Pith. sign in

REVIEW 5 major objections 5 minor 40 references

Enhancing Large-scale UAV Route Planing with Global and Local Features via Reinforcement Graph Fusion

T0 review · 5 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read This paper claims that a Delaunay-triangulation-based divide-and-conquer framework can extend any existing TSP solver to instances of up to 10,000 points without retraining, producing tours that match or beat specialized large-scale TSP…

desk verdict A practical DT-based framework for scaling TSP solvers, but the tables contain impossible numbers and the warm-up's claimed ability to restore non-Delaunay edges is not supported by its own equations. read the letter →

arxiv 2412.15537 v1 pith:YLD6NBVN submitted 2024-12-20 cs.AI cs.RO

classification cs.AIcs.RO MSC 90C2705C85
keywords UAVrouteplanninglarge-scaleTSPDelaunaytriangulationgraphfusionheatmapreinforcementlearningdivide-and-conquercombinatorialoptimization
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

The paper sets out to close the scalability gap between TSP solvers trained on small instances and the large patrol-planning instances that arise in UAV route planning. It proposes DTTGF, a three-stage framework that decomposes a large Euclidean TSP instance along its Delaunay triangulation, solves the resulting overlapping subgraphs with an off-the-shelf TSP solver, and fuses the sub-results into a global heatmap of edge probabilities. A warm-up procedure, cast as pseudo-reinforcement learning, then prunes misleading edges and restores promising ones before a standard search (S+2-OPT or MCTS) produces the final tour. The authors claim that with two very different embedded solvers—the two-stage supervised Att-GCN and the one-stage reinforcement-learning POMO—the framework consistently outperforms state-of-the-art large-scale methods (DIMES, H-TSP, Att-GCN) on TSP-500, TSP-1000, and TSP-10000 benchmarks, with deviations from optimality of about 1–3%. The central payoff of the claim is that any existing or future TSP solver can be reused for large-scale routing without retraining or fine-tuning.

What carries the argument

The Delaunay triangulation of the city set is the load-bearing object. It serves a dual role: as the adjacency guide for sampling overlapping subgraphs (preserving global edges that k-nearest-neighbour methods miss), and as a hard filter on the fused heatmap, setting $P_{ij}=0$ for every edge not in the triangulation (Eq. 3). The warm-up strategy, formulated as pseudo-reinforcement learning, defines a fitness $A_{ij}=P_{ij}\times d_{ij}$, iteratively deletes the highest-fitness edge, re-solves with the S+2-OPT decoder, and back-propagates the tour-length gain via Eq. (5) with $\alpha \in\{-1,0,1\}$ depending on whether the edge belongs to the baseline tour, the new tour, both, or neither. This mechanism is what the paper claims upgrades any embedded solver to large-scale performance without retraining.

What would settle it

Run an exact solver on several 1,000- to 10,000-city Euclidean TSP instances, compare every optimal-tour edge against the Delaunay triangulation, and count the fraction of optimal edges missing from the triangulation; if that fraction is above a small threshold, or if removing the DT filter materially shortens DTTGF's tours on the benchmark sets, the containment premise fails.

Watch

Extended reading notes

Core claim

The paper's core claim is that the Delaunay triangulation of the city set preserves enough of the optimal tour structure to serve as both a decomposition guide and a hard filter for large Euclidean TSP. Specifically, the framework zeroes out the heatmap probability of every edge that is not in the Delaunay triangulation (Eq. 3), on the strength of prior empirical evidence that optimal tours concentrate on triangulation edges. The fused heatmap from subgraph solutions is then refined by a warm-up loop that treats the heatmap as a policy: it repeatedly removes the edge with the largest fitness $A_{ij}=P_{ij}\times d_{ij}$, re-solves with the sampling decoder plus 2-OPT, and back-propagates the tour-length improvement through Eq. (5) to adjust the probabilities of the removed edge and of tour edges. The authors report that this pipeline, with Att-GCN or POMO embedded, beats or matches all compared baselines on the three benchmark sizes, including large-instance specialists, and that the warm-up specifically rescues one-stage solvers like POMO, whose own scaling is otherwise poor.

Load-bearing premise

The framework assumes that every edge of an optimal large-scale Euclidean TSP tour lies within the Delaunay triangulation of the city set, so discarding all other edges cannot remove needed tour edges.

Editorial extensions

If this is right

  • Any existing TSP solver that outputs tours or heatmaps can be embedded into DTTGF and extended to up to 10,000 cities with no additional training.
  • The DT filter plus warm-up yields tours within roughly 1–3% of the optimal benchmark length on TSP-500 and TSP-1000, and outperforms DIMES, H-TSP, and Att-GCN on TSP-10000 in the reported comparisons.
  • POMO, a one-stage solver that fails to scale to TSP-1000 and TSP-10000 on its own, becomes competitive when embedded, which suggests the framework can rescue otherwise unscalable solvers.
  • The framework is agnostic to the solver's training paradigm, as demonstrated by successful embedding of a supervised two-stage model (Att-GCN) and a reinforcement-learning one-stage model (POMO).

Reading between the lines

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

  • The DT filter assumption implies the framework's success is tied to the geometry of Euclidean TSP; on non-Euclidean or clustered instances where optimal edges cross large empty regions, the filter could remove needed edges—a testable limitation.
  • The warm-up's fitness $A_{ij}=P_{ij}d_{ij}$ biases toward deleting long, infrequent edges; a natural extension would be to make the learning rate adaptive per instance or to replace the binary $\alpha$ with a graded reward.
  • Because the framework is training-free, future improved neural solvers could be dropped in as a wrapper without re-engineering, yielding an ongoing performance dividend the paper does not claim explicitly.
  • The same triangulate-solve-fuse-warm-up loop might transfer to other Euclidean combinatorial problems such as capacitated vehicle routing, though the DT optimal-edge property would need re-checking there.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 5 minor

Summary. This paper introduces DTTGF, a training-free divide-and-conquer framework for extending existing TSP solvers to large instances (up to 10,000 nodes). The pipeline is: Delaunay triangulation of the instance, graph sampling into subgraphs, solving subgraphs with an embedded one- or two-stage solver, merging sub-solutions/sub-heatmaps into a global heatmap, a "warm-up" pseudo-reinforcement-learning correction of the heatmap, and a final search by S+2-opt or MCTS. The authors embed Att-GCN, POMO, AM, and GCN into the framework and report average gaps of roughly 1–3% on TSP-500, TSP-1000, and TSP-10000, outperforming recent learning-based baselines.

Significance. The framework addresses a real bottleneck: learned TSP solvers trained on small instances do not scale to thousands of nodes. If the numerical results hold, the contribution is valuable because it is solver-agnostic and does not require retraining or fine-tuning; the use of standard benchmarks, the inclusion of multiple embedded solvers, a public code link, and ablation of the warm-up module are all positive features. However, the paper currently contains load-bearing reporting errors and an internally inconsistent update rule, so the empirical claims cannot be taken at face value until those are fixed.

major comments (5)
  1. [§III-F, Eq. (5)] The warm-up back-propagation formula is internally inconsistent. The text says Tdel is the improved tour and that back-propagation should "enhance" promising edges, yet the assignments α=1 for edges in Tb and α=−1 for edges in Tdel add probability mass to the baseline tour and remove mass from the improved tour, the opposite of the intended reinforcement. The exponent D(Tb)−D(Tdel)/D(Td) also uses an undefined quantity D(Td). Since the warm-up is claimed to improve accuracy and to recover edges removed by the Delaunay filter, this formula must be corrected and the ablation re-run.
  2. [§III-E, Eq. (3)] The Delaunay filter is load-bearing but unvalidated. Setting Pij=0 for every non-DT edge makes it impossible for the fused heatmap to propose any edge outside the DT; the paper cites refs. [28]–[30] and [32] but does not report, for the benchmark suites, what fraction of optimal-tour edges are contained in the DT or what happens on instances where they are not. The warm-up cannot systematically repair this after zeroing because Aij=Pij*dij=0, so the argmax selection in §III-F never picks a zeroed edge; it can only re-enter accidentally via 2-opt moves. A quantitative containment check is required to support the SOTA claim.
  3. [Table I] Several entries in Table I are numerically impossible. Farthest Insertion on TSP-500 shows length 18.30 with drop 0.00% although the stated optimum is 16.55; DTTGF+POMO RL+MCTS shows length 24.77 with drop 9.40% (the correct gap relative to 16.55 is 49.70%); DTTGF+POMO RL+WU+S+2-OPT shows length 1.03, below any feasible Euclidean tour length. These errors make the headline comparisons unreliable and require a full re-computation of the table.
  4. [Table I and §IV-C] The warm-up time reporting is contradictory. The ablation text states warm-up times of 1.22s, 7.23s, and 5.12min for TSP-500/1000/10000, while Table I lists supplementary warm-up times of 2.70m, 15.44m, and 1.58h for the same datasets. If one set is per-instance and the other cumulative, the caption must say so; as written the numbers are incompatible and prevent assessment of the claimed time efficiency.
  5. [§III-C, §III-F, Algorithm 1] The framework depends on hyperparameters that are not specified: subgraph size and overlap, warm-up iteration count or stopping criterion, learning rate β, and the MCTS and S+2-opt budgets. These settings govern both solution quality and total runtime, so the abstract's scalability and efficiency claim cannot be reproduced from the paper alone; report them in the main text or appendix.
minor comments (5)
  1. [Title and throughout] The title contains a typo ("Planing" should be "Planning"), and throughout the text "UA V" appears with an unintended space (e.g., Abstract and Section I).
  2. [References [15] and [35]] References [15] and [35] appear to refer to the same Att-GCN paper with inconsistent author strings and years; please merge them and use one consistent citation.
  3. [Table II] Table II's header says "embedding AM and GCM", but the rows and the surrounding text refer to GCN; this should be corrected.
  4. [§III-E, Eq. (1)] In Eq. (1), the denominator Sij is the number of times edge (i,j) was selected across subgraphs, so for any edge never selected the probability is formally 0/0; please define how such entries are handled.
  5. [§III-F] The phrase "potential edges not in the DT results but filtered during fusion can enhance their P-value via back-propagation" is the only description of the recovery mechanism; after correcting Eq. (5), please also provide a step-by-step numerical example to explain how a zero-probability edge can receive a nonzero update.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: DTTGF is a heuristic composition whose components are defined from the instance and the embedded solver outputs, not from the benchmark results it reports.

full rationale

The paper does not fit any parameter to the reported TSP-500/1000/10000 results, and no equation reduces to its own input by construction. Step 1 computes a Delaunay triangulation; Step 2 samples subgraphs and solves them with an embedded pretrained solver; Step 3 fuses sub-heatmaps via Eq. (1) or Eq. (2); Eq. (3) applies the DT filter; the warm-up in Eqs. (4)-(5) updates the heatmap using tours produced by S+2-opt; and the final tour is obtained by searching the heatmap with S+2-opt or MCTS. Every quantity is derived from the instance and the solver outputs, never from the benchmark solution lengths. The DT-containment assumption is an external empirical premise supported by the paper's cited refs [28]-[30], which are not self-citations; even though ref [32] is a self-citation and is oddly used as an additional citation for the DT-optimality correlation, it is not load-bearing because the external evidence stands independently. The warm-up's repeated use of the S+2-opt decoder is a methodological coupling that adds extra local search, but it is not circular reasoning: the improved tours are new search outputs, not copies of the input heatmap, and the reported ablation honestly shows the added warm-up time. The strongest empirical claim, that DTTGF 'consistently outperforms SOTA,' is a performance claim with possible correctness risks, such as small margins and the unquantified failure cases of the Delaunay filter, but those are empirical robustness concerns, not circularity. Therefore the derivation chain is self-contained and no circular step is exhibited.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The paper does not derive any constants. Its central claim depends on an empirical geometric prior (DT contains optimal edges), a distributional assumption, and several unreported hyperparameters that affect solution quality and runtime.

free parameters (4)
  • Subgraph size and overlap = Not specified
    Section III-C describes graph sampling but gives no target subgraph node count, overlap, or number of subgraphs; these choices directly affect heatmap quality and final tour length.
  • Warm-up learning rate beta = Not specified
    Eq. (5) defines beta as the learning rate for heatmap updates; its value is never reported in the paper.
  • Warm-up iteration count / stopping criterion = Not specified
    Section III-F says the process ends once iteration requirements are met, without defining the requirement or giving a default number of iterations.
  • MCTS and 2-opt search budgets = Not reported
    Runtime and solution quality depend on MCTS iterations, sampling count, and 2-opt improvement budget, none of which are specified.
assumptions (4)
  • domain assumption Optimal TSP tour edges are almost always contained in the Delaunay triangulation graph, so non-DT edges can be deleted.
    Invoked in Section III-E before Eq. (3) to justify zeroing Pij for edges not in DT; supported only by prior empirical studies, not by a guarantee on the benchmark instances.
  • domain assumption Points are uniformly distributed in a unit square for the problem definition.
    Section III-A assumes coordinates in [0,1]^2 to match learning-based benchmarks; real UAV patrol areas may violate this and the paper does not test non-uniform distributions.
  • domain assumption Embedded TSP solver performance on small subgraphs transfers to large instances after fusion.
    The framework maps sub-solutions or sub-heatmaps to a global heatmap without formal guarantees, as described in Section III-D and III-E.
  • standard math Delaunay triangulation properties used for graph construction are standard mathematical facts.
    The empty circumcircle and max-min angle properties in Section III-A are standard, unproved background facts.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Enhancing Large-scale UAV Route Planing with Global and Local Features via Reinforcement Graph Fusion." pith.science (2026). https://pith.science/paper/YLD6NBVN

@misc{pith2026241215537,
  author       = {Pith},
  title        = {Pith review of: Enhancing Large-scale UAV Route Planing with Global and Local Features via Reinforcement Graph Fusion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YLD6NBVN}},
  note         = {Machine review of arXiv:2412.15537}
}
read the original abstract

Numerous remarkable advancements have been made in accuracy, speed, and parallelism for solving the Unmanned Aerial Vehicle Route Planing (UAVRP). However, existing UAVRP solvers face challenges when attempting to scale effectively and efficiently for larger instances. In this paper, we present a generalization framework that enables current UAVRP solvers to robustly extend their capabilities to larger instances, accommodating up to 10,000 points, using widely recognized test sets. The UAVRP under a large number of patrol points is a typical large-scale TSP problem.Our proposed framework comprises three distinct steps. Firstly, we employ Delaunay triangulation to extract subgraphs from large instances while preserving global features. Secondly, we utilize an embedded TSP solver to obtain sub-results, followed by graph fusion. Finally, we implement a decoding strategy customizable to the user's requirements, resulting in high-quality solutions, complemented by a warming-up process for the heatmap. To demonstrate the flexibility of our approach, we integrate two representative TSP solvers into our framework and conduct a comprehensive comparative analysis against existing algorithms using large TSP benchmark datasets. The results unequivocally demonstrate that our framework efficiently scales existing TSP solvers to handle large instances and consistently outperforms state-of-the-art (SOTA) methods. Furthermore, since our proposed framework does not necessitate additional training or fine-tuning, we believe that its generality can significantly advance research on end-to-end UAVRP solvers, enabling the application of a broader range of methods to real-world scenarios.

Figures

Figures reproduced from arXiv: 2412.15537 by the authors.

Figure 1
Figure 1. Pipeline of the proposed approach. Step 1: perform Delaunay Triangulation(DT) on the current instance. Step 2: sampling and solving [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 36 canonical work pages

  1. [28]

    Delaunay-triangulation-based variable neighborhood search to solve large-scale general colored traveling salesman problems,

    X. Xu, J. Li, and M. Zhou, “Delaunay-triangulation-based variable neighborhood search to solve large-scale general colored traveling salesman problems,” IEEE Transactions on Intelligent Transportation Systems, vol. 22, no. 3, pp. 1583–1593, 2021

  2. [30]

    TSP in a Simple Polygon,

    H. Alkema, M. de Berg, M. Monemizadeh, and L. Theocharous, “TSP in a Simple Polygon,” in 30th Annual European Symposium on Algorithms (ESA 2022) (S. Chechik, G. Navarro, E. Rotenberg, and G. Herman, eds.), vol. 244 of Leibniz International Proceedings in Informatics (LIPIcs) , (Dagstuhl, Germany), pp. 5:1–5:14, Schloss Dagstuhl – Leibniz-Zentrum f ¨ur Inf...

  3. [32]

    Cross-modality attack boosted by gradient-evolutionary multiform optimization,

    Y . Gong, Q. Zeng, D. Xu, Z. Wang, and M. Jiang, “Cross-modality attack boosted by gradient-evolutionary multiform optimization,” arXiv preprint arXiv:2409.17977, 2024

  4. [1]

    Optimization approaches for the traveling salesman problem with drone,

    N. Agatz, P. Bouman, and M. Schmidt, “Optimization approaches for the traveling salesman problem with drone,”Transportation Science, vol. 52, no. 4, pp. 965–981, 2018

  5. [2]

    Tsp-based pcr for rapid identification of l and s type strains of sars-cov-2,

    B. Borkakoty and N. K. Bali, “Tsp-based pcr for rapid identification of l and s type strains of sars-cov-2,”Indian journal of medical microbiology, vol. 39, no. 1, pp. 73–80, 2021

  6. [3]

    Spatial-temporal knowledge transfer for dynamic constrained multiobjective optimization,

    Z. Wang, D. Xu, M. Jiang, and K. C. Tan, “Spatial-temporal knowledge transfer for dynamic constrained multiobjective optimization,” IEEE Transactions on Evolutionary Computation , 2024

  7. [4]

    An efficient dynamic resource allocation framework for evolutionary bilevel optimization,

    D. Xu, K. Ye, Z. Zheng, T. Zhou, G. G. Yen, and M. Jiang, “An efficient dynamic resource allocation framework for evolutionary bilevel optimization,” IEEE Transactions on Cybernetics , 2024

  8. [5]

    Fast multilabel feature selection via global relevance and redundancy optimization,

    J. Zhang, Y . Lin, M. Jiang, S. Li, Y . Tang, J. Long, J. Weng, and K. C. Tan, “Fast multilabel feature selection via global relevance and redundancy optimization,” IEEE Transactions on Neural Networks and Learning Systems, vol. 35, no. 4, pp. 5721–5734, 2022

Show all 40 references
  1. [6]

    Fuzzy neural network based dynamic path planning,

    M. Jiang, Y . Yu, X. Liu, F. Zhang, and Q. Hong, “Fuzzy neural network based dynamic path planning,” 2012 International Conference on Machine Learning and Cybernetics , vol. 1, pp. 326–330, 2012

  2. [7]

    Certification of an optimal tsp tour through 85,900 cities,

    D. L. Applegate, R. E. Bixby, V . Chv ´atal, W. Cook, D. G. Espinoza, M. Goycoolea, and K. Helsgaun, “Certification of an optimal tsp tour through 85,900 cities,” Operations Research Letters , vol. 37, no. 1, pp. 11–15, 2009

  3. [8]

    Helsgaun, An Extension of the Lin-Kernighan-Helsgaun TSP Solver for Constrained Traveling Salesman and Vehicle Routing Problems: Technical report

    K. Helsgaun, An Extension of the Lin-Kernighan-Helsgaun TSP Solver for Constrained Traveling Salesman and Vehicle Routing Problems: Technical report. Roskilde Universitet, Dec. 2017

  4. [9]

    Learning 2-opt heuristics for the traveling salesman problem via deep reinforcement learning,

    P. R. de O. da Costa, J. Rhuggenaath, Y . Zhang, and A. Akcay, “Learning 2-opt heuristics for the traveling salesman problem via deep reinforcement learning,” in Proceedings of The 12th Asian Conference on Machine Learning, ACML 2020, 18-20 November 2020, Bangkok, Thailand (S....

  5. [10]

    Attention, learn to solve routing problems!,

    W. Kool, H. van Hoof, and M. Welling, “Attention, learn to solve routing problems!,” in International Conference on Learning Representations , 2019

  6. [11]

    Boosting scalability for large-scale multiobjective optimization via transfer weights,

    H. Hong, M. Jiang, and G. G. Yen, “Boosting scalability for large-scale multiobjective optimization via transfer weights,” Information Sciences, vol. 670, p. 120607, 2024

  7. [12]

    A fast dynamic evolutionary multiobjective algorithm via manifold transfer learning,

    M. Jiang, Z. Wang, L. Qiu, S. Guo, X. Gao, and K. C. Tan, “A fast dynamic evolutionary multiobjective algorithm via manifold transfer learning,” IEEE Transactions on Cybernetics , vol. 51, no. 7, pp. 3417– 3428, 2020

  8. [13]

    Individual- based transfer learning for dynamic multiobjective optimization,

    M. Jiang, Z. Wang, S. Guo, X. Gao, and K. C. Tan, “Individual- based transfer learning for dynamic multiobjective optimization,” IEEE Transactions on Cybernetics , vol. 51, no. 10, pp. 4968–4981, 2020

  9. [14]

    Evolutionary transfer optimization - a new frontier in evolutionary computation research,

    K. C. Tan, L. Feng, and M. Jiang, “Evolutionary transfer optimization - a new frontier in evolutionary computation research,”IEEE Computational Intelligence Magazine, vol. 16, no. 1, pp. 22–33, 2021

  10. [15]

    Generalize a small pre-trained model to arbitrarily large TSP instances,

    Z. Fu, K. Qiu, and H. Zha, “Generalize a small pre-trained model to arbitrarily large TSP instances,” in Thirty-Fifth AAAI Conference on Artificial Intelligence, AAAI 2021, Thirty-Third Conference on Inno- vative Applications of Artificial Intelligence, IAAI 2021, The Eleventh...

  11. [16]

    Multi-view subgraph neural networks: Self-supervised learning with scarce labeled data,

    Z. Wang, Q. Zeng, W. Lin, M. Jiang, and K. Tan, “Multi-view subgraph neural networks: Self-supervised learning with scarce labeled data,” IEEE Transactions on Neural Networks and Learning Systems , 2024

  12. [17]

    Efficiently tackling million- dimensional multiobjective problems: A direction sampling and fine- tuning approach,

    H. Hong, M. Jiang, Q. Lin, and K. C. Tan, “Efficiently tackling million- dimensional multiobjective problems: A direction sampling and fine- tuning approach,” IEEE Transactions on Emerging Topics in Computa- tional Intelligence, 2024

  13. [18]

    Traveling salesman problem: a perspective review of recent research and new results with bio-inspired metaheuristics,

    E. Osaba, X.-S. Yang, and J. Del Ser, “Traveling salesman problem: a perspective review of recent research and new results with bio-inspired metaheuristics,” Nature-inspired computation and swarm intelligence , pp. 135–164, 2020

  14. [19]

    Integration of global and local metrics for domain adaptation learning via dimensionality reduction,

    M. Jiang, W. Huang, Z. Huang, and G. G. Yen, “Integration of global and local metrics for domain adaptation learning via dimensionality reduction,” IEEE Transactions on Cybernetics, vol. 47, no. 1, pp. 38–51, 2017

  15. [20]

    Transfer learning based dynamic multiobjective optimization algorithms,

    M. Jiang, Z. Huang, L. Qiu, H. W, and G. Yen, “Transfer learning based dynamic multiobjective optimization algorithms,” IEEE Transactions on Evolutionary Computation, pp. 1–1, 2017

  16. [21]

    Evolutionary multitask optimization with lower confidence bound-based solution selection strategy,

    Z. Wang, L. Cao, L. Feng, M. Jiang, and K. C. Tan, “Evolutionary multitask optimization with lower confidence bound-based solution selection strategy,” IEEE Transactions on Evolutionary Computation , 2024

  17. [22]

    Pointer networks,

    O. Vinyals, M. Fortunato, and N. Jaitly, “Pointer networks,” Advances in neural information processing systems , vol. 28, 2015

  18. [23]

    Knee point based imbalanced transfer learning for dynamic multi-objective optimization,

    M. JIANG, Z. W ANG, H. HONG, and G. G. YEN, “Knee point based imbalanced transfer learning for dynamic multi-objective optimization,” IEEE Transactions on Evolutionary Computation , 2020

  19. [24]

    Pomo: Policy optimization with multiple optima for reinforcement learning,

    Y .-D. Kwon, J. Choo, B. Kim, I. Yoon, Y . Gwon, and S. Min, “Pomo: Policy optimization with multiple optima for reinforcement learning,” Advances in Neural Information Processing Systems, vol. 33, pp. 21188– 21198, 2020

  20. [25]

    A mixture- of-experts prediction framework for evolutionary dynamic multiobjec- tive optimization,

    R. Rambabu, P. Vadakkepat, K. C. Tan, and M. Jiang, “A mixture- of-experts prediction framework for evolutionary dynamic multiobjec- tive optimization,” IEEE transactions on cybernetics , vol. 50, no. 12, pp. 5099–5112, 2019

  21. [26]

    H- tsp: Hierarchically solving the large-scale traveling salesman problem,

    X. Pan, Y . Jin, Y . Ding, M. Feng, L. Zhao, L. Song, and J. Bian, “H- tsp: Hierarchically solving the large-scale traveling salesman problem,” in AAAI 2023, February 2023

  22. [27]

    DIMES: A differentiable meta solver for combinatorial optimization problems,

    R. Qiu, Z. Sun, and Y . Yang, “DIMES: A differentiable meta solver for combinatorial optimization problems,” in Advances in Neural Informa- tion Processing Systems , 2022

  23. [29]

    Good triangulations yield good tours,

    A. N. Letchford and N. A. Pearson, “Good triangulations yield good tours,” Computers & Operations Research , vol. 35, p. 638–647, Feb 2008

  24. [31]

    An effective heuristic algorithm for the traveling-salesman problem,

    S. Lin and B. W. Kernighan, “An effective heuristic algorithm for the traveling-salesman problem,” Operations Research , p. 498–516, Apr 1973

  25. [33]

    Ask, attend, attack: An effective decision-based black-box targeted attack for image-to-text models,

    Q. Zeng, Z. Wang, Y . Cheung, and et al., “Ask, attend, attack: An effective decision-based black-box targeted attack for image-to-text models,” in The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024

  26. [34]

    Generating diagnostic and actionable explanations for fair graph neural networks,

    Z. Wang, Q. Zeng, W. Lin, and et al., “Generating diagnostic and actionable explanations for fair graph neural networks,” in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 38, pp. 21690– 21698, 2024

  27. [35]

    Generalize a small pre-trained model to arbitrarily large tsp instances,

    Z.-H. Fu, K.-B. Qiu, and H. Zha, “Generalize a small pre-trained model to arbitrarily large tsp instances,” Proceedings of the AAAI Conference on Artificial Intelligence , p. 7474–7482, Sep 2022

  28. [36]

    The traveling sales- man problem: A computational study,

    D. Applegate, R. Bixby, V . Chv ´atal, and W. Cook, “The traveling sales- man problem: A computational study,” Choice Reviews Online,Choice Reviews Online, Feb 2007

  29. [37]

    Learning heuristics for the TSP by policy gradient,

    M. Deudon, P. Cournut, A. Lacoste, Y . Adulyasak, and L. Rousseau, “Learning heuristics for the TSP by policy gradient,” in Integration of Constraint Programming, Artificial Intelligence, and Operations Research - 15th International Conference, CPAIOR 2018, Delft, The Netherla...

  30. [38]

    An efficient graph convolu- tional network technique for the travelling salesman problem,

    C. K. Joshi, T. Laurent, and X. Bresson, “An efficient graph convolu- tional network technique for the travelling salesman problem,” 2019

  31. [39]

    Efficient active search for combi- natorial optimization problems,

    A. Hottung, Y . Kwon, and K. Tierney, “Efficient active search for combi- natorial optimization problems,” in The Tenth International Conference on Learning Representations, ICLR 2022, Virtual Event, April 25-29, 2022, OpenReview.net, 2022

  32. [40]

    H- TSP: hierarchically solving the large-scale traveling salesman problem,

    X. Pan, Y . Jin, Y . Ding, M. Feng, L. Zhao, L. Song, and J. Bian, “H- TSP: hierarchically solving the large-scale traveling salesman problem,” in Thirty-Seventh AAAI Conference on Artificial Intelligence, AAAI 2023, Thirty-Fifth Conference on Innovative Applications of Artifi...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.