Pith. sign in

REVIEW 4 major objections 5 minor 3 cited by

Proof and More Variations of Bellman's Lost-in-a-forest Problem

T0 review · 4 major / 5 minor · reviewed 2026-08-02 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper claims a certified computational solution to Bellman's lost-in-a-forest problem by reducing it to a discrete traveling-salesman-with-neighborhoods problem and proving convergence.

desk verdict Theorem 8's certified bounds are identical by Theorem 5, so the main claim collapses; the one-sided convergence and numerical tables are the only reusable pieces. read the letter →

arxiv 2606.13987 v3 pith:SIIATGVD submitted 2026-06-12 math.OC

Zhipeng Deng This is my paper · ORCID
classification math.OC MSC 49K3049Q1052A40
keywords Bellman'slost-in-a-forestproblemtravelingsalesmanwithneighborhoodsdiscretegeometrycurvecoveringMoser'swormconvexhullcomputationaloptimizationescapepath
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Bellman's lost-in-a-forest problem asks for the shortest path that guarantees escape from a forest of known shape but unknown starting position and orientation. This paper attempts to prove that the problem can be recast as a finite traveling-salesman-with-neighborhoods problem on transformed copies of the forest boundary, solved with certified global optimization, and that the discrete optima converge to the true continuous optimum. If correct, this would turn a long-open geometry problem with only a handful of solved shapes into a computationally tractable approximation problem with rigorous error bounds. The paper also extends the framework to searching for two lines, closed paths, translation-only versions linked to Wetzel's unit-arc cover problem, and three-dimensional sphere inspection.

What carries the argument

The load-bearing construction is the rigid-motion transformed boundary B_g = Rot(-alpha)(B - s), which converts the adversarial minimax search over starting positions and orientations into a fixed constraint: the canonical path must intersect every B_g. Discretizing the parameter space into nested finite sets G_m turns this into a finite TSPN with neighborhoods S_h = B_{g_h}, and the polyline optimality lemma reduces the continuous curve problem to a permutation-based shortest-path problem. Monotone convergence and existence of minimizers are carried by compactness arguments via the Arzelà–Ascoli theorem and lower semicontinuity of path length.

What would settle it

Construct two candidate paths for a simple forest, such as a unit circle: one that has shorter total length but larger worst-case first-hitting time, and another that is longer in total length but hits every possible boundary earlier. If such a pair exists, the minimax escape problem and the total-length boundary-intersection problem cannot be equivalent, and the TSPN discretization would solve a different problem.

Watch

Extended reading notes

Core claim

The paper's central assertion is that every possible starting position and orientation of the hiker can be encoded as a rigidly transformed forest boundary, so that a canonical path escapes if and only if it intersects all transformed boundaries. It then claims that the discretized version of this boundary-intersection problem is exactly a TSPN instance, that the discrete optimal values form a nondecreasing sequence converging to the continuous optimum, and that solving the finite TSPN to zero MIP gap yields a two-sided certificate bracketing the true escape length within any prescribed tolerance. The same discretization is applied to several variants, producing numerical candidate paths for

Load-bearing premise

The central claim rests on passing from the minimax first-hitting-time objective (Eq. 1) to the total-length boundary-intersection problem (P); Theorem 1 establishes pointwise intersection equivalence but not that minimizing total length equals minimizing the worst-case distance traveled before first hitting the boundary.

Editorial extensions

If this is right

  • For any compact forest, a finite set of transformed boundary constraints yields a rigorous lower bound on the true escape length, and refining the grid gives a nondecreasing sequence of lower bounds converging to the continuous optimum.
  • Solving the discretized TSPN with a certified global optimizer would, in principle, produce an epsilon-certificate for the original Bellman problem, making it computationally accessible rather than purely theoretical.
  • The same machinery gives numerical candidate paths for two-line searches, closed escape paths, and three-dimensional sphere inspection, connecting these variations to known constants such as the Zalgaller sphere-inspection value and the 4π closed-curve result.
  • Adding a closing segment to the objective turns the problem into a Hamiltonian-path variant, so the framework covers closed search paths as well as open ones.
  • The translation-only variant matches Wetzel's unit-arc covering problem, giving a TSPN formulation for a problem previously treated separately from Bellman's.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the equivalence between minimax first-hitting time and total-length boundary intersection is made fully rigorous, the framework would make Bellman's problem computationally tractable for arbitrary polygons, not just lines and strips.
  • One could test the convergence rate numerically on shapes with known exact answers, using the monotone lower bounds to measure how quickly the gap closes as the grid refines.
  • The rigid-motion discretization generalizes to other transformation groups, so analogous 'search with unknown pose' problems—rotation-only, translation-only, or in higher dimensions—could be cast as finite TSPN instances.
  • Editorial caution: as printed, Theorem 8's upper-bound half is not derived and appears inconsistent with the monotone lower-bound conclusion of Theorems 5 and 7; the certified two-sided approximation therefore depends on an argument not present in the manuscript.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes a formalization of Bellman's lost-in-a-forest problem as a transformed-boundary intersection problem (P), discretizes the parameter space, proves that the discretized problem is equivalent to TSPN (Theorem 5), proves monotone convergence of discrete optima to the continuous optimum (Theorem 7), and claims a certified approximation scheme (Theorem 8). Additional sections present numerical results for searching two lines, closed-path variants, and 3D extensions, and discuss connections to Wetzel's problem and Moser's worm problem.

Significance. If the central theorems were correct, the paper would provide a rigorous computational solution to a classic unsolved minimax problem, a notable result. It also reproduces known benchmarks and collects numerical data. However, the certification theorem is internally inconsistent, and the equivalence between the original minimax objective and the intersection formulation is not proven. The paper therefore does not establish its headline claim.

major comments (4)
  1. [Theorem 8 (§3)] In Theorem 8 the upper bound is defined as L̄_m := OPT_TSPN(m). Theorem 5 states L*_m = OPT_TSPN(m), so L̄_m = L*_m. Theorem 7 shows L*_m ↑ L*, so L*_m ≤ L* and increases. A number cannot simultaneously be a lower bound and an upper bound unless L*_m = L* for every m, which is false for a generic finite sample. The asserted 'L̄_m ↓ L*' is equally incompatible with Theorem 7. No construction of a path feasible for all g∈G is given, so no upper-bound certificate exists. This invalidates the claimed ε-certificate.
  2. [Eq. (1) → (P), §3] The original minimax objective is a supremum of first-hitting lengths over starting configurations. The reformulation (P) minimizes total length of a curve intersecting every transformed boundary. Theorem 1 proves only pointwise equivalence of intersection; it does not show that the worst-case first hitting time equals the total length of a feasible sweep path. The proof of equivalence therefore stops short of the load-bearing step. Without a truncation/restart argument, the TSPN discretization need not solve Bellman's problem.
  3. [§2, convex-hull equivalence] The statement that 'intersecting all forest boundaries' is identical to the curve's convex hull containing the shape is asserted without proof. This premise is not used in the formal development of Section 3, where the constraint is d(f,B_g)=0 for all g. As stated, it is doubtful for bounded targets under pure translations, since a bounded curve cannot intersect every translate of a bounded set. This unproven assertion is load-bearing for the claimed connection to Moser's worm problem and universal covers.
  4. [Theorem 6] Theorem 6 is not a mathematical proof but a conditional statement: if an MIP solver returns zero gap, then the finite TSPN is solved exactly. As stated, it adds no content beyond assuming the solver is correct; in a proof of a general theorem it should be phrased as an assumption or a computational oracle.
minor comments (5)
  1. [Theorem 5 proof] The proof cites 'Lemma 2' when it means Lemma 4 (polyline optimality).
  2. [Eq. (6), §4] The analytical formula for the arc+tangent path lacks derivation; the expression 1/2 + √3/2 + π/12 + β/2 should be justified.
  3. [Eqs. (9), (11), (14)] The variables γ_i and ϕ_i are used without a precise definition of their feasible ranges, and the constraint 'mod 2π = [0,2π]' is not explained at the numerical level.
  4. [Section 7, Table 3] The Zalgaller 3D length is reported as both 11.605 and 10.605427999164 in the same passage without clear reconciliation.
  5. [Section 5] Weak Form IB is described informally; it lacks a formal definition and a theorem comparable to those in Section 3.

Circularity Check

1 steps flagged · score 8.0 of 10

Theorem 8's certified upper bound is the lower bound by Theorem 5, so the ε-certificate collapses by construction.

  1. self definitional [Section 3, Theorem 8 (Certified approximation scheme), proof defining L_m and L̄_m]
    "Define certified bounds: L_m := L*_m (exact TSPN_m lower bound), L̄_m := OPT_TSPN(m) (exact finite upper approximation). Then L_m ≤ L* ≤ L̄_m. Furthermore, L_m ↑ L*, L̄_m ↓ L*, as discretization is refined. Therefore L̄_m − L_m → 0."

    By Theorem 5, L*_m = OPT_TSPN(m) for every m. Thus the proof defines the upper bound L̄_m to be exactly the lower bound L_m. The asserted two-sided containment L_m ≤ L* ≤ L̄_m then reduces to L* = L*_m, which is not established for finite m; Lemma 3 and Theorem 7 give only L*_m ↑ L*. So the claimed 'upper approximation' is the same quantity as the lower bound, renamed; the error L̄_m − L_m is zero only by assuming the conclusion. No independent feasible path for the continuous problem is constructed, so the ε-certificate is circular by construction.

full rationale

The central two-sided certification claim of Theorem 8 is not an independent derivation: the paper's own Theorem 5 proves OPT_TSPN(m) = L*_m, and Theorem 8 defines L̄_m := OPT_TSPN(m) while L_m := L*_m. Hence the supposed upper bound is identical to the lower bound, and the claimed interval [L_m, L̄_m] collapses to a point. That is a specific, quotable reduction of the main 'certified approximation scheme' to a tautology: the upper bound is the lower bound by definition, and the asserted ε-error bound follows only if one already assumes L* = L*_m, the very equality the scheme is supposed to certify. This is the strongest circular step in the paper. The rest of the convergence argument (Theorem 7, monotone lower bounds) is largely self-contained and does not depend on the author's earlier paper; the external reproduction of Zalgaller's 3D value provides some independent grounding. The self-citation to [8] is present and used for the general formulation, but it is not the main source of the circularity. Other mathematical gaps, such as the unproven minimax-to-path-length equivalence and the unsupported convex-hull identity, are correctness risks rather than circular reductions and are not counted toward the score beyond their role in the invalid certificate.

Assumptions & free parameters 1 free parameters · 7 assumptions · 0 invented entities

The proof relies on standard compactness and semicontinuity results plus several unproven modeling choices: compactness of the starting region, feasibility of some finite-length path, exact MIP solvability, the convex-hull criterion, and the equality of the minimax hitting distance with total path length. The discretization sizes in the numerical sections are unspecified free algorithmic parameters.

free parameters (1)
  • Discretization sizes N, M and segment count k
    The numerical results require choosing the number of orientation samples N, position samples M, and the number of line segments in the candidate path; none are reported, and the 'shortest' values depend on them.
assumptions (7)
  • domain assumption B and R are nonempty compact subsets of R^2 (and R^3 in §7), so G=R×[0,2π] is compact.
    Used throughout §3 for compactness and Arzelà–Ascoli; the original Bellman starting region is typically the open interior of the forest, so compactness is an unstated modeling restriction.
  • domain assumption There exists at least one finite-length feasible curve for (P).
    Theorem 2 only proves existence of a minimizer conditional on a finite-length feasible curve existing; for arbitrary compact B and starting region R feasibility is not automatic (e.g., B a single point with R having interior).
  • standard math Arzelà–Ascoli theorem and lower semicontinuity of the arclength functional under weak W^{1,1} convergence.
    These are the analytical engines of Theorems 2 and 7.
  • standard math The map g↦B_g is Hausdorff-continuous for compact B, and distance functions to closed sets are continuous.
    Used to pass feasibility to the limit in Theorem 7.
  • ad hoc to paper Intersecting every transformed boundary is equivalent to the curve's convex hull containing the target shape.
    Assumed in §2 to motivate the TSPN reduction; no proof or specific citation is supplied, and the formal theorems in §3 do not use it.
  • ad hoc to paper The minimax worst-case escape distance equals the total length of a feasible path minimizing L(f).
    Needed to identify (P) with Bellman's original problem; Theorem 1 proves only pointwise intersection, not equality of objectives.
  • domain assumption The finite TSPN can be formulated exactly as a MIP and solved to zero gap by a global solver.
    Theorem 6 assumes an exact MIP formulation and a zero-gap solver certificate, but no MIP model, solver, or certificate is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Proof and More Variations of Bellman's Lost-in-a-forest Problem." pith.science (2026). https://pith.science/paper/SIIATGVD

@misc{pith2026260613987,
  author       = {Pith},
  title        = {Pith review of: Proof and More Variations of Bellman's Lost-in-a-forest Problem},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SIIATGVD}},
  note         = {Machine review of arXiv:2606.13987}
}
read the original abstract

In this paper, based on our previous general formulation and computational solution to Bellman's Lost-in-a-forest Problem, we provide the proof of general solution and obtained more variations and results related to this problem. This paper provides generalized formalized method connecting curve covering, lost-in-the-forest problem, and traveling salesman problem with neighborhoods. We prove the equivalence and convergence. We also provide more results of searching for two lines, connection to Wetzel's unit arc covering problem, variations with closed path, variations in three dimensions, etc. The results include general calculation equations, partial analytical results, and numerical results.

Figures

Figures reproduced from arXiv: 2606.13987 by the authors.

Figure 1
Figure 1. Length of optimal path in various angles [PITH_FULL_IMAGE:figures/full_fig_p012_1.png] view at source ↗
Figure 2
Figure 2. More examples of detailed results of searching for two lines with from the [PITH_FULL_IMAGE:figures/full_fig_p014_2.png] view at source ↗
Figure 3
Figure 3. Weak Form IB - Starting from unknown points. We convert the path out [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Length of optimal closed path in various angles [PITH_FULL_IMAGE:figures/full_fig_p018_4.png]
Figure 5
Figure 5. Figure 5: More examples of detailed results of searching for two lines with closed path [PITH_FULL_IMAGE:figures/full_fig_p020_5.png]
Figure 6
Figure 6. Figure 6: Reproduced non-closed curve originating from the origin whose convex hull [PITH_FULL_IMAGE:figures/full_fig_p023_6.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. The exact solution of Bellman's lost-in-a-forest problem for the golden gnomon

    math.MG 2026-07 accept novelty 7.5 of 10 partial

    The shortest guaranteed escape path in the golden gnomon has exact length C=1.282676… and is a symmetric seven-piece segment–arc–tangent path.

  2. Universal Triangle Covering Curve and Polygonal Chain: Escaping Forest and Fitting Worm

    math.OC 2026-08 conditional novelty 7.0 of 10

    An exact support-function inequality is derived that characterizes shortest escape paths from arbitrary triangular forests and dual triangle covers for Moser's worm problem.

  3. Revisit escape path for infinite unit strip forest and unit broadworm

    math.OC 2026-07 reject novelty 4.0 of 10

    A convex-optimization reformulation reproduces Zalgaller's unit-strip escape path and argues the closed variant is a length-π constant-width curve, but the key equivalence is not rigorously established.

Reference graph

Works this paper leans on

27 extracted references · 2 linked inside Pith · cited by 3 Pith papers

  1. [1]

    Gross, O. A. (1955). A search problem due to Bellman

    1955

  2. [2]

    T., Falconer, K., & Guy, R

    Croft, H. T., Falconer, K., & Guy, R. K. (2012). Unsolved problems in geometry: unsolved problems in intuitive mathematics (Vol. 2). Springer Science & Business Media

    2012

  3. [3]

    R., Wetzel, J

    Finch, S. R., Wetzel, J. E. (2004). Lost in a forest. The American Mathematical Monthly, 111(8), 645-654

    2004

  4. [4]

    Gluss, B. (1961). The minimax path in a search for a circle in a plane. Naval Research Logistics Quarterly, 8(4), 357-360

    1961

  5. [5]

    Isbell, J. R. (1957). An optimal search pattern. Naval Research Logistics Quar- terly, 4(4), 357-359

    1957

  6. [6]

    Joris, H. (1980). Le chasseur perdu dans la forˆ et. Elemente der Mathematik= Revue de math´ ematiques ´ el´ ementaires= Rivista di matematica elementare, 35(1), 1-14

    1980

  7. [7]

    Zalgaller, V. A. (2005). A question of Bellman. Journal of Mathematical Sciences, 131(1), 5286-5306

    2005

  8. [8]

    Deng, Z. (2024). A General Solution to Bellman’s Lost-in-a-forest Problem. arXiv preprint arXiv:2412.10686

    arXiv 2024

Show all 27 references
  1. [9]

    Gibbs, P. E. (2016). Lost in an isosceles triangle. Working paper

    2016

  2. [10]

    Gibbs, P. (2016). Bellman’s Escape Problem for Convex Polygons

    2016

  3. [11]

    K¨ ubel, D., & Langetepe, E. (2021). On the approximation of shortest escape paths. Computational Geometry, 93, 101709

    2021

  4. [12]

    Wetzel, J. E. (2003). Fits and covers. Mathematics magazine, 76(5), 349-363

    2003

  5. [13]

    Wetzel, J.E. (1973). Sectorial covers for curves of constant length, Canadian Math. Bull. 16 367–375. 26

    1973

  6. [14]

    O., & Pach, J

    Brass, P., Moser, W. O., & Pach, J. (2005). Research problems in discrete geometry (Vol. 18). New York: Springer

    2005

  7. [15]

    Ghomi, M. (2017). Open problems in geometry of curves and surfaces. Available at: people. math. gatech. edu/ghomi/Papers/op. pdf, 4

    2017

  8. [16]

    Melzak, Z. A. (2007). Companion to concrete mathematics. Courier Corpora- tion

    2007

  9. [17]

    D., & Klamkin, M

    Chakerian, G. D., & Klamkin, M. S. (1973). Minimal covers for closed curves. Mathematics Magazine, 46(2), 55-61

    1973

  10. [18]

    F¨ uredi, Z., & Wetzel, J. E. (2011). Covers for closed curves of length two. Periodica Mathematica Hungarica, 63(1), 1-17

    2011

  11. [19]

    Movshovich, Y. (2025). Recent advances in the worm problem. European Jour- nal of Mathematics, 11(4), 71

    2025

  12. [20]

    Zalgaller, V. A. (2003). Shortest inspection curves for a sphere. Zapiski Nauch- nykh Seminarov POMI, 299, 87-108

    2003

  13. [21]

    M., Golynski, A., Lopez-Ortiz, A., & Quimper, C

    Chan, T. M., Golynski, A., Lopez-Ortiz, A., & Quimper, C. G. (2003, June). The asteroid surveying problem and other puzzles. In Proceedings of the nineteenth annual symposium on Computational geometry (pp. 372-373)

    2003

  14. [22]

    Croft, H. T. (1969). Curves intersecting certain sets of great-circles on the sphere. Journal of the London Mathematical Society, 2(1), 461-469

    1969

  15. [23]

    Zalgaller, V. A. (1994). The problem of the shortest space curve of unit width. Mat. Fiz. Anal. Geom, 1(3-4), 454-461

    1994

  16. [24]

    Treeby, D., & Wang, E. (2026). Escaping the unit ball. arXiv preprint arXiv:2601.21867

    2026 arXiv

  17. [25]

    Ghomi, M., & Wenk, J. (2021). Shortest closed curve to inspect a sphere. Journal f¨ ur die reine und angewandte Mathematik (Crelles Journal), 2021(781), 57-84

    2021

  18. [26]

    H., LINUSSON, S., & ASTLUND, J

    ASTAD, J. H., LINUSSON, S., & ASTLUND, J. W. (1999). A SMALLER SLEEPING BAG FOR A BABY SNAKE

    1999

  19. [27]

    Zalgaller, V. A. (1996) Extremal problems on the convex hull of a space curve, Algebra i Analiz, Volume 8, Issue 3, 1–13 27 College of Engineering and Computer Science, University of Central Florida, Or- lando, FL, USA Email: zhipeng.deng@ucf.edu 28

    1996

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.