{"id":"43cb8f8f-f978-4894-869b-05339191e4b7","arxiv_id":"2607.09568","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":4,"one_line_summary":"PUME couples a perturbed-utility Markov route-choice model (allowing zero-probability links) with a cost-space variational inequality that handles asymmetric link costs, plus globally convergent loading and equilibrium algorithms.","lead":"This paper builds a traffic-equilibrium model that lets drivers skip bad roads entirely and still scales to large networks with messy congestion interactions. It matters because city planners and routing systems need models that are both realistic about choice and fast enough to run on real maps.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified","rationale":"The central claim rests on three pillars: (i) well-posedness of the undiscounted PUMCM under Standing Assumptions 1–2, (ii) monotone dual VI existence/uniqueness under coercive strictly monotone supply, and (iii) global convergence of MPI and the merit-safeguarded meta-algorithm. Pillar (i) is the only non-standard modeling hypothesis; the paper treats it carefully (Remark 1 + numerical counter-example). Pillars (ii)–(iii) follow from classical monotone-operator and policy-iteration arguments once (i) is granted, and the deferred proofs in Appendices A–C are complete. Experiments are consistent with the theory and cover both potential and non-potential supply. Consequently the reader's ACCEPT / HIGH-confidence verdict stands; no adjustment is warranted.","tokens_in":39615,"tokens_out":418,"duration_ms":5916,"concrete_test":"Re-run the stylized-network experiment of §6.2 with a non-zero but arbitrarily small positive stage surplus on the zero-cost cycle (e.g., Hs = 10^{-8}); confirm that value iteration still diverges while the translated surplus of Remark 1 restores linear MPI convergence. This verifies that the well-posedness hinge is both necessary and correctly repaired.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's weakest assumption (Standing Assumption 2(B2): strictly negative stage surplus) is correctly identified as the hinge of well-posedness (Proposition 1, Theorems 4–5). The paper itself flags the assumption, supplies a translation fix that leaves the choice map invariant (Remark 1), and demonstrates divergence when it is violated (Figure 3). Under the maintained conditions the dual-VI existence/uniqueness theorems, MPI global convergence, and safeguarded meta-algorithm all follow by standard arguments that are fully written out. No internal inconsistency or hidden gap that would overturn the strongest claim was found.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The paper develops a perturbed utility Markovian equilibrium (PUME) for large-scale traffic assignment. Route choice is modeled as an undiscounted absorbing MDP regularized by a convex surplus function (PUMCM), so that the Bellman optimality operator is the surplus of the state-action values and its gradient is the optimal policy. Under Standing Assumptions 1–2 the MDP is well posed, admits corner solutions, and induces a continuous monotone demand map. Equilibrium is cast as a dual variational inequality on the cost space that allows non-separable and asymmetric supply; existence and uniqueness follow from coercivity and strict monotonicity of supply. Network loading is solved by modified policy iteration with a global convergence guarantee and a local linear rate controlled by evaluation depth; the outer VI is solved by a safeguarded meta-algorithm that preserves global convergence for any acceleration oracle. Numerical experiments on Sioux Falls, Anaheim, Chicago Sketch and synthetic grids illustrate scalability and robustness across surplus families and potential/non-potential supply.","tokens_in":39812,"tokens_out":1050,"duration_ms":11463,"significance":"If the claims hold, the paper supplies a single link-based framework that simultaneously (i) admits zero-flow links without ex-ante choice-set restrictions, (ii) handles undiscounted network loading with a convergent algorithm, and (iii) accommodates asymmetric multi-class cost interactions via a dual VI. These three extensions address well-known limitations of existing Markovian traffic equilibrium models. The theoretical package is complete: well-posedness (Props. 1–3), existence/uniqueness (Thms. 1–2), MPI global and local rates (Thms. 4–5), and meta-algorithm convergence (Thm. 6) are proved with standard convex-analysis and monotone-operator tools, and the appendices contain the deferred arguments. The numerical section shows linear scaling with problem size and stable performance across demand and supply specifications. The contribution is therefore both theoretically coherent and practically relevant for large-scale assignment.","major_comments":[{"comment":"Standing Assumption 2(B2) (strictly negative stage surplus) is load-bearing for well-posedness (Prop. 1) and for the MPI guarantees (Thms. 4–5). Remark 1 correctly notes that a constant translation of stage rewards leaves the choice map invariant and can enforce the condition, and Figure 3 demonstrates divergence when it is violated. The manuscript should state more explicitly, preferably already in the introduction or in §3.3, that the translation is always available under the usual sign convention of traffic costs and that the subsequent theory is therefore not restricted by the assumption in applications. Without that clarification a reader may overestimate the restrictiveness of (B2).","section":null},{"comment":"Theorem 3 establishes equivalence of the dual (cost-space) and primal (flow-space) formulations only for interior equilibria. Remark 3 acknowledges that boundary costs require a constrained inverse supply, but the paper never develops that object or the corresponding VI. Because PUMCM is expressly designed to produce corner solutions, boundary equilibria are not pathological; a short statement of the constrained-inverse VI (or an explicit deferral with a pointer to future work) would close the gap between the modeling ambition and the equivalence result.","section":null}],"minor_comments":[{"comment":"Table 1 claims that α-entmax for α∈(1,2) yields a C² surplus and a Lipschitz choice map; the argument in Appendix D.3 is correct but terse. A one-sentence reminder that the power 1/(α−1)>1 implies C¹ of the thresholded map would help readers who skip the appendix.","section":null},{"comment":"In §5.2 the ST line-search condition (25) and the aGRAAL step-size rule (27) are stated without recalling the precise references for global convergence under mere monotonicity versus local Lipschitz continuity. Adding the theorem numbers from Solodov–Tseng and Malitsky would make the claims self-contained.","section":null},{"comment":"Figure 5 uses ε for coupling strength while the surrounding text uses ι; the notation should be unified.","section":null},{"comment":"The synthetic-grid generation formula (99) appears only in the appendix; a brief mention in §7.3 that demand follows a gravity model with free-flow shortest-path impedance would improve reproducibility of the main text.","section":null},{"comment":"A few typographical slips remain (e.g., “aperturbed”, “the the”, inconsistent spacing around γ=1). A careful proof-reading pass is warranted.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The manuscript is already close to acceptance quality. The two major comments are clarifications rather than technical repairs; either can be handled in a short revision. I see no novelty or citation issues that would affect the editorial decision."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"This is the paper that finally makes Markovian traffic equilibrium usable without the three usual crutches: full-support probabilities, discounting, and symmetric costs. The core is PUMCM—a surplus-function Bellman operator whose gradient is the policy—so corner solutions appear endogenously and unattractive links can carry zero flow. They couple it to a dual VI on the cost space that only needs coercive strictly monotone supply; existence and uniqueness follow by standard monotone-operator arguments (Theorems 1–2). That is the real advance over Baillon–Cominetti and the later GEV-MTE papers.\n\nWhat they do well is the technical hygiene. Standing Assumptions 1–2 plus the Fenchel–Young structure give well-posedness of the undiscounted MDP even at the boundary (Prop. 1–3). Modified policy iteration is shown globally convergent for any fixed evaluation depth m ≥ 1, with a local linear rate controlled by ||P_π*||^m (Thm. 4–5)—the first such analysis I have seen for corner-solution undiscounted PUMCM. The outer meta-algorithm (base VI solver + Anderson/NGMRES with merit safeguard) inherits global convergence from the base method alone (Thm. 6). Appendices contain the deferred proofs; they look complete. Numerics on Sioux Falls, Anaheim, Chicago Sketch and synthetic grids are consistent: linear scaling, robustness across α-entmax / logit / NRL and potential / non-potential supply.\n\nSoft spots are minor and mostly flagged by the authors. The hinge is Standing Assumption 2(B2)—strictly negative stage surplus. If it fails you can get zero-cost cycles and the well-posedness proofs collapse; they show the divergence (Fig. 3) and give the translation fix that leaves the choice map unchanged (Remark 1). No code is released, so independent reproduction is only moderate. Algorithm parameters (m, η, τ, R, μ_s) affect speed, not the theorems. Self-citations are to their own prior PURC work and are appropriate.\n\nThis is for people who actually solve large-scale assignment or who care about the theory of Markovian route choice. It deserves a serious referee; I would accept it for peer review and would cite the dual-VI formulation and the MPI analysis myself.","headline":"Clean dual-VI Markovian equilibrium that actually allows zero flows and asymmetric costs, with global MPI and safeguarded acceleration proofs that hold up.","tokens_in":40413,"tokens_out":575,"would_cite":true,"duration_ms":7608,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["90C33","90B20","90C40"],"pacs":[],"model":"grok-4.5","headline":"A convex surplus turns sequential route choice into a well-posed Markov model that allows zero flow on bad links and solves large asymmetric traffic equilibria by a dual variational inequality.","keywords":["Markovian traffic equilibrium","perturbed utility","variational inequality","modified policy iteration","meta-algorithm","corner solutions","undiscounted MDP"],"falsifier":"Construct a network and surplus that satisfy all other hypotheses but admit a zero-cost cycle; if value iteration or modified policy iteration still converges to a unique finite value and the dual VI still has a unique solution, the necessity claim fails.","tokens_in":40505,"feed_emoji":"🚗","tokens_out":923,"duration_ms":9285,"temperature":0.7,"pith_summary":"Large-scale traffic assignment needs models that look realistic and still run on big networks. This paper replaces the usual random-utility assumption in Markovian route choice with a convex surplus function whose gradient is the optimal link-choice policy. The resulting perturbed-utility Markovian choice model admits both interior and corner probabilities, so unattractive links can carry exactly zero flow without any a-priori pruning of the network. Under a mild strict-negativity condition on stage surplus, every policy consistent with the Bellman equation reaches the destination, the optimal value and demand map are unique and monotone, and modified policy iteration converges globally even when the MDP is undiscounted. Equilibrium is then written as a monotone variational inequality on the dual cost space; existence and uniqueness follow from standard coercivity and strict monotonicity of supply alone, and the formulation covers non-separable and asymmetric cost interactions that fall outside classic potential-based Markovian models. A safeguarded meta-algorithm that mixes first-order VI steps with acceleration oracles is proved globally convergent and shown to scale linearly on benchmark and synthetic grids.","feed_headline":"Zero-flow routes and asymmetric costs in one Markov model","feed_subtitle":"A convex surplus yields corner solutions and a dual VI that scales on large networks.","key_machinery":"The surplus function Hs whose gradient is the optimal policy: it turns the Bellman optimality operator into a convex conjugate, forces every admissible policy to be proper once stage surplus is strictly negative, and yields a monotone demand map that is the gradient of a convex potential.","core_discovery":"Under two standing assumptions on the surplus and a coercive strictly monotone supply map, there exists a unique cost vector that solves the dual variational inequality of excess supply; the induced demand map is continuous and monotone, admits corner solutions, and can be computed by globally convergent modified policy iteration for any fixed evaluation depth.","pith_inferences":["The same surplus construction should transfer to other undiscounted sequential discrete-choice settings (dynamic discrete choice, inventory routing) where corner solutions and non-contractive Bellman operators appear.","Because the demand map is the gradient of a convex potential, standard sensitivity and comparative-static arguments for monotone operators become available for Markovian traffic equilibria.","The strict-negativity translation trick suggests a practical pre-processing step that can be applied to any existing recursive-logit codebase before switching to a sparse surplus."],"forward_implications":["Unattractive links receive exactly zero flow without any ex-ante choice-set restriction, giving endogenous consideration sets inside a Markovian model.","Asymmetric multi-class or spillover cost interactions can be treated inside a single dual VI without requiring a potential function.","Network loading remains globally convergent for the entire surplus family, including the sparse α-entmax maps that produce corner solutions.","The same safeguarded meta-algorithm works for any acceleration oracle once a continuous merit function is available, so future first-order or quasi-Newton schemes can be swapped in without re-proving global convergence."],"fun_headline_variants":["Markov model unifies zero-flow links with asymmetric costs","Convex surplus enables corner solutions in traffic equilibrium","Unique dual VI for PUME with general non-separable costs","Boundary choice probs and link interactions in one Markov eq","Scalable perturbed utility equilibrium admits zero link flows"],"cache_read_input_tokens":32896,"weakest_assumption_plain":"Stage surplus must be strictly negative at every non-terminal node (or made so by a constant translation that leaves choices unchanged); without it, zero-cost cycles can produce infinite or non-unique values and the whole well-posedness argument collapses.","fun_headline_variants_meta":{"raw":{"variants":["Markov model unifies zero-flow links with asymmetric costs","Convex surplus enables corner solutions in traffic equilibrium","Unique dual VI for PUME with general non-separable costs","Boundary choice probs and link interactions in one Markov eq","Scalable perturbed utility equilibrium admits zero link flows"]},"model":"grok-4.5","effort":"low","cost_usd":0.005008,"raw_usage":{"total_tokens":1440,"prompt_tokens":812,"num_sources_used":0,"completion_tokens":79,"cost_in_usd_ticks":50080000,"prompt_tokens_details":{"text_tokens":812,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":549,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":812,"tokens_out":79,"duration_ms":6136,"temperature":1.0,"reasoning_tokens":549,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-13T02:04:37.460538+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Construct a network and surplus that satisfy all other hypotheses but admit a zero-cost cycle; if value iteration or modified policy iteration still converges to a unique finite value and the dual VI still has a unique solution, the necessity claim fails.","supporting_citations":[],"review_version":1}