{"id":"7b9863c2-0778-46f7-a44e-2cd500b8a125","arxiv_id":"2611.17047","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Weighted oriented paths and cycles with all sink/source vertices, and trees with a unique sink, have IDP polytopes, while certain weighted subgraphs obstruct IDP.","lead":"This paper studies edge polytopes built from vertex-weighted oriented graphs and asks when they have the integer decomposition property (IDP), meaning every lattice point in a dilated polytope splits into lattice points of the original. It proves new IDP results for paths, cycles, and trees with a unique sink, and it identifies forbidden subgraphs whose presence destroys the property.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek/deepseek-v4-flash","headline":"Converse of Theorem 5.11 relies on an unproven structural claim about balanced uniformly weighted cycles.","rationale":"The reader's weakest_assumption correctly identifies the unproven structural claim in the converse of Theorem 5.11. The paper's own sentence [b297] states the theorem, and the proof paragraph (starting 'We now prove the converse...') contains the unsupported assertion. The reader's concern is precise: a uniformly weighted 6-cycle with alternating sinks, sources, and mixed vertices can satisfy the balance equation without containing the odd-length Prop 4.2 shape as an induced subgraph. My analysis confirms this gap. The paper acknowledges no limitation here; the proof simply declares the structure without justification. This is the single most load-bearing concern because the 'if and only if' classification of cycles is a headline result, and the 'only if' direction is essential to the claim of classification. If this direction fails, the paper's strongest claim is weakened from a classification to a sufficient condition. The concrete test I propose would either falsify the converse or force the authors to provide the missing argument. I agree with the reader's assessment, and the verdict should remain CONDITIONAL because the gap is real but potentially fixable with additional reasoning. I do not recommend REJECT because the positive results (paths, unique-sink trees) are independent and valuable, and the 'if' direction of Theorem 5.11 is correctly proved. The paper's self-citations are appropriate and not circular. No ad hominem or theatrical language is used.","tokens_in":29874,"tokens_out":777,"duration_ms":7258,"concrete_test":"Construct a balanced uniformly weighted 6-cycle with vertices: x1 (sink), x2 (source), x3 (mixed, weight w), x4 (sink), x5 (source), x6 (mixed, weight w), edges oriented accordingly. Check whether the balance condition (Definition 5.10(b)) holds. If it does, compute the edge polytope and test IDP via Macaulay2 or by checking whether the point a = (1,1,1,1,1,1) (if w=2) lies in 2P but cannot be expressed as a sum of two lattice points. If such a cycle has IDP, the converse of Theorem 5.11 is false. If it does not have IDP, the authors must provide a general proof that any balanced uniformly weighted cycle with a non-sink, non-source vertex contains the Proposition 4.2 structure.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The 'only if' direction of Theorem 5.11 asserts that if a balanced uniformly weighted cycle has the IDP then all vertices are sinks or sources. The proof states: 'Since Cn is balanced and uniformly weighted, then the number of clockwise and counterclockwise vertices are the same, and all non-sink, non-source vertices cannot be all counter-clockwise or all clockwise, therefore Cn has the structure of Proposition 4.2.' This is not a derivation; it is an assertion. Proposition 4.2 describes a specific odd-length path-like structure (2r+1 vertices) with a particular alternating pattern of sinks and sources. A balanced uniformly weighted cycle with mixed vertices could have a different pattern (e.g., a 6-cycle with alternating sinks, sources, and mixed vertices) that satisfies the balance condition but does not contain the induced subgraph of Proposition 4.2. The paper provides no argument that the presence of a non-sink, non-source vertex forces the existence of that specific induced subgraph. Without this, the converse is unsupported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek/deepseek-v4-flash","summary":"The paper studies the integer decomposition property (IDP) of edge polytopes of vertex-weighted oriented graphs. It characterizes when such polytopes have the IDP by identifying forbidden induced subgraphs (Proposition 4.2, Theorem 4.5, Proposition 4.6) and proving positive results for several families: weighted oriented paths with all vertices sinks or sources under either uniform weight or coprime sink weights (Propositions 5.3, 5.7, Theorem 5.8), weighted oriented trees with a unique sink (Theorem 5.9), and balanced uniformly weighted cycles where all vertices are sinks or sources (Theorem 5.11). The main tool is the IDP criterion of Hà–Lin (Proposition 2.12) and explicit description of non-vertex lattice points (Section 3).","tokens_in":30209,"tokens_out":2076,"duration_ms":13774,"significance":"If correct, the results provide a near-classification of IDP for edge polytopes of weighted oriented graphs within the studied families. The identification of forbidden induced subgraphs and the positive results for trees and cycles are nontrivial extensions of prior work on 0-1 polytopes. The paper is clearly written and the proofs are largely detailed. The main significance lies in connecting combinatorial graph structure (sink/source patterns, coprimality) to the IDP, which is a central property in Ehrhart theory and toric geometry.","major_comments":[{"comment":"The converse (only if) direction of Theorem 5.11 is not proved. The proof states: 'Since Cn is balanced and uniformly weighted, then the number of clockwise and counterclockwise vertices are the same, and all non-sink, non-source vertices cannot be all counter-clockwise or all clockwise, therefore Cn has the structure of Proposition 4.2.' This is an assertion without derivation. Proposition 4.2 describes a specific odd-length path-like structure (2r+1 vertices) with a particular alternating pattern. A balanced uniformly weighted cycle with mixed vertices could have a different pattern (e.g., a 6-cycle with alternating sinks, sources, and mixed vertices) that satisfies the balance condition but does not contain the induced subgraph of Proposition 4.2. The paper provides no argument that the presence of a non-sink, non-source vertex forces the existence of that specific induced subgraph. Without this, the converse is unsupported. This is a load-bearing gap because the 'if and only if' claim is a central result.","section":"Theorem 5.11 (converse direction, proof paragraph)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The proof of Theorem 4.5 for even n is incomplete. The text says: 'The proof is similar to the odd case, and the details are left to the reader.' Given that the odd case already involves nontrivial sign analysis and the even case has a different sign pattern (det(C_n) negative), leaving the verification to the reader is insufficient for a rigorous proof. The claim that the even case follows similarly is not justified, especially because the sign of the determinant changes the direction of inequalities. This gap affects the classification of forbidden structures for even cycles, which is used in the overall characterization.","section":"Theorem 4.5, even n case","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Proposition 3.10 claims that for a naturally oriented cycle with uniform weight w≥2, the non-vertex lattice points are exactly the set L = {λ∈N_0^n | λ_i ≥ 1, |λ| = w+1}. The proof uses the inverse matrix C_n^{-1} and argues that a = C_n^{-1} λ has nonnegative entries and sum 1. However, the argument that a_i ≥ 0 for all i relies on bounding inequalities that assume n≥3 and w≥2, but the derivation of the bounds (e.g., λ_i ≤ w+2-n) is not fully justified. Moreover, the proof that every λ in L is indeed a lattice point of P_{C_n} requires showing that a_i ≥ 0, but the inequalities given are not rigorous for all possible λ. For example, the bound λ_i ≤ w+2-n is derived from |λ|=w+1 and λ_i ≥ 1, but this is only valid if n≥3; for n=3, w+2-n = w-1, which is fine, but the argument that a_1 ≥ 0 uses a sum that may not be nonnegative for all choices of λ. The proof appears to assume a particular ordering of λ_i that may not hold. This proposition is used in Theorem 4.5, so its correctness is important.","section":"Proposition 3.10, characterization of non-vertex lattice points for naturally oriented cycles","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The proof of Proposition 5.3 reduces the system to independent subsystems (6) and claims that nonnegative integer solutions exist based on the values T_l. The construction of solutions when T_l = k_l w + t_l with 0 < t_l < w sets d_{l,t}=1 and others 0. However, it is not verified that the resulting d_{l,t} are integers between 0 and w-1 (they are 0 or 1, which is fine) and that the corresponding c_{2l-1}, c_{2l} are nonnegative integers. The argument that a_{2l} - k_l - 1 ≥ 0 is not explicitly justified; it relies on the inequality a_{2l} ≥ k_l + 1, which follows from T_l < w a_{2l}? The proof states 'If instead T_l = k_l w + t_l with 0 ≤ k_l < a_{2l} and 1 ≤ t_l < w', but it is not shown that k_l < a_{2l} always holds. This needs clarification.","section":"Proposition 5.3, proof using row reduction and subsystems","authors_acknowledge":null,"authors_acknowledge_p":null}],"minor_comments":[{"comment":"The paper uses the term 'normal' interchangeably with IDP, but Remark 2.9 notes that some authors distinguish them. It would be helpful to state explicitly which definition is adopted (IDP as defined in Definition 2.8(b)).","section":"General","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The determinant formula det(C_n) = 1 + (-1)^{n-1} w^n is given without proof. While it is straightforward, a brief derivation or reference would improve readability.","section":"Lemma 2.6","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The statement says 'with weight w_1 on vertex x_1 and w_3 on vertex x_3'. It might be clearer to explicitly state that the path is x_2 → x_1 and x_2 → x_3, with vertex x_2 having weight 1 (or unspecified). The weights on x_2 are not mentioned; presumably they are 1 (since it is a source with out-edges). This should be clarified.","section":"Proposition 3.2","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The example compares H and H' with different orientations and claims that H' has IDP based on Macaulay2 computation. It would be helpful to state whether this is a general phenomenon or just an example.","section":"Example 4.4","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The statement assumes 'P_H has no non-vertex lattice points' but the proof does not explicitly verify this condition; it relies on the structure and coprimality conditions. The condition should be stated as part of the hypothesis or justified.","section":"Proposition 4.6","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The proof is very brief: it argues that because G has a unique sink, there are no non-vertex lattice points, and then uses a leaf elimination argument to show that any a in A[P_G] must have c_i = 0. However, the step 'we can pick another leaf of G \\ {e_1} and continue this process to conclude that c_i = 0 for all i' is not fully explained. It assumes that after removing a leaf, the remaining graph still has a unique sink? This needs more justification.","section":"Theorem 5.9","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The proof uses Lemma 5.5 which requires pairwise coprime weights. The statement says 'all the weights of the sink vertices are pairwise coprime'. It should be clarified whether the source vertices also have weights that are coprime to each other or not; the proof only uses coprimality of sink weights (odd-indexed vertices).","section":"Proposition 5.12","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The paper uses both N and N_0; it would be helpful to define them explicitly at the start.","section":"Notation","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Reference [6] is Hà and Lin, 'Normal 0-1 polytopes', which is published. The paper cites it for Proposition 2.12 and Theorem 4.1. This is appropriate.","section":"References","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"major_revision"},"author_rebuttal":null,"desk_editor":{"model":"deepseek/deepseek-v4-flash","letter":"This paper extends the IDP (integer decomposition property) literature from 0-1 polytopes and unweighted edge polytopes to vertex-weighted oriented graphs. The main contributions are: (1) a description of non-vertex lattice points in these polytopes (Section 3), (2) a catalogue of forbidden induced subgraphs (Section 4), and (3) positive IDP results for weighted oriented paths, unique-sink trees, and balanced uniformly weighted cycles (Section 5). The positive theorems for paths and unique-sink trees are proved in detail and appear correct. The paper is clearly written and the examples are helpful.\n\nThe soft spot is Theorem 5.11, the balanced-cycle classification. The 'if' direction (all vertices sinks or sources implies IDP) is argued via a reduction to the path case and a linear algebra argument that looks plausible. The 'only if' direction, however, is not proved. The proof states: 'Since Cn is balanced and uniformly weighted, then the number of clockwise and counterclockwise vertices are the same, and all non-sink, non-source vertices cannot be all counter-clockwise or all clockwise, therefore Cn has the structure of Proposition 4.2.' This is an assertion, not a derivation. Proposition 4.2 describes a specific odd-length path-like structure. A balanced uniformly weighted 6-cycle with alternating sinks, sources, and mixed vertices satisfies the balance condition but does not contain the induced subgraph of Proposition 4.2. The paper provides no argument that a non-sink, non-source vertex forces that specific induced subgraph. Without this, the converse is unsupported. The authors should either supply a proof or downgrade the claim to a conjecture.\n\nMinor issues: Theorem 4.5 (naturally oriented cycles) leaves the even-n verification to the reader, which is acceptable for a journal but should be checked. The abstract says 'classify' but the classification is partial (paths, cycles, unique-sink trees). That is fine for a paper, but the abstract overstates slightly.\n\nWho this is for: researchers in algebraic combinatorics, Ehrhart theory, and toric geometry. The paper is a solid step within its subfield, not a breakthrough. It deserves a serious referee: the positive results are new and the gap in Theorem 5.11 is fixable. I would accept peer review and recommend the editor send it out, with the expectation that the authors address the converse of Theorem 5.11.","headline":"Solid IDP classification for weighted oriented graphs, but the converse of the balanced-cycle theorem is not proved.","tokens_in":30719,"tokens_out":591,"duration_ms":6490},"pith_extraction":{"msc":["52B20","05C31","15B36","52C07","13F65","05C20"],"pacs":[],"model":"deepseek/deepseek-v4-flash","headline":"This paper classifies when edge polytopes of weighted oriented graphs have the integer decomposition property (IDP).","keywords":["integer decomposition property","edge polytope","weighted oriented graph","normal polytope","lattice polytope","forbidden induced subgraph","sink","source"],"falsifier":"Construct a uniformly weighted 6-cycle with vertices weighted 2, alternating sinks, sources, and mixed vertices (e.g., clockwise, counter-clockwise, sink, source, clockwise, counter-clockwise) such that the balance condition holds (even number of clockwise and counter-clockwise vertices) but the induced subgraph of Proposition 4.2 does not appear. If such a cycle's edge polytope still fails IDP, the 'only if' direction of Theorem 5.11 would be supported; if it has IDP, the theorem's claim that IDP requires all vertices to be sinks or sources would be false.","tokens_in":31358,"feed_emoji":"","tokens_out":1475,"duration_ms":12021,"temperature":0.7,"pith_summary":"The paper studies edge polytopes associated to vertex-weighted oriented graphs and asks when these polytopes have the integer decomposition property (IDP)—meaning every lattice point in any dilation can be expressed as a sum of lattice points from the original polytope. The authors identify forbidden induced subgraphs whose presence destroys the IDP, and they prove positive results for several families: weighted oriented paths and cycles whose vertices are all sinks or sources (with either uniform weights or coprime sink weights), and any weighted oriented tree with a unique sink. The central claim is that for a balanced uniformly weighted cycle, the IDP holds if and only if every vertex is a sink or source.","feed_headline":"Edge polytopes of weighted oriented graphs: when do they have the IDP?","feed_subtitle":"Forbidden subgraphs and sink/source conditions classify the integer decomposition property in several families.","key_machinery":"The central object is the edge polytope P_G of a weighted oriented graph G, defined as the convex hull of the columns of its incidence matrix. The key machinery includes the characterization of non-vertex lattice points in P_G (Proposition 3.1, Lemma 3.5, Proposition 3.10), the identification of forbidden induced subgraphs that prevent IDP (Proposition 4.2, Theorem 4.5, Proposition 4.6), and the use of row reduction and combinatorial arguments to show that certain families (paths, cycles, trees with a unique sink) satisfy IDP.","core_discovery":"The paper establishes that for edge polytopes of vertex-weighted oriented graphs, the integer decomposition property (IDP) is characterized by the absence of certain induced subgraphs and, in the case of uniformly weighted balanced cycles, by the condition that every vertex is either a sink or source. It provides positive results for paths and cycles under sink/source orientations and coprimality or uniform weight conditions, and proves that any weighted oriented tree with a unique sink has the IDP.","pith_inferences":["The 'only if' direction of Theorem 5.11 depends on the assertion that a balanced uniformly weighted cycle with a non-sink, non-source vertex must contain the induced subgraph of Proposition 4.2. This structural claim is not fully justified in the paper; if it fails, the classification would be incomplete.","The results suggest that the IDP property for edge polytopes of weighted oriented graphs is tightly controlled by the orientation pattern (sink/source structure) and the coprimality of weights, rather than by the overall graph topology. One could test this by constructing a cycle with mixed orientations that is not balanced but still has IDP.","The paper's approach of characterizing non-vertex lattice points and then using them to prove IDP (or its failure) could be extended to other families of polytopes arising from monomial maps, such as those from hypergraphs or directed graphs with multiple edge types.","The explicit construction of a point in A[P_H]\\K[P_H] in Proposition 4.2 provides a concrete witness for the failure of IDP, which could be used to algorithmically check IDP for small graphs."],"forward_implications":["If correct, these results provide a near-complete classification of when edge polytopes of weighted oriented graphs have the IDP, extending previous work on 0-1 polytopes and unweighted graphs.","The forbidden subgraph characterization can be used to test whether a given weighted oriented graph yields an IDP polytope by checking for the presence of the described induced subgraphs.","The positive results for trees with a unique sink and for paths/cycles with sink/source orientations give explicit families of IDP polytopes, which may be useful in Ehrhart theory and integer programming.","The coprimality conditions (Proposition 5.7, Proposition 5.12) show that IDP can hold even when non-vertex lattice points are absent, relying only on number-theoretic properties of the weights.","The balanced-cycle classification (Theorem 5.11) gives a necessary and sufficient condition for IDP in a natural family, showing that the presence of any vertex that is neither sink nor source destroys IDP."],"supporting_citations":[{"why":"Provides the criterion (Proposition 2.12) used to determine IDP via lattice points, originally for 0-1 polytopes but extended here.","marker":"[6]"},{"why":"Background on Oda's conjecture and the difficulty of determining IDP.","marker":"[1]"},{"why":"Establishes the connection between graph structure and polyhedral geometry for unweighted graphs, which this paper extends to weighted oriented graphs.","marker":"[12]"},{"why":"Used for computational verification of examples (Macaulay2).","marker":"[5]"},{"why":"Provides the definition of the monomial edge ideal that generates the toric ideal studied here.","marker":"[2]"},{"why":"Background on toric edge rings of unoriented graphs, which motivates the definition for oriented graphs.","marker":"[11]"}],"fun_headline_variants":["Which weighted oriented graphs yield IDP edge polytopes?","Forbidden subgraphs for IDP in edge polytopes","Weighted oriented trees with unique sink have IDP","Vertex weights and edge orientations meet IDP classification"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The proof that a balanced uniformly weighted cycle with a non-sink, non-source vertex must contain the specific induced subgraph of Proposition 4.2 is stated without derivation, and this structural claim is not forced by the balance condition alone.","fun_headline_variants_meta":{"raw":{"variants":["Which weighted oriented graphs yield IDP edge polytopes?","Forbidden subgraphs for IDP in edge polytopes","Weighted oriented trees with unique sink have IDP","Vertex weights and edge orientations meet IDP classification"]},"model":"deepseek/deepseek-v4-flash","effort":"low","cost_usd":6.7e-05,"raw_usage":{"cost":6.69942e-05,"is_byok":false,"cost_details":{"upstream_inference_cost":6.69942e-05,"upstream_inference_prompt_cost":5.49822e-05,"upstream_inference_completions_cost":1.2012e-05},"total_tokens":875,"prompt_tokens":809,"completion_tokens":66,"prompt_tokens_details":{"audio_tokens":0,"video_tokens":0,"cached_tokens":256,"cache_write_tokens":0},"completion_tokens_details":{"audio_tokens":0,"image_tokens":0,"reasoning_tokens":0}},"tokens_in":809,"tokens_out":66,"cost_source":"openrouter_usage","duration_ms":2574,"image_pages":[],"temperature":1.0,"billed_attempts":1,"reasoning_tokens":0,"transport_provider":"openrouter","cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-11-13T07:29:57.294856+00:00","model_set":{"reader":"deepseek/deepseek-v4-flash"},"pipeline_version":"pith-pipeline@v1.4.1-daily-deepseek-or-medium","input_kind":null,"input_chars":null,"input_total_chars":null,"passes_missing":null,"published_standing":null,"published_confidence":null,"objection_audit":null,"run_record":null,"basis":{"kind":"not_recorded","source":null,"chars_given":null,"chars_total":null},"receipt":null,"falsifier":"Construct a uniformly weighted 6-cycle with vertices weighted 2, alternating sinks, sources, and mixed vertices (e.g., clockwise, counter-clockwise, sink, source, clockwise, counter-clockwise) such that the balance condition holds (even number of clockwise and counter-clockwise vertices) but the induced subgraph of Proposition 4.2 does not appear. If such a cycle's edge polytope still fails IDP, the 'only if' direction of Theorem 5.11 would be supported; if it has IDP, the theorem's claim that IDP requires all vertices to be sinks or sources would be false.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the criterion (Proposition 2.12) used to determine IDP via lattice points, originally for 0-1 polytopes but extended here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Background on Oda's conjecture and the difficulty of determining IDP."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the connection between graph structure and polyhedral geometry for unweighted graphs, which this paper extends to weighted oriented graphs."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Used for computational verification of examples (Macaulay2)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the definition of the monomial edge ideal that generates the toric ideal studied here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Background on toric edge rings of unoriented graphs, which motivates the definition for oriented graphs."}],"lane_accuracy":{"status":"not yet measured","lane":{"model":"deepseek/deepseek-v4-flash","input_kind":null},"url":"https://pith.science/accuracy"},"review_version":1}