Preference-Conditioned Multi-Objective Reinforcement Learning for Runtime-Tunable Transit Signal Priority

Philip-Roman Adam; Stefanie Schmidtner

REVIEW 3 major objections 4 minor 24 references

Preference-Conditioned Multi-Objective Reinforcement Learning for Runtime-Tunable Transit Signal Priority

T0 review · 3 major / 4 minor · reviewed 2026-08-02 · deepseek-v4-flash

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

Pith's one-line read A single trained traffic-signal controller can be tuned at runtime, through one preference dial, to emphasize bus priority or general traffic delay, without retraining.

desk verdict A reproducible, well-executed empirical MORL-for-TSP paper that is let down mostly by an overbroad abstract and a few structural evaluation choices; the central runtime-tunability claim holds up in the simulated regime. read the letter →

arxiv 2607.18286 v1 pith:JV5TFSLL submitted 2026-06-30 cs.LG cs.AIcs.SYeess.SY

Philip-Roman Adam , Stefanie Schmidtner This is my paper

classification cs.LGcs.AIcs.SYeess.SY

keywords transitsignalprioritypreference-conditionedreinforcementlearningmulti-objectiveruntimetuningtrafficcontroltail-delayfairnesstrade-offfrontierconstrainedphase

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

Transit signal priority needs to balance two goals that pull against each other: moving buses through an intersection quickly and not punishing everyone else. Most learning-based controllers hard-wire that balance into a fixed reward, so changing the priority later means retraining. This paper claims that one trained policy, parameterized by a single preference weight, can be tuned at runtime to sweep continuously between bus-priority emphasis and general-traffic delay, while still respecting real signal constraints like minimum and maximum green times. In simulation across thirteen intersections, that single policy traces a smooth trade-off frontier, often beating fixed-time and rule-based baselines at moderate settings, while exposing that aggressive bus-priority settings can push large delays onto a small tail of non-bus vehicles. If true, agencies could swap many trained controllers for one tunable controller whose dial they can turn at run time.

What carries the argument

The load-bearing object is the preference-conditioned action-value function Q(s,a,w), from which the controller selects the phase that maximizes the inner product of the preference vector and the Q-values. A single scalar weight on bus priority blends a two-dimensional per-step reward vector: negative instantaneous bus delay on one entry, and negative all-vehicle delay plus a tail-delay penalty on the other. Around this is a feasibility layer that enforces minimum and maximum green times and transition rules, so every controller, learned or rule-based, is judged under the same operational constraints. To make sparse bus-priority events learnable, the training scenarios insert timetable-based

What would settle it

Run the same 11-setting preference sweep on the same intersections but replace the synthetic bus timetables with recorded arrival and headway data from a real transit agency; the claim survives only if the single conditioned policy still traces a smooth bus-delay versus non-bus-delay frontier and beats the fixed-time and rule-based baselines at moderate priorities. A sharper version: if any realistic headway distribution collapses the frontier to a single point or makes non-bus tail delay explode at moderate settings, runtime tunability as described is not portable.

Watch

Extended reading notes

Core claim

The central claim is that a preference-conditioned controller, trained once per intersection, makes the bus-versus-all-vehicle delay trade-off a continuous runtime parameter rather than a training-time choice. Learning a single action-value function over state, action, and preference lets the same controller arbitrate conflicting priority requests differently as the dial moves. The paper reports that the conditioned policy spans a smooth empirical trade-off frontier, dominates both fixed-time and rule-based transit-priority overlays for most preferences at moderate weights, and stays within minimum and maximum green and transition constraints. At high bus-priority weights, the same diagnosti

Load-bearing premise

The entire demonstration leans on the assumption that the synthetic timetable-based bus insertion (headways uniformly drawn from 300-900 seconds over four concurrent routes) produces a representative distribution of real transit-priority events; if actual bus operations differ in headway correlation, route structure, or demand, the learned policy's trade-off frontier and its dominance over the baselines may not transfer.

Editorial extensions

If this is right

A single trained policy can replace a set of fixed-weight specialists while matching or exceeding them in frontier quality on the tested intersections, removing the need to train and store many controllers.
Operators can adjust the bus-versus-traffic emphasis at runtime for time-of-day shifts, incidents, or special events without retraining.
Moderate preference settings capture most bus-delay gains with limited non-bus and tail-delay costs, suggesting a safe operating range for normal use.
The approach shows qualitative robustness when bus prevalence drops, indicating that runtime tunability may transfer to less transit-saturated settings with scenario-dependent offsets.
The constrained benchmark and shared feasibility layer give future controllers a common ground for comparison under realistic signal constraints.

Reading between the lines

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

Beyond the paper's stated results, the practical product is likely to be a bounded dial: exposing only moderate weights as the normal operating range and treating high bus-priority weights as an emergency mode with tail-delay guardrails.
The same conditioning idea should extend to richer preference inputs than one scalar, since the paper deliberately left bus occupancy and schedule deviation out of the observation; adding them would let the dial control headway adherence or passenger-weighted delay instead of raw bus delay.
Because the paper's strongest gains appear on low-complexity intersections and degrade with phase count and cycle length, agencies should expect runtime tunability to be worth less at complex intersections and should validate per site rather than citywide.

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

Candid take: this is a genuinely useful empirical contribution, not a breakthrough. The new thing is the systematic application of preference-conditioned MORL (Envelope Q-learning) to constrained transit signal priority, plus an IntersectionZoo wrapper and a baseline pack that makes results comparable. They ship code, pin SUMO/IntersectionZoo versions, run 10 seeds across 13 intersections, and report Pareto frontiers, hypervolume, and tail-delay diagnostics. That is real work, and it is mostly done carefully. The single-policy runtime tuning claim is supported in the simulated regime: the preference sweep produces a smooth trade-off, and at moderate weights MORL often beats fixed-time and rule-based TSP.

The biggest gap is the abstract, which says 'outperforms fixed-time and rule-based baselines' without qualification. The paper's own Table III shows CHI758 where MORL dominates FixedTime 1.0/11 and RuleTSP 0.0/11, and SEA543 is also weak. The Discussion rightly limits the claim to moderate preferences, so the abstract needs to match.

I also share the reader's concern about CHI758 being both the tuning site and an evaluation site. It is not fatal — the parameters were held fixed afterward — but it should be stated more clearly. And the three case-study intersections (NYC10802, LA2114, CHI2412) are presented without any selection criterion; a reader will wonder whether they were picked for favorable behavior. The distribution-shift experiment only changes bus prevalence, not route structure or headway correlations, which is a genuine limitation for the operational claim. The paper acknowledges this in Section VI but does not test it.

One thing the reader flagged that I do not weight as heavily is the synthetic bus insertion distribution. It is unvalidated, yes, but this is a benchmark paper; the synthetic regime is clearly specified and reproducible. The problem is only when they make an operational claim. That is more about framing than the method.

What is missing is an empirical comparison with [6] (Zhou et al.), the closest prior constrained-TSP RL method. They cite it but never run it. Given the paper's own emphasis on benchmarking, this is a clear soft spot.

Bottom line: the paper deserves a serious referee. It is a solid empirical study with released code, honest limitations, and a defensible central claim in the simulated regime. I would send it to review, with the expectation that the abstract be fixed, the case-study selection justified, and a baseline against the most relevant prior method added.

Referee Report

3 major / 4 minor

Summary. The paper presents a preference-conditioned multi-objective reinforcement learning controller for transit signal priority (TSP). The controller π(a|s,w) is trained once per intersection and can be tuned at runtime through a scalar preference w to trade off bus delay against all-vehicle delay, under a shared feasibility layer enforcing min/max green and transition constraints. The authors build on IntersectionZoo and SUMO-RL, add bus-prevalence and timetable-insertion wrappers, and compare against fixed-time, a rule-based TSP overlay, and fixed-weight PPO specialists. Evaluation is carried out on three case-study intersections plus ten additional intersections, with 10 seeds per setting, pinned simulator versions, and released code. The paper reports empirical Pareto fronts, hypervolume, dominance counts, operating-point delays, tail-delay diagnostics, and a distribution-shift experiment that varies bus prevalence.

Significance. If the reported results hold, the paper makes a useful contribution: it provides a reproducible benchmark for preference-conditioned TSP and demonstrates that a single conditioned policy can approximate the Pareto front between bus priority and general-traffic delay, with a runtime knob that does not require retraining. Strengths include the open-source implementation with pinned IntersectionZoo/SUMO versions, the use of 10 seeds and 13 intersections, a shared feasibility layer for fair baseline comparison, and the inclusion of tail-delay and fairness diagnostics. These are concrete assets for the RL-for-traffic-signal-control community. However, the central claim is currently overstated in the abstract and the evaluation rests on a synthetic timetable-insertion regime whose representativeness is not validated; both issues need to be addressed before the operational claims can be taken at face value.

major comments (3)

[Abstract and Section V-D / Table III] The abstract states that the controller 'outperforms fixed-time and rule-based baselines' without qualification, but Table III shows CHI758 with FT-dom=1.0 and TSP-dom=0.0, and SEA543 with TSP-dom=3.5±4.1 — meaning the learned policy does not dominate either baseline for most preference settings at these intersections. The body text is more careful, saying 'often dominates' and 'at moderate preferences.' The abstract and the concluding claims should carry the same qualifier, or the dominance claim should be restricted to the case-study intersections at moderate preferences.
[Section II-D and Section V-B] The entire empirical claim is evaluated under the synthetic timetable-insertion regime described in Section II-D: headways U[300,900] s, departure offsets U[−120,600] s, route durations U[10,800,36,000] s, and 4 concurrent routes per episode. The distribution-shift experiment in Section V-B changes only bus prevalence, not headway correlations, route structures, dwell times, or demand patterns. Since the learned policy and baselines are all measured in this constructed regime, the reported trade-off frontier and dominance numbers are not shown to transfer to real transit operations. The limitation paragraph in Section VI acknowledges 'dependence on scenario realism' but does not test it. To support the operational 'runtime-tunable TSP' claim, the authors should validate the timetable model against real GTFS/AVI data, or at minimum add a sensitivity analysis sweeping the headway, route-co
[Section IV-B and Table III] Section IV-B states that hyperparameters and the preference curriculum were tuned on CHI758 and then held fixed for all remaining intersections. However, Table III and Section V-D include CHI758 in the ten additional intersections used to demonstrate cross-intersection reproducibility, and Table IV uses all 13 intersections for the structural correlation analysis. This violates the holdout principle: CHI758 is not an independent test site. The paper should either exclude CHI758 from the additional-intersection evaluation and the correlation analysis, or explicitly label it as a development intersection whose results are not out-of-sample. The same concern applies to the RuleTSP parameter pilot sweep on CHI758.

minor comments (4)

[Section II-D] The interval '[10,800,36,000]s' is ambiguous because of the comma separator. Use 10 800–36 000 s or write 10800–36000 to avoid confusion with a three-element list.
[Section III-D] The sentence 'We apply a homotopy schedule with λ increasing from 0.0 to 0.3 over 200k steps' introduces λ without defining it. Please specify what λ controls (e.g., reward-shaping weight, Envelope Q-learning parameter) and how it interacts with the preference curriculum.
[Table IV] The cell color encoding described in the caption may be invisible in grayscale print. Add explicit numeric labels or hatching so the table is readable without color.
[Section V-A] The statement 'MORL has slightly higher HV on all three case studies' is supported by the table, but the standard deviations overlap substantially (e.g., CHI2412: 83.7±1.3 vs 83.1±2.2). Consider adding a significance test or explicitly stating that the difference is not statistically assessed.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the paper reports empirical evaluations of trained policies; hyperparameters were fixed on a pilot site and baselines are externally specified, with limitations acknowledged.

full rationale

The paper makes no first-principles derivation that could be self-referential. Its central claim—that a single preference-conditioned policy spans a smooth trade-off frontier and outperforms fixed-time and rule-based baselines—is supported by simulation experiments under a specified timetable-insertion regime (Section II-D) and evaluated with measured episode-level delays (Section IV-C). Reward-shaping constants (λ_cvar=0.25, T_cvar=120), MORL hyperparameters, and RuleTSP parameters were chosen on the pilot intersection CHI758 and held fixed; the paper even reports CHI758 as a weak case (FT-dom 1.0, TSP-dom 0.0, Table III), which is inconsistent with tuning on the evaluation target. Baselines are externally defined (fixed-time from IntersectionZoo/Thunig; rule-based GE/EG per standard TSP practice), not defined in terms of the MORL outputs. There are no self-citations; all references are to external work such as IntersectionZoo, SUMO-RL, morl-baselines, and Envelope Q-learning. The only link between the training objective and an evaluation metric is the CVaR reward-shaping term in Eq. (4) and the tail diagnostic J^tail_nb, but the latter is an episode-level measured quantity and the paper does not present it as an independent prediction; it is an evaluation of the trained policy's behavior. The acknowledged limitation 'dependence on scenario realism' (Section VI) concerns external validity, not circularity. Therefore no step reduces to its own inputs by construction.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central result is an empirical comparison; no derivation introduces hidden constants. The load-bearing hand-chosen quantities are the reward-shaping parameters and the synthetic bus-route distributions that define the evaluation regime. The policy itself is a standard Envelope-Q function conditioned on w; no new entities are postulated.

free parameters (6)

λ_cvar = 0.25
Reward-shaping weight for the non-bus CVaR tail penalty in Eq. (4); hand-chosen, affects learned policy.
T_cvar = 120 s
Threshold above which the CVaR penalty activates in Eq. (4); hand-chosen.
α = 0.1
Quantile for CVaR tail-delay diagnostics and reward penalty; hand-chosen.
P_inf = 120.0
Infeasibility penalty applied to MORL's all-vehicle reward component when it requests an infeasible phase change; hand-chosen (Section II-E).
Preference curriculum bounds = U(0.1,0.9) then U(0,1)
Training-time preference sampling distribution; hand-chosen to shape coverage of the trade-off surface (Section II-C).
Bus route sampling distributions = headways U[300,900] s; offsets U[-120,600] s; durations U[10,800,36,000] s
Parameters of the timetable-based bus insertion wrapper that defines the in-distribution evaluation regime; hand-chosen, no external calibration (Section II-D).

assumptions (5)

domain assumption SUMO v1.25.0 faithfully models signal timing, vehicle delays, and TSP effects
All results are produced inside SUMO; realism of the conclusions depends on simulator fidelity.
domain assumption IntersectionZoo scenarios at revision 912d102 provide realistic intersection geometries and traffic demand
The 13 evaluation intersections come from this dataset; no independent validation of scenario realism is provided.
domain assumption Envelope Q-learning converges to an adequate preference-conditioned value function in 500k steps
The reported trade-off frontier presumes that training did not prematurely stall; no convergence diagnostics are shown.
ad hoc to paper The timetable-based bus insertion wrapper produces a representative distribution of transit priority events
Uniform headways and durations are invented for this study; transferability to real transit operations is untested (Section II-D).
domain assumption The shared feasibility layer correctly enforces min/max green and transition constraints
Constraint feasibility is enforced by the wrapper rather than learned; the paper does not report violation rates.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Preference-Conditioned Multi-Objective Reinforcement Learning for Runtime-Tunable Transit Signal Priority." pith.science (2026). https://pith.science/paper/JV5TFSLL

@misc{pith2026260718286,
  author       = {Pith},
  title        = {Pith review of: Preference-Conditioned Multi-Objective Reinforcement Learning for Runtime-Tunable Transit Signal Priority},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JV5TFSLL}},
  note         = {Machine review of arXiv:2607.18286}
}

abstract

Transit signal priority (TSP) requires balancing competing objectives: reducing bus delay while limiting adverse impacts on non-bus traffic and avoiding extreme waits for a subset of vehicles. Existing reinforcement-learning (RL) approaches to TSP typically encode transit-aware features (e.g., occupancy and schedule deviation) but optimize a fixed reward or fixed scalarization, which limits operational flexibility when agency priorities change across time-of-day or disruption conditions. We present a preference-conditioned TSP controller, $\pi(a \mid s,w)$, that selects the next signal phase under minimum/maximum green and transition-feasibility constraints and can be tuned at runtime via a preference parameter $w$ to trade off bus-priority emphasis against overall traffic delay without retraining. We implement this on top of IntersectionZoo by introducing a constrained signal-control/TSP wrapper, and we extend scenario generation with bus-prevalence augmentation and timetable-based bus insertion to address sparse transit-priority events during training. Experiments against fixed-time control, a rule-based TSP overlay, and fixed-weight PPO specialists show that a single learned conditioned policy spans a smooth empirical trade-off frontier across runtime preferences, outperforms fixed-time and rule-based baselines, and maintains constraint feasibility, while tail-delay diagnostics reveal that non-bus externalities remain limited for moderate preference settings but can increase substantially under high bus-priority weights. The source code of this work is available at https://github.com/urbanAIthi/morl-tsp.

Figures

Figures reproduced from arXiv: 2607.18286 by the authors.

**Figure 1.** Pareto trade-off summaries across three case-study intersections (NYC10802, LA2114, CHI2412). Points show MORL mean outcomes across the 11- [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗

**Figure 2.** Distribution-shift evaluation (see Section II-D) across NYC10802, LA2114, and CHI2412. RC4 denotes the in-distribution timetable-augmented regime [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

**Figure 3.** Runtime tunability on NYC10802: mean bus and non-bus delay as [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

24 extracted references · 1 canonical work pages

[6]

Constrained traffic signal control under competing public transport priority requests via safe reinforcement learning,

R. Zhou, T. Nousch, L. Wei, and M. Wang, “Constrained traffic signal control under competing public transport priority requests via safe reinforcement learning,”Expert Systems with Applications, vol. 284, p. 127676, 2025. [Online]. Available: https://www.sciencedirect.com/ science/article/pii/S0957417425012989

2025
[1]

Transit signal priority (tsp): A planning and implementation handbook,

H. R. Smith, B. Hemily, and M. Ivanovic, “Transit signal priority (tsp): A planning and implementation handbook,” U.S. Department of Transportation, Washington, DC, Tech. Rep., 2005. [Online]. Available: https://nacto.org/wp-content/uploads/transit signal priority handbook smith.pdf

2005
[2]

Accessed 2026-05-13

American Association of State Highway and Transportation Officials, Institute of Transportation Engineers, and National Electrical Manufacturers Association,NTCIP 1211 v02: National Transportation Communications for ITS Protocol: Object Definitions for Signal Control and Prioritization (SCP), American Association of State Highway and Transportation Offici...

2014
[3]

Deep reinforcement learning for transit signal priority in a connected environment,

M. Long, X. Zou, Y . Zhou, and E. Chung, “Deep reinforcement learning for transit signal priority in a connected environment,” Transportation Research Part C: Emerging Technologies, vol. 142, p. 103814, 2022. [Online]. Available: https://www.sciencedirect.com/ science/article/pii/S0968090X22002388

2022
[4]

Deep reinforcement learning two-way transit signal priority algorithm for opti- mizing headway adherence and speed,

W. X. Hu, H. Ishihara, C. Chen, A. Shalaby, and B. Abdulhai, “Deep reinforcement learning two-way transit signal priority algorithm for opti- mizing headway adherence and speed,”IEEE Transactions on Intelligent Transportation Systems, vol. 24, no. 8, pp. 7920–7931, 2023

2023
[5]

Multi-agent reinforcement learning for co- operative transit signal priority to promote headway adherence,

M. Long and E. Chung, “Multi-agent reinforcement learning for co- operative transit signal priority to promote headway adherence,”IEEE Transactions on Intelligent Transportation Systems, vol. 26, no. 3, pp. 3588–3602, 2025

2025
[7]

Universal value function approximators,

T. Schaul, D. Horgan, K. Gregor, and D. Silver, “Universal value function approximators,” inProceedings of the 32nd International Conference on Machine Learning, ser. Proceedings of Machine Learning Research, F. Bach and D. Blei, Eds., vol. 37. Lille, France: PMLR, 07–09 Jul 2015, pp. 1312–1320. [Online]. Available: https://proceedings.mlr.press/v37/schaul15.html

2015
[8]

Dynamic weights in multi-objective deep reinforcement learning,

A. Abels, D. Roijers, T. Lenaerts, A. Now ´e, and D. Steckelmacher, “Dynamic weights in multi-objective deep reinforcement learning,” in Proceedings of the 36th International Conference on Machine Learning, ser. Proceedings of Machine Learning Research, K. Chaudhuri and R. Salakhutdinov, Eds., vol. 97. PMLR, 09–15 Jun 2019, pp. 11–20. [Online]. Available:...

2019

Show all 24 references

[9]

Pareto conditioned net- works,

M. Reymond, E. Bargiacchi, and A. Now ´e, “Pareto conditioned net- works,” inProceedings of the 21st International Conference on Au- tonomous Agents and Multiagent Systems, ser. AAMAS ’22. Richland, SC: International Foundation for Autonomous Agents and Multiagent Systems, 202...

2022
[10]

Reinforcement learning benchmarks for traffic signal control,

J. Ault and G. Sharon, “Reinforcement learning benchmarks for traffic signal control,” inProceedings of the Thirty-fifth Conference on Neural Information Processing Systems (NeurIPS 2021) Datasets and Bench- marks Track, December 2021

2021
[11]

Reinforcement learning for traffic signal control: Comparison with commercial systems,

A. Cabrejas-Egea, R. Zhang, and N. Walton, “Reinforcement learning for traffic signal control: Comparison with commercial systems,” Transportation Research Procedia, vol. 58, pp. 638–645, 2021, xIV Conference on Transport Engineering, CIT2021. [Online]. Available: https://www....

2021
[12]

Libsignal: an open library for traffic signal control,

H. Mei, X. Lei, L. Da, B. Shi, and H. Wei, “Libsignal: an open library for traffic signal control,”Mach. Learn., vol. 113, no. 8, p. 5235–5271, Nov
[13]

Intersectionzoo: Eco-driving for benchmarking multi- agent contextual reinforcement learning,

V . Jayawardana, B. Freydt, A. Qu, C. Hickert, Z. Yan, and C. Wu, “Intersectionzoo: Eco-driving for benchmarking multi- agent contextual reinforcement learning,” 2024. [Online]. Available: https://arxiv.org/abs/2410.15221

2024 arXiv
[14]

SUMO-RL,

L. N. Alegre, “SUMO-RL,” https://github.com/LucasAlegre/sumo-rl, 2019

2019
[15]

Microscopic traffic simulation using sumo,

P. A. Lopez, M. Behrisch, L. Bieker-Walz, J. Erdmann, Y .-P. Fl ¨otter¨od, R. Hilbrich, L. L ¨ucken, J. Rummel, P. Wagner, and E. Wiessner, “Microscopic traffic simulation using sumo,” in2018 21st International Conference on Intelligent Transportation Systems (ITSC), 2018, pp....

2018
[16]

Domain randomization for transferring deep neural networks from sim- ulation to the real world,

J. Tobin, R. Fong, A. Ray, J. Schneider, W. Zaremba, and P. Abbeel, “Domain randomization for transferring deep neural networks from sim- ulation to the real world,” in2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2017, pp. 23–30

2017
[17]

Bridging the reality gap of reinforcement learning based traffic signal control using domain randomization and meta learning,

A. M ¨uller and M. Sabatelli, “Bridging the reality gap of reinforcement learning based traffic signal control using domain randomization and meta learning,” in2023 IEEE 26th International Conference on Intelli- gent Transportation Systems (ITSC), 2023, pp. 5271–5278

2023
[18]

Generalight: Improving environment generalization of traffic signal control via meta reinforcement learning,

H. Zhang, C. Liu, W. Zhang, G. Zheng, and Y . Yu, “Generalight: Improving environment generalization of traffic signal control via meta reinforcement learning,” inProceedings of the 29th ACM International Conference on Information & Knowledge Management, ser. CIKM ’20. New Yor...

2020 doi
[19]

Optimization and simulation of fixed-time traffic signal control in real-world applications,

T. Thunig, R. Scheffler, M. Strehler, and K. Nagel, “Optimization and simulation of fixed-time traffic signal control in real-world applications,”Procedia Computer Science, vol. 151, pp. 826–833, 2019, the 10th International Conference on Ambient Systems, Networks and Technolo...

2019
[20]

Stable-baselines3: Reliable reinforcement learning implementations,

A. Raffin, A. Hill, A. Gleave, A. Kanervisto, M. Ernestus, and N. Dormann, “Stable-baselines3: Reliable reinforcement learning implementations,”Journal of Machine Learning Research, vol. 22, no. 268, pp. 1–8, 2021. [Online]. Available: https://jmlr.org/papers/v22/ 20-1364.html

2021
[22]

A generalized algorithm for multi-objective reinforcement learning and policy adaptation,

R. Yang, X. Sun, and K. Narasimhan, “A generalized algorithm for multi-objective reinforcement learning and policy adaptation,” in Advances in Neural Information Processing Systems, H. Wallach, H. Larochelle, A. Beygelzimer, F. d'Alch ´e-Buc, E. Fox, and R. Garnett, Eds., vol....

2019
[23]

A toolkit for reliable benchmarking and research in multi-objective reinforcement learning,

F. Felten, L. N. Alegre, A. Now ´e, A. L. C. Bazzan, E.-G. Talbi, G. Danoy, and B. C. da Silva, “A toolkit for reliable benchmarking and research in multi-objective reinforcement learning,” inProceedings of the 37th International Conference on Neural Information Processing Sys...

2023
[2017]

Available: http://arxiv.org/abs/1707.06347

[Online]. Available: http://arxiv.org/abs/1707.06347

arXiv
[2023]

Available: https://doi.org/10.1007/s10994-023-06412-y

[Online]. Available: https://doi.org/10.1007/s10994-023-06412-y

doi

Pith tools

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

[1] [6]

Constrained traffic signal control under competing public transport priority requests via safe reinforcement learning,

R. Zhou, T. Nousch, L. Wei, and M. Wang, “Constrained traffic signal control under competing public transport priority requests via safe reinforcement learning,”Expert Systems with Applications, vol. 284, p. 127676, 2025. [Online]. Available: https://www.sciencedirect.com/ science/article/pii/S0957417425012989

2025

[2] [1]

Transit signal priority (tsp): A planning and implementation handbook,

H. R. Smith, B. Hemily, and M. Ivanovic, “Transit signal priority (tsp): A planning and implementation handbook,” U.S. Department of Transportation, Washington, DC, Tech. Rep., 2005. [Online]. Available: https://nacto.org/wp-content/uploads/transit signal priority handbook smith.pdf

2005

[3] [2]

Accessed 2026-05-13

American Association of State Highway and Transportation Officials, Institute of Transportation Engineers, and National Electrical Manufacturers Association,NTCIP 1211 v02: National Transportation Communications for ITS Protocol: Object Definitions for Signal Control and Prioritization (SCP), American Association of State Highway and Transportation Offici...

2014

[4] [3]

Deep reinforcement learning for transit signal priority in a connected environment,

M. Long, X. Zou, Y . Zhou, and E. Chung, “Deep reinforcement learning for transit signal priority in a connected environment,” Transportation Research Part C: Emerging Technologies, vol. 142, p. 103814, 2022. [Online]. Available: https://www.sciencedirect.com/ science/article/pii/S0968090X22002388

2022

[5] [4]

Deep reinforcement learning two-way transit signal priority algorithm for opti- mizing headway adherence and speed,

W. X. Hu, H. Ishihara, C. Chen, A. Shalaby, and B. Abdulhai, “Deep reinforcement learning two-way transit signal priority algorithm for opti- mizing headway adherence and speed,”IEEE Transactions on Intelligent Transportation Systems, vol. 24, no. 8, pp. 7920–7931, 2023

2023

[6] [5]

Multi-agent reinforcement learning for co- operative transit signal priority to promote headway adherence,

M. Long and E. Chung, “Multi-agent reinforcement learning for co- operative transit signal priority to promote headway adherence,”IEEE Transactions on Intelligent Transportation Systems, vol. 26, no. 3, pp. 3588–3602, 2025

2025

[7] [7]

Universal value function approximators,

T. Schaul, D. Horgan, K. Gregor, and D. Silver, “Universal value function approximators,” inProceedings of the 32nd International Conference on Machine Learning, ser. Proceedings of Machine Learning Research, F. Bach and D. Blei, Eds., vol. 37. Lille, France: PMLR, 07–09 Jul 2015, pp. 1312–1320. [Online]. Available: https://proceedings.mlr.press/v37/schaul15.html

2015

[8] [8]

Dynamic weights in multi-objective deep reinforcement learning,

A. Abels, D. Roijers, T. Lenaerts, A. Now ´e, and D. Steckelmacher, “Dynamic weights in multi-objective deep reinforcement learning,” in Proceedings of the 36th International Conference on Machine Learning, ser. Proceedings of Machine Learning Research, K. Chaudhuri and R. Salakhutdinov, Eds., vol. 97. PMLR, 09–15 Jun 2019, pp. 11–20. [Online]. Available:...

2019