{"paper_id":"2610.01585","title":"Distribution-constrained maximum stopping of maximum type","authors":["Shuoqing Deng","Xin Zhang"],"paper_date":"2026-10-01","paper_type":"theoretical","paper_url":"https://arxiv.org/pdf/2610.01585","page_url":"https://thequant.space/flowcharts/distribution-constrained-maximum-stopping-of-maximum-type/","scores":{"empirical_rigor":2,"math_complexity":9,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper characterizes the optimal stopping boundary for the distribution‑constrained stopping problem as the solution of a variational inequality and provides a novel probabilistic representation of this inequality as a time‑reversed optimal stopping problem.","quote":"In this paper, we characterize this boundary by a variational inequality. In the spirit of Cox et al. (PTRF, 2019), we provide a novel probabilistic representation for the variational inequality as a time-reversed optimal stopping problem.","summary":"We consider the distribution-constrained optimal stopping problem _τ~ μ E[B^*_τ], where μ is a probability distribution on R_+, and (B^*_t) denotes the running maximum of a standard Brownian motion. This problem was introduced in Beiglbock et al. (PTRF, 2018), where a monotonicity principle is used…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The target distribution µ is a probability distribution on ℝ₊ and its survival function p(t)=µ((t,∞)) is continuous, non‑increasing with p(0)=1.","type":"data","stated":"explicit","quote":"We consider the distribution-constrained optimal stopping problem $\\sup_{τ\\sim μ} \\mathbb E[B^*_τ]$, where $μ$ is a probability distribution on $\\mathbb R_+$, and $(B^*_t)$ denotes the running maximum of a standard Brownian motion. Let p : [0, ∞) →[0, 1] be continuous and nonincreasing, with p(0) = 1.","support":"full"},{"assumption":"The boundary function b(t) is smooth, strictly negative, upper‑semicontinuous and satisfies b(t)=b⁻(t) (no upward jumps).","type":"model","stated":"explicit","quote":"2.1.1 The PDE formulation In this formal derivation, we assume that b is smooth and strictly negative and that the law of Yt restricted to {τb > t} has a density u, namely P(τb > t, Yt ∈dx) = u(t, x) dx, for b(t) < x < 0. Suppose b : (0, ∞) →[−∞, 0] is upper semicontinuous and, for 0 < t ≤T, b(t) = b−(t).","support":"full"},{"assumption":"The density u(t,x) of the law of Yₜ on the continuation region exists and satisfies the forward heat equation with Neumann condition at 0 and Dirichlet condition at the moving boundary.","type":"model","stated":"explicit","quote":"It follows that the density function u satisfies the forward equation  ut = 1 2uxx, b(t) < x < 0, ux(t, 0) = 0, u(t, b(t)) = 0, t > 0, u(0, ·) = δ0. (2.1) 3 For b(t) < x ≤0, define function F(t, x) := P(τb > t, Yt > x) = R 0 x u(t, y) dy; and for x ≤b(t) let F(t, x) = P(τb > t).","support":"full"},{"assumption":"The initial and boundary conditions are incompatible at the origin (0,0), i.e., Y₀=0 while F(0,x)=1 for x<0 and F(t,0)=0 for t>0.","type":"other","stated":"explicit","quote":"(2.3) The initial and boundary conditions are incompatible at (0, 0) because Y0 = 0. This motivates the following variational inequality formulation of the classical free-boundary problem: max ( Ft −1 2Fxx, F −p(t) ) = 0, for all x < 0, t > 0, (2.4) with the initial and boundary conditions F(0, x) = 1, F(t, 0) = 0, for all x < 0, t > 0.","support":"full"},{"assumption":"Viscosity solutions are sought in the class of functions satisfying p(t)H(t,x) ≤ U(t,x) ≤ min{p(t),H(t,x)}.","type":"model","stated":"explicit","quote":"Motivated by the above lemma, we will restrict in the next section the admissible sub- and supersolu-tions U to satisfy p(t)H(t, x) ≤U(t, x) ≤min{p(t), H(t, x)}. In the comparison argument below, when we consider sub- and supersolutions u and v satisfying satisfy the above bounds, it will naturally imply that u(t, x) −v(t, x) ≤(1 −p(t))H(t, x) ≤1 −p(t) →0 (2.10) 5 as (t, x) →(0, 0).","support":"full"},{"assumption":"The stopping time τ̂ defined by the constructed boundary satisfies P(τ̂>0)=1 (no atom at zero).","type":"model","stated":"explicit","quote":"Assume also P(τb > 0) = 1. Moreover, P(τb = t) = 0 for 0 < t ≤T, and P(τb > t) = lim x→−∞Gb(t, x) = Gb(t, b(t)).","support":"full"}]},"datasets":{"status":"none_found","items":[],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"none"},"transaction_costs":{"status":"not_applicable","modeled":"n/a","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R3","candidates":{"front_matter":"n/a","evidence_card":"n/a","evidence_card_quoted":false},"meaning":"the paper reports no trading or execution result"},"results":{"status":"available","evaluation_type":"none","findings":[{"finding":"The variational inequality (2.4) admits a unique viscosity solution, establishing well‑posedness of the free‑boundary problem.","kind":"theoretical","quote":"First, we analyze the wellposedness of a vari-ational inequality which determines the boundary of the inverse first passage problem. 2.2 Wellposedness of the variational inequality In this section, we justify the existence and uniqueness of viscosity solution to the system (2.4)–(2.5).","support":"full"},{"finding":"The constructed boundary ˆb(t) is upper‑semicontinuous and yields a stopping time τ̂ whose law matches the target distribution µ (i.e., τ̂∼µ).","kind":"theoretical","quote":"Let p : [0, ∞) →[0, 1] be continuous and nonincreasing, with p(0) = 1. Then P(τˆb > t) = p(t), t ≥0.","support":"full"},{"finding":"The hitting time τ̂ is optimal among all stopping times with law µ, i.e., for any ρ∼µ, E[M_ρ] ≤ E[M_{τ̂}].","kind":"theoretical","quote":"For every stopping time ρ ∈T µ, E[Mρ] ≤E[Mˆτ]. Consequently, sup ρ∈T µ E[Mρ] = E[Mˆτ], where the value +∞is allowed.","support":"full"},{"finding":"A discrete‑time Snell envelope construction together with a stability argument provides an alternative proof of optimality without requiring a moment assumption on µ.","kind":"theoretical","quote":"We first work in discrete time and add a deterministic function of the stopping time to the reward. We verify optimality using the Snell envelope and then pass to continuous time through the stability of the associated stopping times.","support":"full"},{"finding":"Quantitative bounds near the singular corner (0,0) allow comparison of sub‑ and supersolutions without assigning a value at the origin.","kind":"theoretical","quote":"A key ingredient for proving the viscosity solution property and comparison principle is a quantitative estimate near the singular corner of the time-space domain, where the initial and boundary conditions are incompatible. We address this issue by es-tablishing quantitative bounds that determine an admissible solution class and allow comparison without assigning a value at the origin (0, 0).","support":"full"}],"reported_performance":null,"out_of_sample":{"method":"none","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":24,"citations":{"cited_by_in_corpus":0,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":0},"independent_papers":[],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":false,"code_url":null,"note":"No repository presented by the authors as their own was found in the paper. Absence here is not proof that no code exists."},"replication":{"status":"none_found","card":null,"note":"No replication of this paper has been run by thequant.space yet."},"extraction":{"text_source":"fulltext","model":"openai/gpt-oss-120b","version":2,"extracted_at":"2026-11-14T19:39:10Z"},"provenance":{"claims":"Everything here is what the authors state in the paper. Nothing was re-run or reproduced by the platform.","quotes":"Assumptions, datasets, cost detail and findings were extracted by a language model. Every quote is verbatim paper text. Where an item has `support`, \"full\" means every figure in it appears in the quoted sentences and \"partial\" means a figure is stated elsewhere in the paper; items with a figure found nowhere in the paper were discarded.","datasets_known":"datasets.known lists well-known vendors and public datasets by a stable id, found by a dictionary over the full text where the sentence says the data is used; it is independent of the model and covers only names in the dictionary.","flags":"transaction_costs.modeled, results.evaluation_type and results.out_of_sample.method are platform classifications (structured-decision models); see https://thequant.space/score-guide/#penalty-flags","comparisons":"supports / contradicts / qualifies are a language model's reading of two papers' quoted findings, restricted to closely related papers that share no author. They are pointers to read, not verified replications, and coverage is partial.","implementation":"code_url is a repository the paper itself presents as its own (rule-scored, ambiguous cases judged by a model), checked for liveness when it was recorded."},"docs":"https://thequant.space/agents/"}