SORREL: Suboptimal-Demonstration-Guided Reinforcement Learning for Learning to Branch

Shengyu Feng; Yiming Yang

REVIEW 3 major objections 6 minor 58 references

SORREL: Suboptimal-Demonstration-Guided Reinforcement Learning for Learning to Branch

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read SORREL shows a branching policy for mixed-integer linear programs can be trained from suboptimal demonstrations and match an imitation learner trained on high-quality demonstrations.

desk verdict SORREL is a genuinely new and well-tested recipe for learning branching policies from cheap, suboptimal demonstrations, but the evidence for why it works is thinner than the claims. read the letter →

arxiv 2412.15534 v2 pith:4KVGWRX3 submitted 2024-12-20 cs.LG

Shengyu Feng , Yiming Yang This is my paper · ORCID

classification cs.LG

keywords mixedintegerlinearprogrammingbranch-and-boundlearningtobranchofflinereinforcementself-imitationtreeMarkovdecisionprocessimitationneuralbranching

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

Mixed-integer linear program (MILP) solvers rely on branching heuristics that are either hand-crafted or learned from high-quality demonstrations produced by near-optimal rules. High-quality demonstrations are expensive or impossible to obtain for every problem family. This paper claims that a branching policy can be trained effectively from suboptimal demonstrations by selectively imitating only the good decisions among them. It proposes SORREL, a two-stage reinforcement learning method that pretrains offline on suboptimal demonstrations and then finetunes with self-imitation learning. The reported experiments show SORREL beating imitation and reinforcement baselines that use the same weak demonstrations, and matching an imitation learner trained on full strong branching data.

What carries the argument

The central object is a tree Markov decision process for branch-and-bound variable selection: a state is a sub-MILP, an action picks a variable to branch on, and the action produces two child states whose rewards are the local dual-bound improvements. The return from a state is defined as the expected discounted reward over a random walk down the tree, and the paper proves the resulting Bellman operator is a contraction. On this MDP, SORREL trains a policy with offline reinforcement learning, regularizing the policy update toward the demonstration actions to control distributional shift, then finetunes online with Proximal Policy Optimization augmented by self-imitation learning that replays high-return trajectories kept in per-instance priority queues.

What would settle it

Run SORREL on a MILP family where the dual bound improves quickly but the search tree remains large, or replace the reward with the actual reduction in node count per branching step and compare policies; if the tree-size objective yields clearly better node counts, the dual-bound proxy is the load-bearing assumption.

Watch

Extended reading notes

Core claim

SORREL establishes that the need for near-optimal branching demonstrations can be removed: a policy trained on suboptimal demonstrations, with a value filter that keeps only decisions leading to above-expected returns, performs as well as a policy trained on high-quality demonstrations. The method models branching as a tree Markov decision process in which rewards are local dual-bound improvements and returns are expectations over random walks down the tree; the associated Bellman operator is a contraction. Offline reinforcement learning with a behavior-cloning regularizer provides the pretrained starting policy, and self-imitation learning from a priority queue of the best trajectories provides the finetuning signal. On five MILP benchmarks, SORREL consistently reduces solving time and search-tree size relative to comparable neural baselines, and its offline-only variant already beats imitation learning on the same demonstrations.

Load-bearing premise

The load-bearing premise is that pushing the lower bound of relaxed subproblems upward quickly, as measured along a random walk of the search tree, will make the solver explore fewer nodes overall.

Editorial extensions

If this is right

MILP branching policies can be trained without near-optimal heuristics, removing the chicken-and-egg problem in data collection for new problem families.
The offline pretraining stage provides a strong starting point, so online finetuning requires far fewer solver interactions than prior RL branching methods.
On problem families with uninformative linear relaxations, SORREL can outperform imitation learning trained on full strong branching data.
Trained on small instances, SORREL generalizes to larger instances, cutting search-tree size and solving time on transfer benchmarks.

Reading between the lines

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

The value-filtered use of suboptimal demonstrations could be transferred to other solver heuristics such as node selection and cutting-plane selection, where good demonstrations are similarly expensive.
The dual-bound reward proxy may be replaceable by direct node-count objectives or learned reward models; comparing those would isolate whether the reward choice or the selective imitation mechanism drives the gains.
The reported training-efficiency gain suggests that self-imitation acts as a conservative exploration strategy, trading asymptotic performance for stability; the paper's stated limitation of not benefiting from longer RL training time supports this reading.

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.

Desk Editor's Note

read the letter

SORREL is a sensible engineering contribution, not a deep theoretical one. The two-stage combination of TD3+BC-style offline RL on cheap heuristic demonstrations, followed by PPO with self-imitation learning, appears new for learning-to-branch, and the empirical results on five benchmarks are strong enough to take seriously. The paper is honest about its main limitation: it explicitly says reward design and returns are open alternatives.

What is actually new: using offline RL to avoid the need for expert demonstrations, plus SIL to improve online sample efficiency. That is a useful practical recipe. The tree-MDP formulation is a minor generalization of Scavuzzo et al. 2022, and the propositions are standard contraction and stationary-distribution facts—correct but not deep. The experiments support the central claim that SORREL beats the included neural baselines on most benchmarks, transfers to larger instances, and that offline pretraining plus online finetuning both help. The MK result, where SORREL beats GCNN-FSB from suboptimal demonstrations, is the most convincing single datapoint.

Now the soft spots. The reward design is asserted to reduce tree size via "improve the dual bound as soon as possible," but there is no ablation that varies the reward, and the link between discounted local-dual-bound improvement along a random walk and total B&B node count is genuinely heuristic. That is a real gap, but it does not sink the paper: the method works empirically even if its mechanism is not isolated. Also missing: no comparison to the parallel Zhang et al. 2024 hybrid augmentation approach, no code, and no hyperparameter values. The training-efficiency claim is under-supported—Figure 3 only shows SIL versus no SIL, not wall-clock comparisons to TreeREINFORCE or GCNN. Error bars are per-instance rather than across seeds, so "consistent improvement" is a bit overstated.

Net: this is a solid paper that deserves peer review. It is not revolutionary, but it addresses a real bottleneck and the experiments are careful enough to engage with. A serious referee should ask for code, a reward ablation, and a direct efficiency comparison. I would accept it for review and would cite it as related work if I were working in this subfield.

Referee Report

3 major / 6 minor

Summary. This paper proposes SORREL, a two-stage reinforcement learning method for variable selection (branching) in MILP branch-and-bound. Stage 1 performs offline RL on demonstrations collected by a suboptimal hybrid heuristic (VHB), using a TD3+BC-style objective. Stage 2 finetunes the agent online with PPO and augments it with self-imitation learning from per-instance priority queues. The RL formulation is a tree MDP with a random-walk discounted local-dual-bound reward. Experiments on five MILP benchmarks and a size-transfer setting report that SORREL reduces B&B tree size and solving time relative to GCNN, TreeREINFORCE, TreeDQN, and the underlying suboptimal heuristic, with ablations attributing gains to offline pretraining and SIL.

Significance. If the empirical results hold, the paper makes a useful practical contribution: it shows that suboptimal demonstrations can be used to train branching policies that match or beat expert-trained imitation on several benchmarks while improving sample efficiency. The evaluation is broad (five benchmark families, transfer to larger instances, and ablations) and the core algorithmic recipe is plausible. The main weakness is that the link between the random-walk local-dual-bound reward and the actual evaluation metric (B&B node count) is asserted heuristically and never tested by a reward ablation, so the mechanism behind the reported gains is not established. The paper would be significantly strengthened by isolating the reward design and by tightening the theoretical claims.

major comments (3)

[Method, 'Tree MDP for B&B' (after Proposition 3)] The sentence 'Since γ ∈ [0, 1), maximizing the return would encourage the brancher to improve the dual bound as soon as possible, and thus reduce the size of the search tree as well' is the load-bearing link between the RL objective and the evaluation metric, but it is asserted rather than proved. Total B&B node count depends on the entire tree shape, the node-selection policy, and the evolution of the primal bound; maximizing discounted local dual-bound improvement can in principle increase the number of nodes by producing lopsided subtrees or interacting badly with node selection. The experiments never vary the reward (Table 3 ablates offline pretraining, online finetuning, and SIL, while Figure 4 varies demonstration quality), so the reported gains could come from the BC/offline/SIL machinery instead of the tree-MDP reward. Please add either a formal argument or an ablation that isolates the reward design (for example, a sparse -1-per-node reward, an undiscounted LDB reward, or different settings of κ and γ).
[Eq. (5), 'Pretraining with Offline Reinforcement Learning'] The Bellman target in Eq. (5) writes E_{schi, achi}[rchi + γQθ′(schi, achi)] but does not specify the distribution of the next action achi. If achi is drawn from the current policy πφ, the equation is a policy-evaluation target; if it is drawn from the behavior policy that generated D1, the learned Q is not the value of the actor. The subsequent actor update (Eq. (6)) assumes the former, but the text should state it explicitly, because the choice affects both correctness and reproducibility of the offline stage.
[Conclusion & Limitation] The paper's own limitation paragraph concedes that 'Design alternatives, including the reward function, returns, and priority queues, are still open for study.' This is consistent with the heuristic status of the reward objective described in my first comment, but it also means the abstract's and introduction's claims that SORREL is 'grounded' on the tree MDP should be narrowed: the theoretical results (Propositions 2 and 3) establish properties of the random-walk Bellman operator, not that the LDB reward is the right objective for minimizing B&B tree size. Please separate these claims and support the reward choice with evidence.

minor comments (6)

[Main Results paragraph] The text contains a typo: 'TreeREINFROCE' should be 'TreeREINFORCE'.
[Table 3] The table header 'SORRE-offline' should be 'SORREL-offline' for consistency with the text.
[Experimental Setup, Metrics] The sentence 'we always keep the results from the hard instances and the solving is not interrupted only if the 1-hour time limit is reached' is unclear; please state explicitly how timeouts are handled in the reported time metrics (e.g., whether a timeout is recorded as 3600 seconds).
[Experimental Setup] Hyperparameters α, κ, γ, ϵ, the network architecture, the number of training iterations, and the priority-queue size are not reported, which makes the training-efficiency comparison hard to reproduce.
[Figure 4] The y-axis label is missing; the caption should state what quantity is plotted (tree size or solving time).
[Method, Proposition 1] Proposition 1 ('Any B&B process following Definition 1 can be formulated as a tree MDP') is true by construction because Definition 1 already assigns the tree-MDP components; the 'universal model' claim would be stronger if the paper showed how arbitrary node-selection policies are handled or how the random-walk abstraction covers the actual B&B queue.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: SORREL's core training loop is a standard offline/online RL pipeline, and the LDB-reward-to-tree-size link is a heuristic proxy rather than a circular derivation.

full rationale

Score 0. I walked the claimed derivation chain: (i) tree MDP modeling (Definition 1, Propositions 1-3, Eq. 4); (ii) offline RL pretraining (Eqs. 5-7); (iii) online PPO finetuning and SIL (Eqs. 8-11); (iv) evaluation on node count and solving time. No step fits a parameter to the evaluation metric and then reports it as a prediction. The offline critic is trained by Bellman backup on demonstration transitions with LDB-improvement reward; the actor is BC-regularized, following external prior work. The online stage is standard PPO plus SIL with positive-advantage filtering; the SIL priority queue uses tree size to select trajectories, but the policy update (Eqs. 10-11) uses returns, so this is training-data selection, not a fitted parameter renamed as a prediction. Proposition 1 is essentially a formal restatement of Definition 1, but it is a modeling equivalence, not an empirical 'prediction', so it is not circular in the prohibited sense. The only unproven link is the claim that maximizing discounted LDB return reduces B&B node count; the paper itself flags reward design as open in the Conclusion and Limitation section. That is a heuristic assumption and a correctness risk, not circularity. Citations to tree MDP (Etheve 2020, Scavuzzo 2022), SIL (Oh 2018), PPO (Schulman 2017), and offline BC regularization (Fujimoto & Gu 2021) are to external prior work; there is no load-bearing self-citation or imported uniqueness theorem. The paper is self-contained against the external benchmarks it compares to, so no circular step is exhibited.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central claim relies on an unproven reward-proxy assumption (LDB improvement to tree size) and on standard RL assumptions (Markov state, contraction). The method itself uses no new entities. Main free parameters are unreported hyperparameters, which weakens reproducibility but not the logical derivation.

free parameters (5)

α (BC regularization scale)
Controls λ in the offline RL objective (Eq. 7); hand-chosen, not reported, and it modulates how strongly the policy is tied to suboptimal demonstrations.
κ (lower-LDB child probability)
Sets pch in the tree MDP random walk; defines the return objective's preference over leaf nodes and is not reported.
γ (discount factor)
Discounts future LDB improvements in the return; part of the objective the agent is trained to maximize, value not stated.
ϵ (PPO clip threshold)
Standard PPO clipping hyperparameter; value not reported, affects online update stability.
priority queue size limit
Determines how many best trajectories are retained per instance for SIL; not specified.

assumptions (5)

domain assumption Maximizing expected discounted LDB improvement reduces B&B tree size.
Stated in Method: 'maximizing the return would encourage the brancher to improve the dual bound as soon as possible, and thus reduce the size of the search tree as well.' No proof; the entire RL objective rests on this proxy.
domain assumption The sub-MILP at a node is a sufficient Markov state for branching decisions.
Definition 1 sets state si = MILPi and ignores search tree history and node selection context; GNN policies then map only the local MILP. This is standard in learning-to-branch but unproven.
standard math The Bellman operator T^π_tree is a γ-contraction in L∞.
Proposition 2; follows from discounting and standard contraction arguments, treated as correct.
ad hoc to paper The random-walk return with pch(schi|si)=κ for lower-LDB child is a valid objective for branching.
Introduced to create dense rewards and to allow flexible leaf preferences (Proposition 3). The specific κ is not derived from a tree-size minimization principle.
domain assumption VHB (FSB with probability 0.05, otherwise pseudo-cost) is a representative suboptimal demonstration heuristic.
Used as the only demonstration source in main experiments; the sensitivity study varies p but does not cover other heuristic families.

how reviews work

0 comments

Cite this review

Pith. "Pith review of SORREL: Suboptimal-Demonstration-Guided Reinforcement Learning for Learning to Branch." pith.science (2026). https://pith.science/paper/4KVGWRX3

@misc{pith2026241215534,
  author       = {Pith},
  title        = {Pith review of: SORREL: Suboptimal-Demonstration-Guided Reinforcement Learning for Learning to Branch},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4KVGWRX3}},
  note         = {Machine review of arXiv:2412.15534}
}

read the original abstract

Mixed Integer Linear Program (MILP) solvers are mostly built upon a Branch-and-Bound (B\&B) algorithm, where the efficiency of traditional solvers heavily depends on hand-crafted heuristics for branching. The past few years have witnessed the increasing popularity of data-driven approaches to automatically learn these heuristics. However, the success of these methods is highly dependent on the availability of high-quality demonstrations, which requires either the development of near-optimal heuristics or a time-consuming sampling process. This paper averts this challenge by proposing Suboptimal-Demonstration-Guided Reinforcement Learning (SORREL) for learning to branch. SORREL selectively learns from suboptimal demonstrations based on value estimation. It utilizes suboptimal demonstrations through both offline reinforcement learning on the demonstrations generated by suboptimal heuristics and self-imitation learning on past good experiences sampled by itself. Our experiments demonstrate its advanced performance in both branching quality and training efficiency over previous methods for various MILPs.

Figures

Figures reproduced from arXiv: 2412.15534 by the authors.

**Figure 2.** Tree MDP & its associated rewards. MILPi is the corresponding local sub-MILP, action ai as the selected variable for branching and reward r(si , ai , schi ) as the local-dual-bound improvement minch LDBch − LDBi . Proposition 1. Any B&B process following Definition 1 can be formulated as a tree MDP. Our proposed model does not exactly follow the actual B&B process but instead encourages local dual-bound improvement… view at source ↗

**Figure 4.** Comparative results on the MIS datset under differ [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗

Figures from the paper (1 more)

**Figure 3.** Figure 3: Effect of SIL on the sample efficiency. Training [PITH_FULL_IMAGE:figures/full_fig_p007_3.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 33 canonical work pages

[1]

, " * write output.state after.block = add.period write newline

ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...
[2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...
[3]

Achterberg, T. 2008. Constraint Integer Programming. ZIB-Report: Konrad-Zuse-Zentrum f \"u r Informationstechnik. Verlag Dr. Hut. ISBN 9783899638929

work page 2008
[4]

Achterberg, T. 2009. SCIP: solving constraint integer programs. Mathematical Programming Computation, 1: 1--41

2009
[5]

Achterberg, T.; Koch, T.; and Martin, A. 2005 a . Branching Rules Revisited. Oper. Res. Lett., 33(1): 42–54

work page 2005
[6]

Achterberg, T.; Koch, T.; and Martin, A. 2005 b . Branching rules revisited. Operations Research Letters, 33(1): 42--54

work page 2005
[7]

Achterberg, T.; and Wunderling, R. 2013. Mixed Integer Programming: Analyzing 12 Years of Progress, 449--481. ISBN 978-3-642-38188-1

work page 2013
[8]

M.; Louveaux, Q.; and Wehenkel, L

Alvarez, A. M.; Louveaux, Q.; and Wehenkel, L. 2017. A Machine Learning-Based Approximation of Strong Branching. INFORMS Journal on Computing, 29(1): 185--195

work page 2017

Show all 58 references

[9]

Bengio, Y.; Lodi, A.; and Prouvost, A. 2021. Machine learning for combinatorial optimization: A methodological tour d’horizon. European Journal of Operational Research, 290(2): 405--421

2021
[10]

Bertsekas, D.; and Tsitsiklis, J. 1996. Neuro-Dynamic Programming, volume 27

1996
[11]

B.; Gleixner, A

Chmiela, A.; Khalil, E. B.; Gleixner, A. M.; Lodi, A.; and Pokutta, S. 2021. Learning to Schedule Heuristics in Branch-and-Bound. In Neural Information Processing Systems

2021
[12]

Chopra, S.; and Meindl, P. 2001. Strategy, planning, and operation. Supply Chain Management, 15(5): 71--85

2001
[13]

Cplex, I. I. 2009. V12. 1: User’s Manual for CPLEX. International Business Machines Corporation, 46(53): 157

2009
[14]

T.; Jiang, H.; Krishnamoorthy, M.; and Sier, D

Ernst, A. T.; Jiang, H.; Krishnamoorthy, M.; and Sier, D. 2004. Staff scheduling and rostering: A review of applications, methods and models. European journal of operational research, 153(1): 3--27

2004
[15]

Etheve, M.; Al \`e s, Z.; Bissuel, C.; Juan, O.; and Kedad-Sidhoum, S. 2020. Reinforcement Learning for Variable Selection in a Branch and Bound Algorithm. In Hebrard, E.; and Musliu, N., eds., Integration of Constraint Programming, Artificial Intelligence, and Operations Rese...

2020
[16]

Fujimoto, S.; and Gu, S. S. 2021. A Minimalist Approach to Offline Reinforcement Learning. In Thirty-Fifth Conference on Neural Information Processing Systems

2021
[17]

Gamrath, G.; Berthold, T.; and Salvagnin, D. 2020. An exploratory computational analysis of dual degeneracy in mixed-integer programming. EURO J. Comput. Optim., 8: 241--261

2020
[18]

Gasse, M.; Chételat, D.; Ferroni, N.; Charlin, L.; and Lodi, A. 2019. Exact Combinatorial Optimization with Graph Convolutional Neural Networks. In Advances in Neural Information Processing Systems 32

2019
[19]

B.; Kumar, M

Gupta, P.; Gasse, M.; Khalil, E. B.; Kumar, M. P.; Lodi, A.; and Bengio, Y. 2020. Hybrid Models for Learning to Branch. In Advances in Neural Information Processing Systems 33

2020
[20]

B.; Ch \'e telat, D.; Gasse, M.; Lodi, A.; Bengio, Y.; and Kumar, M

Gupta, P.; Khalil, E. B.; Ch \'e telat, D.; Gasse, M.; Lodi, A.; Bengio, Y.; and Kumar, M. P. 2022. Lookback for Learning to Branch. Transactions on Machine Learning Research. Expert Certification

2022
[21]

Gurobi Optimization, L. 2021. Gurobi optimizer reference manual

2021
[22]

Han, Q.; Yang, L.; Chen, Q.; Zhou, X.; Zhang, D.; Wang, A.; Sun, R.; and Luo, X. 2023. A GNN -Guided Predict-and-Search Framework for Mixed-Integer Linear Programming. In The Eleventh International Conference on Learning Representations

2023
[23]

Hansknecht, C.; Joormann, I.; and Stiller, S. 2018. Cuts, Primal Heuristics, and Learning to Branch for the Time-Dependent Traveling Salesman Problem. arXiv: Optimization and Control

2018
[24]

He, H.; Daume III, H.; and Eisner, J. M. 2014. Learning to Search in Branch and Bound Algorithms. In Ghahramani, Z.; Welling, M.; Cortes, C.; Lawrence, N.; and Weinberger, K., eds., Advances in Neural Information Processing Systems, volume 27. Curran Associates, Inc

2014
[25]

Hendel, G.; Miltenberger, M.; and Witzig, J. 2019. Adaptive Algorithmic Behavior for Solving Mixed Integer Programs Using Bandit Algorithms. In Fortz, B.; and Labb \'e , M., eds., Operations Research Proceedings 2018, 513--519. Cham: Springer International Publishing. ISBN 978...

2019
[26]

Huang, T.; Ferber, A.; Tian, Y.; Dilkina, B.; and Steiner, B. 2023 a . Searching Large Neighborhoods for Integer Linear Programs with Contrastive Learning. In International conference on machine learning. PMLR

2023
[27]

Huang, Z.; Chen, W.; Zhang, W.; Shi, C.; Liu, F.; Zhen, H.-L.; Yuan, M.; Hao, J.; Yu, Y.; and Wang, J. 2023 b . Branch Ranking For Efficient Mixed-Integer Programming Via Offline Ranking-Based Policy Learning. In Machine Learning and Knowledge Discovery in Databases: European ...

2023
[28]

B.; Bodic, P

Khalil, E. B.; Bodic, P. L.; Song, L.; Nemhauser, G.; and Dilkina, B. 2016. Learning to Branch in Mixed Integer Programming. In Proceedings of the Thirtieth AAAI Conference on Artificial Intelligence, AAAI'16, 724–731. AAAI Press

2016
[29]

B.; Dilkina, B.; Nemhauser, G

Khalil, E. B.; Dilkina, B.; Nemhauser, G. L.; Ahmed, S.; and Shao, Y. 2017. Learning to Run Heuristics in Tree Search. In Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence, IJCAI-17 , 659--666

2017
[30]

Konda, V.; and Tsitsiklis, J. 1999. Actor-Critic Algorithms. In Solla, S.; Leen, T.; and M\" u ller, K., eds., Advances in Neural Information Processing Systems, volume 12. MIT Press

1999
[31]

Kostrikov, I.; Tompson, J.; Fergus, R.; and Nachum, O. 2021. Offline Reinforcement Learning with Fisher Divergence Critic Regularization. ArXiv, abs/2103.08050

2021 arXiv
[32]

Kumar, A.; Fu, J.; Tucker, G.; and Levine, S. 2019. Stabilizing Off-Policy Q-Learning via Bootstrapping Error Reduction. In Neural Information Processing Systems

2019
[33]

H.; and Doig, A

Land, A. H.; and Doig, A. G. 1960. An Automatic Method of Solving Discrete Programming Problems. Econometrica, 28(3): 497--520

1960
[34]

Levine, S.; Kumar, A.; Tucker, G.; and Fu, J. 2020. Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problems. ArXiv, abs/2005.01643

2020 arXiv
[35]

Mashayekh, S.; Stadler, M.; Cardoso, G.; and Heleno, M. 2017. A mixed integer linear programming approach for optimal DER portfolio, sizing, and placement in multi-energy microgrids. Applied Energy, 187

2017
[36]

Nair, A.; McGrew, B.; Andrychowicz, M.; Zaremba, W.; and Abbeel, P. 2018. Overcoming Exploration in Reinforcement Learning with Demonstrations. In 2018 IEEE International Conference on Robotics and Automation (ICRA), 6292–6299. IEEE Press

2018
[37]

Nair, V.; Bartunov, S.; Gimeno, F.; von Glehn, I.; Lichocki, P.; Lobov, I.; O'Donoghue, B.; Sonnerat, N.; Tjandraatmadja, C.; Wang, P.; Addanki, R.; Hapuarachchi, T.; Keck, T.; Keeling, J.; Kohli, P.; Ktena, I.; Li, Y.; Vinyals, O.; and Zwols, Y. 2020. Solving Mixed Integer Pr...

2020 arXiv
[38]

Oh, J.; Guo, Y.; Singh, S.; and Lee, H. 2018. Self-Imitation Learning. In Dy, J.; and Krause, A., eds., Proceedings of the 35th International Conference on Machine Learning, volume 80 of Proceedings of Machine Learning Research, 3878--3887. PMLR

2018
[39]

Parsonson, C. W. F.; Laterre, A.; and Barrett, T. D. 2023. Reinforcement Learning for Branch-and-Bound Optimisation Using Retrospective Trajectories. Proceedings of the AAAI Conference on Artificial Intelligence, 37(4): 4061--4069

2023
[40]

B.; and Krause, A

Paulus, M. B.; and Krause, A. 2023. Learning To Dive In Branch And Bound. arXiv:2301.09943

2023 arXiv
[41]

B.; Zarpellon, G.; Krause, A.; Charlin, L.; and Maddison, C

Paulus, M. B.; Zarpellon, G.; Krause, A.; Charlin, L.; and Maddison, C. 2022. Learning to Cut by Looking Ahead: Cutting Plane Selection via Imitation Learning. In Chaudhuri, K.; Jegelka, S.; Song, L.; Szepesvari, C.; Niu, G.; and Sabato, S., eds., Proceedings of the 39th Inter...

2022
[42]

Qu, Q.; Li, X.; Zhou, Y.; Zeng, J.; Yuan, M.; Wang, J.; Lv, J.; Liu, K.; and Mao, K. 2022. An Improved Reinforcement Learning Algorithm for Learning to Branch. ArXiv, abs/2201.06213

2022 arXiv
[43]

Y.; Ch \'e telat, D.; Gasse, M.; Lodi, A.; Yorke-Smith, N.; and Aardal, K

Scavuzzo, L.; Chen, F. Y.; Ch \'e telat, D.; Gasse, M.; Lodi, A.; Yorke-Smith, N.; and Aardal, K. 2022. Learning to Branch with Tree MDP s. In Oh, A. H.; Agarwal, A.; Belgrave, D.; and Cho, K., eds., Advances in Neural Information Processing Systems

2022
[44]

Schulman, J.; Wolski, F.; Dhariwal, P.; Radford, A.; and Klimov, O. 2017. Proximal Policy Optimization Algorithms. arXiv:1707.06347

2017 arXiv
[45]

Song, J.; Lanka, R.; Zhao, A.; Yue, Y.; and Ono, M. 2018. Learning to Search via Retrospective Imitation. arXiv: Learning

2018
[46]

Sonnerat, N.; Wang, P.; Ktena, I.; Bartunov, S.; and Nair, V. 2021. Learning a Large Neighborhood Search Algorithm for Mixed Integer Programs. ArXiv, abs/2107.10201

2021 arXiv
[47]

Sorokin, D.; and Kostin, A. 2023. TreeDQN: Learning to minimize Branch-and-Bound tree. arXiv:2306.05905

2023 arXiv
[48]

Sun, H.; Chen, W.; Li, H.; and Song, L. 2020. Improving Learning to Branch via Reinforcement Learning. In Learning Meets Combinatorial Algorithms at NeurIPS2020

2020
[49]

Tang, Y.; Agrawal, S.; and Faenza, Y. 2020. Reinforcement Learning for Integer Programming: Learning to Cut. In Proceedings of the 37th International Conference on Machine Learning, ICML'20. JMLR.org

2020
[50]

Trofin, M.; Qian, Y.; Brevdo, E.; Lin, Z.; Choromanski, K.; and Li, D. 2021. Mlgo: a machine learning guided compiler optimizations framework. arXiv preprint arXiv:2101.04808

2021 arXiv
[51]

Turner, M.; Koch, T.; Serrano, F.; and Winkler, M. 2023. Adaptive Cut Selection in Mixed-Integer Linear Programming . Open Journal of Mathematical Optimization, 4: 5

2023
[52]

Wang, Z.; Li, X.; Wang, J.; Kuang, Y.; Yuan, M.; Zeng, J.; Zhang, Y.; and Wu, F. 2023. Learning Cut Selection for Mixed-Integer Linear Programming via Hierarchical Sequence Model. In The Eleventh International Conference on Learning Representations

2023
[53]

Wu, Y.; Song, W.; Cao, Z.; and Zhang, J. 2021. Learning Large Neighborhood Search Policy for Integer Programming. In Advances in Neural Information Processing Systems

2021
[54]

Yoon, T. 2022. Confidence Threshold Neural Diving. arXiv:2202.07506

2022 arXiv
[55]

Zarpellon, G.; Jo, J.; Lodi, A.; and Bengio, Y. 2021. Parameterizing Branch-and-Bound Search Trees to Learn Branching Policies. Proceedings of the AAAI Conference on Artificial Intelligence, 35(5): 3931--3939

2021
[56]

Zhang, C.; Ouyang, W.; Yuan, H.; Gong, L.; Sun, Y.; Guo, Z.; Dong, Z.; and Yan, J. 2024. Towards Imitation Learning to Branch for MIP : A Hybrid Reinforcement Learning based Sample Augmentation Approach. In The Twelfth International Conference on Learning Representations

2024
[57]

Zhang, T.; Banitalebi-Dehkordi, A.; and Zhang, Y. 2022. Deep Reinforcement Learning for Exact Combinatorial Optimization: Learning to Branch. arXiv:2206.06965

2022 arXiv
[58]

P.; et al

Zheng, L.; Li, Z.; Zhang, H.; Zhuang, Y.; Chen, Z.; Huang, Y.; Wang, Y.; Xu, Y.; Zhuo, D.; Xing, E. P.; et al. 2022. Alpa: Automating inter-and \ Intra-Operator \ parallelism for distributed deep learning. In 16th USENIX Symposium on Operating Systems Design and Implementation...

2022

Pith tools

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

[1] [1]

, " * write output.state after.block = add.period write newline

ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

[2] [2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

[3] [3]

Achterberg, T. 2008. Constraint Integer Programming. ZIB-Report: Konrad-Zuse-Zentrum f \"u r Informationstechnik. Verlag Dr. Hut. ISBN 9783899638929

work page 2008

[4] [4]

Achterberg, T. 2009. SCIP: solving constraint integer programs. Mathematical Programming Computation, 1: 1--41

2009

[5] [5]

Achterberg, T.; Koch, T.; and Martin, A. 2005 a . Branching Rules Revisited. Oper. Res. Lett., 33(1): 42–54

work page 2005

[6] [6]

Achterberg, T.; Koch, T.; and Martin, A. 2005 b . Branching rules revisited. Operations Research Letters, 33(1): 42--54

work page 2005

[7] [7]

Achterberg, T.; and Wunderling, R. 2013. Mixed Integer Programming: Analyzing 12 Years of Progress, 449--481. ISBN 978-3-642-38188-1

work page 2013

[8] [8]

M.; Louveaux, Q.; and Wehenkel, L

Alvarez, A. M.; Louveaux, Q.; and Wehenkel, L. 2017. A Machine Learning-Based Approximation of Strong Branching. INFORMS Journal on Computing, 29(1): 185--195

work page 2017