backed by any experiment. The central feasibility claim—that low-energy EQC samples correspond to feasible flows—is asserted, not shown.\n\nBottom line: This is for readers who want a first look at applying a specific quantum device to space logistics. It should not be cited as evidence of quantum advantage, or even of quantum feasibility beyond this toy instance. As a referee, I'd ask for a classical baseline and a feasibility check on the repaired solution before publishing. The core question—does EQC produce feasible MCNF solutions?—remains unanswered.","headline":"A transparent but tiny demonstration of a standard QUBO encoding on Dirac-3; the overreaching conclusion is not supported by the evidence.","tokens_in":8156,"tokens_out":2601,"would_cite":false,"duration_ms":26011,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["90B10","90C27","90C11","81P68"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper attempts to show that entropy quantum computing, running on the Dirac-3 photonic device, can produce a feasible—though admittedly non-optimal—schedule for a time-dependent multicommodity network flow model of an interplanetary…","keywords":["quantum computing","entropy quantum computing","space logistics","multicommodity network flow","mission planning","Hamiltonian optimization","interplanetary logistics","Dirac-3"],"falsifier":"Solve the paper's 7-node, 2-commodity, 6-time-step instance to proven optimality with a classical mixed-integer linear program; if the solver finishes in under a second, the paper's concluding claim that such networks are intractable for classical methods fails, and the only remaining claim is the narrower feasibility demonstration.","tokens_in":7177,"feed_emoji":"🚀","tokens_out":12094,"duration_ms":103214,"temperature":0.7,"pith_summary":"The paper attempts to establish that a quantum optimization device can generate a feasible logistics plan for an interplanetary supply network, a problem class it argues classical methods cannot efficiently handle. The authors cast the problem as a time-expanded multicommodity network flow on a 7-node graph spanning Earth, lunar orbits, the lunar surface, and Mars, with two cargo types and six time steps. They encode the total transportation cost and all flow and capacity constraints into a single Hamiltonian, then sample low-energy states on the Dirac-3 entropy quantum computer. The best of 40 samples, after a small repair step, yields a schedule that meets every demand with an objective value of 62.08 km/s; the paper openly states this is not optimal. A sympathetic reader should see the result as a proof-of-feasibility: quantum hardware of this kind can produce workable schedules for small logistics networks, with the scalability claims left for future work.","feed_headline":"Quantum hardware finds a feasible interplanetary logistics plan","feed_subtitle":"An entropy quantum computer schedules two cargo types across a 7-node lunar-to-Mars network, meeting all demands.","key_machinery":"The load-bearing object is the Hamiltonian $H(x,z)=P_1(x,z)+\\alpha P_2(x,z)$ of Eq. (7): $P_1$ accumulates arc costs $c_{ij}z^t_{ij}$ over vehicles, and $P_2$ is a quadratic penalty that converts the flow-conservation and vehicle-capacity constraints into energy terms, following the QUBO recipe. The multiplier $\\alpha$ is chosen manually to make infeasible solutions carry higher energy than feasible ones, while the 116-variable Hamiltonian (89 decision variables and 29 slack variables) is sampled on the Dirac-3 photonic device using the software-configurable quantum fluctuation coefficient $1/\\sqrt{7}$ and relaxation schedule 2. The device's entropy-driven dissipation acts like imaginary time evolution, amplifying low-energy eigenstates over repeated samples; the argument is that low-energy samples correspond to good candidate plans, requiring only the small repair of zeroing commodity flows on arcs where the sampler scheduled no vehicle.","core_discovery":"The central claim is that a time-dependent multicommodity network flow formulation of a space logistics campaign can be mapped to a polynomial Hamiltonian of the form $H(x,z)=P_1(x,z)+\\alpha P_2(x,z)$, where $P_1$ is the transportation cost, $P_2$ is a quadratic penalty built from flow-conservation and capacity constraints in the QUBO style, and $\\alpha$ is a hand-tuned multiplier that lifts infeasible solutions to higher energies. Sampling this Hamiltonian on the Dirac-3 entropy quantum computer produced, in the best of 40 samples, a vehicle schedule whose repair (setting commodity flows to zero wherever no vehicle was scheduled) satisfies all supplies and demands with an objective value of 62.08 km/s. The authors emphasize that this solution is not optimal—one extra vehicle traverses the $N_3$–$N_4$ arc at time step 4—and they attribute the gap partly to the Hamiltonian's 30.78 dB dynamic range approaching the device's limit. The paper presents the result as a feasibility demonstration: entropy quantum computing can explore the solution space of a logistics network and return a workable plan, which the authors take as evidence that the approach could scale to problems that resist classical optimization.","pith_inferences":["A natural extension the paper does not pursue is to benchmark this exact instance against a classical MILP solver; the 7-node, 6-time-step problem is likely trivial for branch-and-cut, so the concluding 'intractable for classical methods' phrasing goes beyond what this single experiment establishes.","The manual choice of $\\alpha$ and the one-off repair step suggest reproducibility could be strengthened by an automatic penalty-selection rule and by reporting statistics over all samples, not just the best-energy one; the paper currently shows a distribution of energies but not a distribution of feasibility rates.","If dynamic range is the binding constraint, then a testable improvement would be to rescale or drop the two smallest cost coefficients (below 0.86 km/s) and re-run; the paper mentions this trade-off but does not explore it.","The time-expanded formulation with $\\Delta t_{ij}=1$ for all arcs is a strong idealization; extending the approach to variable flight times and stochastic demands would stress the Hamiltonian encoding and reveal whether the method generalizes beyond this stylized instance."],"forward_implications":["A feasible plan for the 7-node, 2-commodity, 6-time-step network is obtained from low-energy samples, implying that the Hamiltonian encoding captures the constraints well enough for the device to land in the feasible region.","The reported runtime of about 2.5 seconds per sample, with roughly 1.5 seconds on the device, indicates that the method can return candidate schedules quickly for instances of this size.","Because the returned solution is explicitly non-optimal, the paper's claim is about feasibility only; optimality would require additional algorithmic or hardware improvements.","The dynamic-range discussion implies that coefficient scaling directly affects solution quality, so the method's practical range is bounded by the hardware's ability to resolve small energy differences."],"supporting_citations":[{"why":"Introduces entropy quantum computing and the Dirac-3 hardware, the device used for all sampling in this paper.","marker":"[10]"},{"why":"Provides the QUBO penalty formulation used to turn constraints into the quadratic penalty $P_2$.","marker":"[11]"},{"why":"Defines the integrated space logistics mission planning problem with spacecraft, ISRU, and scheduling that motivates this work.","marker":"[3]"},{"why":"Formulates space transportation as a network flow with multiple commodities, the model this paper time-expands and solves.","marker":"[9]"},{"why":"Presents the reinforcement-learning alternative that the quantum approach is positioned against for solving the same class of planning problems.","marker":"[4]"}],"fun_headline_variants":["Entropy quantum computer schedules lunar-Mars cargo","Quantum device finds workable interplanetary cargo plan","Quantum sampling yields feasible space logistics route","Quantum computer meets lunar-Mars logistics demands","Entropy quantum computing explores space logistics"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The demonstration rests on the assumption that the hand-picked penalty weight $\\alpha$ makes low-energy samples of the Hamiltonian correspond to near-feasible logistics plans, so that raw outputs need only a small repair step to become valid schedules.","fun_headline_variants_meta":{"raw":{"variants":["Entropy quantum computer schedules lunar-Mars cargo","Quantum device finds workable interplanetary cargo plan","Quantum sampling yields feasible space logistics route","Quantum computer meets lunar-Mars logistics demands","Entropy quantum computing explores space logistics"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000147,"raw_usage":{"total_tokens":1151,"prompt_tokens":877,"completion_tokens":274,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":493,"completion_tokens_details":{"reasoning_tokens":207}},"tokens_in":493,"tokens_out":274,"duration_ms":3364,"temperature":1.0,"reasoning_tokens":207,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:20:06.037128+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Solve the paper's 7-node, 2-commodity, 6-time-step instance to proven optimality with a classical mixed-integer linear program; if the solver finishes in under a second, the paper's concluding claim that such networks are intractable for classical methods fails, and the only remaining claim is the narrower feasibility demonstration.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces entropy quantum computing and the Dirac-3 hardware, the device used for all sampling in this paper."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the integrated space logistics mission planning problem with spacecraft, ISRU, and scheduling that motivates this work."},{"cited_title":"W., and Ho, K., Space Transportation System and Mission Planning for Regular Interplanetary Missions , Journal of Spacecraft and Rockets, Vol","cited_arxiv_id":null,"evidence_quote":"Formulates space transportation as a network flow with multiple commodities, the model this paper time-expands and solves."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Presents the reinforcement-learning alternative that the quantum approach is positioned against for solving the same class of planning problems."}],"review_version":1}