Pith. sign in

REVIEW 4 major objections 6 minor 48 references

RideAgent: An LLM-Enhanced Optimization Framework for Automated Taxi Fleet Operations

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read RideAgent claims an LLM that reads historical optimal decisions can fix low-impact variables in a taxi-fleet MIP, cutting solve time by over half while holding the profit gap near 2%.

desk verdict RideAgent's LLM-guided variable fixing is a legitimate speedup heuristic, but the paper's 'near-optimal operational profit' and 'joint pricing' claims outrun what the surrogate-objective and no-demand-response model can support. read the letter →

arxiv 2505.06608 v2 pith:DE2VBQLC submitted 2025-05-10 math.OC

Xinyu Jiang , Haoyu Zhang , Mengyi Sha , Zihao Jiao , Long He , Junbo Zhang , Wei Qi This is my paper · ORCID
classification math.OC MSC 90C1190B06
keywords LLM-guidedoptimizationmixed-integerprogrammingvariablefixingheuristicrandomforestembeddedobjectivetaxipre-allocationdynamicpricingelectricfleetsfeature-driven
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

RideAgent is an optimization framework for electric taxi fleets that gives a non-expert fleet manager a way to describe an operational goal in plain language and receive a near-optimal pre-allocation and pricing plan. The paper claims that a large language model can perform two tasks that usually require a human optimization expert: translate qualitative goals into objective-function code, and learn from a small sample of historically optimal decisions which decision variables barely matter and can be fixed. That second step, called Small-Sample Guided Optimization, shrinks the mixed-integer program enough that a standard solver runs about 53% faster than on the full model while the predicted-profit objective worsens by only about 2%. If the paper is right, LLM-OR pipelines need not be one-shot parsers; they can serve as heuristic guides that keep the rigor of MIP while making it fast and accessible enough for peak-hour fleet control.

What carries the argument

The key machinery is Small-Sample Guided Optimization: an iterative loop in which a prompted LLM, called the Problem Tailor, sees the current variable set and a satisfaction score, proposes a subset of low-sensitivity decision variables to fix to historical averages, and hands the reduced problem to an MIP solver. The solver returns the best solution for the reduced problem, the satisfaction score is updated, and the loop repeats until that score stops improving. Supporting this is a random-forest-to-MIP conversion that turns each tree's leaf predictions into linear big-M logical constraints and binary leaf-selection variables, so maximizing the average predicted profit over the forest is a bona fide MIP objective rather than a separate prediction step.

What would settle it

Backtest on held-out days using realized profit rather than the forest's predictions: solve the full model and RideAgent for the same days, simulate or observe the trips that result, and compare the profits actually earned. If the realized-profit gap between the two solutions is materially larger than the reported 2.42%, the near-optimality claim is an artifact of the profit model.

Watch

Extended reading notes

Core claim

On its own terms, the paper's discovery is that a small sample of optimal solutions can substitute for a human heuristic expert. Given historical scenarios paired with their optimal decisions, an LLM proposes a subset of variables whose values barely affect the optimal solution; the solver then optimizes only the remaining variables, with the fixed ones pinned to historical averages. RideAgent couples this with an objective that is itself learned: a random forest predicts operational profit from exogenous features and decisions, its tree structure is embedded into the MIP as linear constraints, and the primary objective maximizes the average predicted profit over all trees. A secondary, LLM-generated objective reflecting the manager's query is then optimized in a second stage. On a real-world 2016 dataset with 50 zones and three state-of-charge levels, the paper reports a 53.15% reduction in solve time over the full model with a 2.42% average optimality gap, and a 42.3% time saving over five standard cutting-plane configurations.

Load-bearing premise

The central claim assumes that maximizing the random forest's predicted profit is a trustworthy stand-in for maximizing profit actually earned; since the forest explains only about 61% of profit variation on held-out data, the reported small gaps are measured against that estimate rather than against real money.

Editorial extensions

If this is right

  • Fleet managers can submit qualitative goals such as reducing idle taxis or improving high-charge taxi availability and receive a concrete pre-allocation and pricing plan without writing a model.
  • Variable fixing makes the MIP solvable in roughly half the time, so decisions can be re-run more often during peak hours or as conditions change.
  • The LLM-generated heuristic beat five standard cutting-plane methods in total time by about 42% with a similarly small profit gap, so the speedup comes from smarter model reduction rather than from the solver alone.
  • Fixing more variables widens the profit gap, especially when the user's secondary objective conflicts with the profit objective, which gives operators a tunable speed-quality trade-off.
  • The framework generalizes, as the paper argues, to other feature-driven decision problems such as supply chain, logistics, and resource scheduling wherever a small set of past optimal decisions can be produced.

Reading between the lines

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

  • The reported 2.42% gap is measured against the random forest's predicted profit, not against profit actually realized in the city; because the forest explains only 60.7% of profit variation on held-out data, the real operational gap could be larger and should be measured in a deployment.
  • The same small-sample variable-fixing recipe could be lifted out of taxi operations and applied to any large MIP whose optimal solutions can be generated offline, and a cheaper non-LLM sensitivity metric might reproduce most of the speedup, making the core claim testable independently of any particular language model.
  • Improving the embedded profit model's accuracy would probably matter more than further solver acceleration: at 60.7% test R-squared, the quality ceiling of every decision in the framework is set by the predictor, not the optimizer.
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

4 major / 6 minor

Summary. The paper proposes RideAgent, an LLM-based framework for electric taxi fleet pre-allocation and pricing. It converts natural-language operator objectives into mathematical objective functions, embeds a random-forest profit predictor directly into a mixed-integer program via the RF-to-MIP reformulation of Biggs et al., and uses an LLM to fix a subset of decision variables based on a small sample of historical optimal solutions. In a New York City taxi case study, the authors report 86% zero-shot text similarity for generated objectives, a 53.15% average time reduction versus the full RF-embedded MIP with a 2.42% average RF-Obj Gap, and a 42.3% average time advantage over five cutting-plane baseline settings.

Significance. The paper's clear strengths are the out-of-sample evaluation of LLM-generated objectives against human-annotated labels and the careful empirical setup for measuring solver-time changes on a realistically sized RF-embedded MIP. The RF-to-MIP embedding itself follows existing work, but the integration with an LLM-guided variable-fixing heuristic is a plausible applied contribution. The significance of the headline claims, however, depends on an evaluation metric that is only a proxy for operational profit: the random forest's test-set R^2 is 60.7%, and all reported optimality gaps measure deviation from the FULL model's RF-predicted profit rather than from realized profit. If the authors reframe the claims as acceleration of the RF-embedded decision model and add validation against realized profit, the framework is a useful practical tool; as written, the near-optimal operational profit claim is not established.

major comments (4)
  1. [Section V-A and Section V-C-2-b] The headline '2.42% average optimality gap' is the RF-Obj Gap, defined as the percentage deviation of RideAgent's primary objective from the FULL model's optimal objective, where the primary objective (7) is the average predicted profit across 200 random-forest trees. This is not realized profit. Since Section V-A reports a test-set R^2 of only 60.7%, the RF explains less than two-thirds of profit variance, and a small gap in RF-predicted profit does not imply near-optimal operational profit. The abstract and conclusion statements about 'minimal compromise to solution quality' and 'near-optimal results' are therefore not supported by the reported metrics. An out-of-sample evaluation against realized profit, or at least a sensitivity analysis using an alternative profit model, is needed.
  2. [Section III-A, Eqs. (1)-(6)] In the base model, the pricing variable \hat u_jk appears only in the objective through R(\hat u,d) = \sum_{j,k} u_jk d_jk, with u_jk = \theta \hat u_jk + b_j, while the satisfied demand d_jk is determined in constraint (3) entirely by the exogenous anticipated demand z_jk and the allocation variables. Price therefore has no effect on demand, so the 'dynamic pricing to modulate demand' claim is not part of the stated model. Moreover, because \hat u_jk has no bounds in (1)-(6), the base model is unbounded in the pricing variables; the RF leaf constraints bound the price only through historical training ranges and do not create a price-demand response. The authors should either add a price-dependent demand model or remove the pricing claim.
  3. [Section V-A and Section V-C-2] The variable-fixing heuristic is learned from 14 days of 'historical optimal decisions' obtained by solving the same RF-embedded FULL model, and its quality is then evaluated as the gap to that same FULL model. This makes the reported time reduction and optimality gap in-sample with respect to the RF objective. The evaluation does not test whether the variables identified by the LLM generalize to a different profit model or to realized profit. A control experiment that fixes the same number of variables by a simple data-driven rule, such as lowest variance or lowest marginal contribution to the RF objective, is needed to attribute the acceleration to LLM reasoning rather than to the variable-fixing operation itself.
  4. [Section V-C-2-b] The metric 'CPU Time' is defined as the time required to solve the optimization problem, so the reported time savings exclude LLM inference time, prompt construction, and code-safeguard checking. For a real-time fleet operations setting, end-to-end wall-clock time is the relevant quantity, and the abstract's 'reduces computation time by 53.15%' may overstate the actual latency improvement. Please report the LLM overhead separately or include it in an end-to-end timing comparison.
minor comments (6)
  1. [Section V-C-2-b] The Time Gap is defined as the absolute difference (CPU Time_FULL - CPU Time_RideAgent), but Table VI reports values such as '39.46s (30.49%)'. The percentage should be defined explicitly as relative to the FULL model's CPU time, and the term 'Gap' should be used consistently.
  2. [Section V-C-2-b and Table V] The Fixed Variable Scale is defined as the number of decision variables fixed, but Table V reports it as a percentage of 1032 total decision variables. Please define both conventions in one place.
  3. [Appendix B] The example for Result Similarity refers to 'maximizing the total number of accessible e-bikes', which appears to be a leftover from a different application. This should be replaced with the taxi pre-allocation example used elsewhere in the paper.
  4. [Section IV-A, Eq. (16)] The codomain notation R^{|Y| \times |W|} does not match the stated mapping to R_+; the function f(y; w) should be written as a map to the nonnegative reals.
  5. [Section III-B] In the definition of tree nodes, 'let pn,l n and rn be the immediate parent, the left and right children' should be written with subscripts as p_n, l_n, and r_n, and the root node needs a boundary convention for the parent relationship in constraint (10).
  6. [Section V-A] The text says the historical optimal decisions are solved for the 'profit-maximizing (obj 13)' model, but Eq. (13) is the reduced model after variable fixing; the full RF-based profit-maximizing model is Eq. (7). Please correct the equation reference.

Circularity Check

1 steps flagged · score 3.0 of 10

The speedup claim is empirical, but 'near-optimal solution quality' is measured against the same RF objective used to generate the variable-fixing guidance, making the reported gaps a self-consistency check rather than an external validation.

  1. fitted input called prediction [Section V-A (dataset), Section V-C-2-b (evaluation metrics), Section V-C-2-d (discussion of nonlinear objectives)]
    "To create a small sample of historical optimal decisions, we solve for the profit-maximizing (obj 13) allocation and pricing optimal decisions for 14 randomly selected days. These optimal decisions are provided as input to the agent as small sample optimal decision data. ... RF-Obj Gap: The percentage deviation of RideAgent’s primary objective (operational profit) from the FULL model’s optimal profit. ... The historical guidance data is optimized for the primary profit objective (RF-Obj)."

    The variable-fixing heuristic is fitted to decisions that are optimal for objective (13), which maximizes the RF's average predicted profit. The reported near-optimality metric, RF-Obj Gap, is defined as the percentage deviation from the FULL model's optimum of that same RF objective. Thus the small reported gaps (e.g., 0.41%–2.42%) show only that RideAgent's fixing rule reproduces the objective on which it was trained; they do not validate realized operational profit. Because the RF achieves only 60.7% test R^2 (Section V-A), both FULL and RideAgent optimize the same imperfect proxy, so the 'minimal compromise to solution quality' claim remains a self-consistency check on the RF objective rather than an independent prediction about real fleet operations.

full rationale

The paper's objective-generation claim is not circular: it is benchmarked against human-annotated ground-truth objective code (Section V-C-1 and Appendix A), including zero-shot out-of-sample tests, so that part has independent content. The RF-to-MIP embedding is imported from Biggs et al. [4], an external published method, and the cited model setup from Hao et al. [14] is prior work, not a self-justifying uniqueness claim. The headline time-reduction result is an empirical algorithmic comparison against the FULL model and five standard cutting-plane methods; the 53.15% time saving does not reduce by definition. However, the 'solution quality' component is partially circular: the historical optimal decisions used to teach the Problem Tailor are generated by solving objective (13), and the RF-Obj Gap measures deviation from that same objective's optimum. With test R^2 = 60.7%, this evaluation demonstrates self-consistency on the RF profit proxy, not proximity to realized operational profit. This is a partial circularity in the evaluation loop rather than a definitional collapse of the entire contribution, so the score is 3.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central claim depends on case-study constants, on the assumption that the RF profit model is a valid proxy for real profit, and on the representativeness of 14 historical optimal instances. The pricing variable is not grounded in a demand response, which is an implicit domain assumption.

free parameters (6)
  • theta (operator revenue share) = 0.2
    Set manually in the case study; scales the variable-fare revenue term in objective (1) and therefore changes the profit-maximizing allocation.
  • bj (fixed booking fee) = $5 per trip
    Set manually; enters ujk = theta * uhat_jk + bj, affecting the revenue term in the objective.
  • w_hat (unit inconvenience cost) = $0.5 per kilometer
    Set manually; defines the pre-allocation cost C(x) in the objective.
  • RF hyperparameters = 200 trees, max depth 150
    Chosen for the case study; the embedded trees define the profit landscape the solver maximizes.
  • Historical optimal sample size = 14 days
    Chosen by hand as the small sample for LLM variable-fixing guidance; the paper's small-sample effectiveness claim depends on this number.
  • Few-shot prompt count = 8 examples
    Fixed for all efficiency tests; affects the quality of LLM objective formulation and the reported similarities.
assumptions (4)
  • domain assumption Demand z_jk is exogenous and independent of the fare decision uhat_jk.
    In Eqs. (1)-(6), demand enters only as a parameter; the fare variable has no constraint or demand response, so the joint pricing claim relies on this assumption.
  • domain assumption The RF profit model is a valid proxy for true operational profit.
    The MIP maximizes average predicted profit from the random forest; test R^2 is 60.7%, so the optimization may not maximize real profit.
  • domain assumption The 14 historical optimal decision instances are representative of future optimal decisions.
    The Problem Tailor learns variable-fixing rules from these instances; if they are not representative, the fixed variables may be wrong.
  • standard math The Biggs et al. RF-to-MIP encoding exactly represents the trained random forest.
    The paper relies on the external encoding without reproducing its proof.

how reviews work

0 comments
Cite this review

Pith. "Pith review of RideAgent: An LLM-Enhanced Optimization Framework for Automated Taxi Fleet Operations." pith.science (2026). https://pith.science/paper/DE2VBQLC

@misc{pith2026250506608,
  author       = {Pith},
  title        = {Pith review of: RideAgent: An LLM-Enhanced Optimization Framework for Automated Taxi Fleet Operations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DE2VBQLC}},
  note         = {Machine review of arXiv:2505.06608}
}
read the original abstract

Efficient management of electric ride-hailing fleets, particularly pre-allocation and pricing during peak periods to balance spatio-temporal supply and demand, is crucial for urban traffic efficiency. However, practical challenges include unpredictable demand and translating diverse, qualitative managerial objectives from non-expert operators into tractable optimization models. This paper introduces RideAgent, an LLM-powered agent framework that automates and enhances electric ride-hailing fleet management. First, an LLM interprets natural language queries from fleet managers to formulate corresponding mathematical objective functions. These user-defined objectives are then optimized within a Mixed-Integer Programming (MIP) framework, subject to the constraint of maintaining high operational profit. The profit itself is a primary objective, estimated by an embedded Random Forest (RF) model leveraging exogenous features. To accelerate the solution of this MIP, a prompt-guided LLM analyzes a small sample of historical optimal decision data to guide a variable fixing strategy. Experiments on real-world data show that the LLM-generated objectives achieve an 86% text similarity to standard formulations in a zero-shot setting. Following this, the LLM-guided variable fixing strategy reduces computation time by 53.15% compared to solving the full MIP with only a 2.42% average optimality gap. Moreover, this variable fixing strategy outperforms five cutting plane methods by 42.3% time reduction with minimal compromise to solution quality. RideAgent offers a robust and adaptive automated framework for objective modeling and accelerated optimization. This framework empowers non-expert fleet managers to personalize operations and improve urban transportation system performance.

Figures

Figures reproduced from arXiv: 2505.06608 by the authors.

Figure 1
Figure 1. Training and Usage Process of the random forest [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. The agent framework new solution y ∗ t compared to a baseline, where all decision variables are set to their historical average values y¯hist: St = f(yˆ ∗ t ,y¯ ′ ;w)−f(y¯hist;w) f(y¯hist;w) . This module operates inside the iterative loop. The prompt P PP t (yˆt−1, St−1) in iteration t contains previously remaining variables yˆt−1 and satisfaction score St−1. Based on this information, the LLM provides guidance on … view at source ↗
Figure 3
Figure 3. Effect of Fixed Variable Scale on linear objective function gaps [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Agent performance on different user queries [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Objective function gaps and time advantage of RideAgent compared to different cutting methods [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Brief introduction of prompt [PITH_FULL_IMAGE:figures/full_fig_p014_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 33 canonical work pages

  1. [1]

    Gpt-4 technical report

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774, 2023

    arXiv 2023

  2. [2]

    Optimus: Optimization modeling using mip solvers and large language models

    Ali AhmadiTeshnizi, Wenzhi Gao, and Madeleine Udell. Optimus: Optimization modeling using mip solvers and large language models. arXiv preprint arXiv:2310.06116, 2023

    arXiv 2023

  3. [3]

    Conflict graphs in solving integer pro- gramming problems

    Alper Atamt ¨urk, George L Nemhauser, and Martin WP Savelsbergh. Conflict graphs in solving integer pro- gramming problems. European Journal of Operational Research, 121(1):40–55, 2000

    work page 2000

  4. [4]

    Con- strained optimization of objective functions determined from random forests

    Max Biggs, Rim Hariss, and Georgia Perakis. Con- strained optimization of objective functions determined from random forests. Production and Operations Man- agement, 32(2):397–415, 2022. doi: 10.1111/poms. 13877

    doi:10.1111/poms 2022

  5. [5]

    Open problems and fundamental limitations of reinforcement learning from human feedback

    Stephen Casper, Xander Davies, Claudia Shi, Thomas Krendl Gilbert, J ´er´emy Scheurer, Javier Rando, Rachel Freedman, Tomasz Korbak, David Lindner, Pedro Freire, et al. Open problems and fundamental limitations of reinforcement learning from human feedback. arXiv preprint arXiv:2307.15217 , 2023

    arXiv 2023

  6. [6]

    Cutting planes for integer programs with general integer variables

    Sebastian Ceria, C ´ecile Cordier, Hugues Marchand, and Laurence A Wolsey. Cutting planes for integer programs with general integer variables. Mathematical program- ming, 81:201–214, 1998

    work page 1998

  7. [7]

    Large language models for the automated analysis of optimization algorithms

    Camilo Chac ´on Sartori, Christian Blum, and Gabriela Ochoa. Large language models for the automated analysis of optimization algorithms. In Proceedings of the Genetic and Evolutionary Computation Conference , pages 160–168, 2024

    work page 2024

  8. [8]

    Dispatching through pricing: modeling ride-sharing and designing dynamic prices

    Mengjing Chen, Weiran Shen, Pingzhong Tang, and Song Zuo. Dispatching through pricing: modeling ride-sharing and designing dynamic prices. In Proceedings of the 28th International Joint Conference on Artificial Intelligence , IJCAI’19, page 165–171, 2019. ISBN 9780999241141

    work page 2019

Show all 48 references
  1. [9]

    Real-time spatial–intertemporal pricing and relocation in a ride- hailing network: Near-optimal policies and the value of dynamic pricing

    Qi Chen, Yanzhe Lei, and Stefanus Jasin. Real-time spatial–intertemporal pricing and relocation in a ride- hailing network: Near-optimal policies and the value of dynamic pricing. Operations Research, 2023

    2023

  2. [10]

    Joint relocation and pricing in electric car-sharing systems

    Ulrik Eilertsen, Olav M Falck-Pedersen, Jone V Hen- riksen, Kjetil Fagerholt, and Giovanni Pantuso. Joint relocation and pricing in electric car-sharing systems. European Journal of Operational Research, 315(2):553– 566, 2024

    2024

  3. [11]

    Innovative approaches for electric vehicles relocation in sharing systems

    Maria Pia Fanti, Agostino Marcello Mangini, Michele Roccotelli, and Biagio Silvestri. Innovative approaches for electric vehicles relocation in sharing systems. IEEE Transactions on Automation Science and Engineering, 18 (3):1116–1130, 2021

    2021

  4. [12]

    Innovative ap- proaches for electric vehicles relocation in sharing sys- tems

    Maria Pia Fanti, Agostino Marcello Mangini, Michele Roccotelli, and Bartolomeo Silvestri. Innovative ap- proaches for electric vehicles relocation in sharing sys- tems. IEEE Transactions on Automation Science and Engineering, 19(1):21–36, 2022

    2022

  5. [13]

    Outline of an algorithm for integer solutions to linear programs and an algorithm for the mixed integer problem

    Ralph E Gomory. Outline of an algorithm for integer solutions to linear programs and an algorithm for the mixed integer problem. Springer, 2010. 12

    2010

  6. [14]

    Robust vehicle pre-allocation with uncertain covariates

    Zhaowei Hao, Long He, Zhenyu Hu, and Jun Jiang. Robust vehicle pre-allocation with uncertain covariates. Production and Operations Management, 29(4):955–972, 2020

    2020

  7. [15]

    Hasija, Z

    S. Hasija, Z. J. M. Shen, and C. P. Teo. Smart city operations: Modeling challenges and opportunities. Man- ufacturing & Service Operations Management , 22(1): 203–213, 2020

    2020

  8. [16]

    Can large language models understand real-world complex instructions? In Proceedings of the AAAI Conference on Artificial Intelligence , volume 38, pages 18188–18196, 2024

    Qianyu He, Jie Zeng, Wenhao Huang, Lina Chen, Jin Xiao, Qianxi He, Xunzhe Zhou, Jiaqing Liang, and Yanghua Xiao. Can large language models understand real-world complex instructions? In Proceedings of the AAAI Conference on Artificial Intelligence , volume 38, pages 18188–18196, 2024

    2024

  9. [17]

    Introduction to operations research

    Frederick S Hillier and Gerald J Lieberman. Introduction to operations research. McGraw-Hill, 2015

    2015

  10. [18]

    Orlm: A customizable framework in train- ing large models for automated optimization modeling

    Chenyu Huan, Zhengyang Tang, Shixi Hu, Ruoqing Jiang, Xin Zheng, Dongdong Ge, Benyou Wang, and Zizhuo Wang. Orlm: A customizable framework in train- ing large models for automated optimization modeling. Operations Research, 2025

    2025

  11. [19]

    When large language model meets optimization

    Sen Huang, Kaixiang Yang, Sheng Qi, and Rui Wang. When large language model meets optimization. arXiv preprint arXiv:2405.10098, 2024

    2024 arXiv

  12. [20]

    New york city taxi trip - hourly weather data

    Kaggle. New york city taxi trip - hourly weather data. https://www.kaggle.com/datasets/meinertsen/ new-york-city-taxi-trip-hourly-weather-data, 2017

    2017

  13. [21]

    Large language models are zero-shot reasoners

    Takeshi Kojima, Shixiang Shane Gu, Machel Reid, Yu- taka Matsuo, and Yusuke Iwasawa. Large language models are zero-shot reasoners. Advances in neural information processing systems , 35:22199–22213, 2022

    2022

  14. [22]

    Optimizing relocation cost in free-floating car-sharing systems

    Damianos Kypriadis, Grammati Pantziou, Charalampos Konstantopoulos, and Damianos Gavalas. Optimizing relocation cost in free-floating car-sharing systems. IEEE Transactions on Intelligent Transportation Systems , 21 (9):4017–4030, 2020. doi: 10.1109/TITS.2020.2995197

    2020

  15. [23]

    Large language mod- els for supply chain optimization

    Beibin Li, Konstantina Mellou, Bo Zhang, Jeevan Pathuri, and Ishai Menache. Large language mod- els for supply chain optimization. arXiv preprint arXiv:2307.03875, 2023

    2023 arXiv

  16. [24]

    Syn- thesizing mixed-integer linear programming models from natural language descriptions

    Qingyang Li, Lele Zhang, and Vicky Mak-Hau. Syn- thesizing mixed-integer linear programming models from natural language descriptions. arXiv preprint arXiv:2311.15271, 2023

    2023 arXiv

  17. [25]

    Joint order dispatching and vehicle repositioning for dynamic ridesharing

    Zhidan Liu, Guofeng Ouyang, Bolin Zhang, Bo Du, Chao Chen, and Kaishun Wu. Joint order dispatching and vehicle repositioning for dynamic ridesharing. IEEE Transactions on Mobile Computing , 24(4):2628–2643,

  18. [26]

    Aggregation and mixed integer rounding to solve mips

    Hugues Marchand and Laurence A Wolsey. Aggregation and mixed integer rounding to solve mips. Operations research, 49(3):363–371, 2001

    2001

  19. [27]

    Using large language models to improve query- based constraint acquisition

    Younes Mechqrane, Christian Bessiere, and Ismail Elab- bassi. Using large language models to improve query- based constraint acquisition. In Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, IJCAI-24 , pages 1916–1925, 2024. doi: 10.249...

    1916 doi

  20. [28]

    The importance of directional feedback for llm-based optimizers

    Allen Nie, Ching-An Cheng, Andrey Kolobov, and Adith Swaminathan. The importance of directional feedback for llm-based optimizers. arXiv preprint arXiv:2405.16434, 2024

    2024 arXiv

  21. [29]

    Joint pricing and matching in ride-sharing systems

    Erhun ¨Ozkan. Joint pricing and matching in ride-sharing systems. European Journal of Operational Research, 287 (3):1149–1160, 2020

    2020

  22. [30]

    Exact solutions to a carsharing pricing and relocation problem under uncertainty

    Giovanni Pantuso. Exact solutions to a carsharing pricing and relocation problem under uncertainty. Computers & Operations Research, 144:105802, 2022

    2022

  23. [31]

    Integrating pre- diction/estimation and optimization with applications in operations management

    Mengyi Qi and Zuo-Jun (Max) Shen. Integrating pre- diction/estimation and optimization with applications in operations management. In Tutorials in Operations Research: Emerging and Impactful Topics in Operations, pages 36–58. INFORMS, 2022

    2022

  24. [32]

    A smart-city scope of operations management

    Wei Qi and Zuo-Jun Max Shen. A smart-city scope of operations management. Production and Operations Management, 28(2):393–406, 2019

    2019

  25. [33]

    Optimization methods for personalizing large language models through retrieval augmentation

    Alireza Salemi, Surya Kallumadi, and Hamed Zamani. Optimization methods for personalizing large language models through retrieval augmentation. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval , pages 752–762, 2024

    2024

  26. [34]

    Joint pricing and matching for city-scale ride- pooling

    Sanket Shah, Meghna Lowalekar, and Pradeep Varakan- tham. Joint pricing and matching for city-scale ride- pooling. In Proceedings of the International Conference on Automated Planning and Scheduling , volume 32, pages 499–507, 2022

    2022

  27. [35]

    Hugginggpt: Solving ai tasks with chatgpt and its friends in hugging face

    Yongliang Shen, Kaitao Song, Xu Tan, Dongsheng Li, Weiming Lu, and Yueting Zhuang. Hugginggpt: Solving ai tasks with chatgpt and its friends in hugging face. Advances in Neural Information Processing Systems , 36, 2024

    2024

  28. [36]

    Tlc trip record data

    Taxi & Limousine Commission. Tlc trip record data. https://www.nyc.gov/site/tlc/about/tlc-trip-record-data. page, 2024

    2024

  29. [37]

    A survey of contextual optimization methods for decision- making under uncertainty

    Sadana Utsav, Abhilash Chenreddy, Erick Delage, Alexandre Forel, Emma Frejinger, and Thibaut Vidal. A survey of contextual optimization methods for decision- making under uncertainty. European Journal of Opera- tional Research, 2024

    2024

  30. [38]

    Chain-of-thought prompting elicits reasoning in large language models

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. Chain-of-thought prompting elicits reasoning in large language models. Advances in neural information processing systems, 35:24824–24837, 2022

    2022

  31. [39]

    Valid inequalities for 0–1 knapsacks and mips with generalised upper bound constraints

    Laurence A Wolsey. Valid inequalities for 0–1 knapsacks and mips with generalised upper bound constraints. Dis- crete Applied Mathematics , 29(2-3):251–261, 1990

    1990

  32. [40]

    Chain-of-experts: When llms meet complex operations research problems

    Ziyang Xiao, Dongxiang Zhang, Yangjun Wu, Lilin Xu, Yuan Jessica Wang, Xiongwei Han, Xiaojin Fu, Tao Zhong, Jia Zeng, Mingli Song, et al. Chain-of-experts: When llms meet complex operations research problems. In The Twelfth International Conference on Learning Representations, 2023

    2023

  33. [41]

    Large language models syn- ergize with automated machine learning

    Jinglue Xu, Jialong Li, Zhen Liu, Nagar An- 13 thel Venkatesh Suryanarayanan, Guoyuan Zhou, Jia Guo, Hitoshi Iba, and Kenji Tei. Large language models syn- ergize with automated machine learning. arXiv preprint arXiv:2405.03727, 2024

    2024 arXiv

  34. [42]

    Electric vehicle fleet size and trip pricing for one-way carsharing services considering vehicle relocation and personnel assignment

    Min Xu, Qiang Meng, and Zhiyuan Liu. Electric vehicle fleet size and trip pricing for one-way carsharing services considering vehicle relocation and personnel assignment. Transportation Research Part B: Methodological , 111: 60–82, 2018

    2018

  35. [43]

    Large language models as optimizers

    Chengrun Yang, Xuezhi Wang, Yifeng Lu, Hanxiao Liu, Quoc V Le, Denny Zhou, and Xinyun Chen. Large language models as optimizers. arXiv preprint arXiv:2309.03409, 2023

    2023 arXiv

  36. [44]

    Or-llm-agent: Au- tomating modeling and solving of operations research op- timization problem with reasoning large language model

    Bowen Zhang and Pengcheng Luo. Or-llm-agent: Au- tomating modeling and solving of operations research op- timization problem with reasoning large language model. arxiv preprint arxiv:2503.10009 , 2025

    2025 arXiv

  37. [45]

    Future aware pricing and matching for sustainable on- demand ride pooling

    Xianjie Zhang, Pradeep Varakantham, and Hao Jiang. Future aware pricing and matching for sustainable on- demand ride pooling. In Proceedings of the AAAI Conference on Artificial Intelligence , volume 37, pages 14628–14636, 2023

    2023

  38. [46]

    Decision information meets large language models: The future of explainable operations research

    Yansen Zhang, Qingcan Kang, Wing Yin Yu, Hailei Gong, Xiaojin Fu, Xiongwei Han, Tao Zhong, and Chen Ma. Decision information meets large language models: The future of explainable operations research. arXiv preprint arXiv:2502.09994, 2025

    2025 arXiv

  39. [47]

    Results Similarity

    Andrew Zhao, Daniel Huang, Quentin Xu, Matthieu Lin, Yong-Jin Liu, and Gao Huang. Expel: Llm agents are experiential learners. In Proceedings of the AAAI Conference on Artificial Intelligence , volume 38, pages 19632–19642, 2024. APPENDIX A. Objective Function Standard Answer ...

    2024

  40. [2025]

    doi: 10.1109/TMC.2024.3493974

    2024

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.