pith. sign in

arxiv: 2604.19813 · v1 · submitted 2026-04-17 · 💻 cs.GT · cs.MA· quant-ph

Evolution of Lane-Changing Behavior in Mixed Traffic: A Quantum Game Theory Approach

Sungyong Chung , Tina Radvand , Alireza Talebpour This is my paper

Pith reviewed 2026-05-10 08:04 UTC · model grok-4.3

classification 💻 cs.GT cs.MAquant-ph
keywords lane changingmixed trafficquantum game theoryentanglement parameterautomated vehiclescooperation rateevolutionary game theorywaymo dataset
0
0 comments X

The pith

A quantum entanglement parameter of 0.52 reproduces the 42 percent cooperation rate observed in real lane changes

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

This paper applies quantum game theory to model lane-changing decisions in mixed traffic with human drivers and automated vehicles. Classical evolutionary game theory assumes independent agent choices and predicts convergence to full cooperation, yet empirical data from thousands of interactions shows a stable 42 percent cooperation rate. By embedding latent correlations via an entanglement parameter in the Marinatto-Weber quantization scheme, the model matches the observed mixed equilibrium and enables simulations of how different AV algorithms influence human adaptation over time. A sympathetic reader cares because these simulations let stakeholders anticipate behavioral shifts at varying AV penetration rates without waiting for full deployment.

Core claim

The central claim is that quantizing the lane-changing game with the Marinatto-Weber scheme and a human entanglement parameter of |b|^2_HDV approximately 0.52 reproduces the empirical mixed equilibrium of 42 percent cooperation extracted from 7636 Waymo interactions via quantal response equilibrium, whereas classical models predict full cooperation; simulations of classical, entangled, and inverted AV strategies then show that human cooperation evolves differently depending on the AV algorithm and market penetration.

What carries the argument

The Marinatto-Weber quantization scheme with a single entanglement parameter, which embeds latent correlations directly into the payoff structure of the lane-changing interaction game.

Load-bearing premise

A single entanglement parameter in the Marinatto-Weber scheme is sufficient to capture all relevant latent correlations in human lane-changing decisions and generalizes beyond the calibration dataset.

What would settle it

Introduce automated vehicles running known classical or inverted strategies into real mixed traffic at controlled low or high penetration rates and measure whether the human lane-change cooperation rate shifts exactly as the model predicts.

Figures

Figures reproduced from arXiv: 2604.19813 by Alireza Talebpour, Sungyong Chung, Tina Radvand.

Figure 1
Figure 1. Figure 1: FIGURE 1: Quantum circuit diagram for the EWL game quantization scheme [15]. This scheme uses an [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIGURE 2: Quantum circuit diagram for the MW game quantization scheme [8]. This scheme removes the [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIGURE 3: Calibration of the HDV-HDV entanglement parameter. The solid blue line represents the mean [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIGURE 4: Evolution of cooperation in 100% HDV traffic under different entanglement regimes. The [PITH_FULL_IMAGE:figures/full_fig_p012_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIGURE 5: Comparison of evolutionary dynamics among classical AVs (dashed lines), entangled AVs [PITH_FULL_IMAGE:figures/full_fig_p013_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIGURE 6: Sensitivity analysis for Case 1 (Classical AVs). Box plots show the distribution of the final [PITH_FULL_IMAGE:figures/full_fig_p014_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIGURE 7: Sensitivity analysis for Case 2 (Entangled AVs). Box plots show the distribution of the final [PITH_FULL_IMAGE:figures/full_fig_p015_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: FIGURE 8: Sensitivity analysis for Case 3 (Inverted AVs). Box plots show the distribution of the final [PITH_FULL_IMAGE:figures/full_fig_p016_8.png] view at source ↗
read the original abstract

As automated vehicles (AVs) enter mixed traffic, proactively anticipating the evolution of human driving behavior during critical interactions, such as lane changes, is essential. However, classical Evolutionary Game Theory (EGT) fails to capture the complexity of human decision-making during lane changes. Specifically, by strictly assuming independence between agents, classical models calibrated on empirical payoffs predict a convergence to unrealistic full cooperation, contradicting the stable 42% cooperation rate observed in real-world data. To resolve this discrepancy, this study introduces a Quantum Game Theory (QGT) framework. We analyze 7,636 lane-changing interactions from the Waymo Open Motion Dataset (WOMD) to derive empirical payoff matrices via a Quantal Response Equilibrium (QRE) model. Utilizing the Marinatto-Weber (MW) quantization scheme, we introduce an entanglement parameter to mathematically embed latent correlations directly into the payoff structure of a single interaction. Our results identify a human entanglement parameter of $|b|^2_{HDV} \approx 0.52$ that accurately reproduces the observed mixed equilibrium. Furthermore, simulations of three AV deployment strategies (classical, entangled, and inverted) reveal that human adaptation depends critically on the underlying AV algorithm: while cooperative classical AVs maximize system-wide cooperation at high market penetration rates, defective inverted AVs paradoxically yield higher overall cooperation at low penetration rates by prompting more cooperative behaviors from human drivers. Consequently, rather than waiting for large scale deployment to observe these effects, stakeholders can utilize this framework to simulate repeated interactions and proactively anticipate how human driver behavior will evolve in response to specific AV software designs.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper claims that classical evolutionary game theory (EGT), when calibrated on empirical payoffs from 7,636 Waymo lane-change interactions via quantal response equilibrium (QRE), incorrectly converges to full cooperation, contradicting the observed stable 42% cooperation rate. It introduces a quantum game theory (QGT) model using the Marinatto-Weber quantization scheme with a single entanglement parameter |b|^2_HDV ≈ 0.52 that reproduces the mixed equilibrium; simulations of classical, entangled, and inverted AV deployment strategies then show that human adaptation depends on the AV algorithm and penetration rate.

Significance. If the QGT reproduction is shown to be non-tautological and the entanglement parameter generalizes, the framework could enable proactive simulation of human-AV behavioral evolution for traffic planning. The empirical grounding in Waymo data and the three AV strategy simulations provide concrete, testable outputs that could inform AV software design, though the single-parameter fit limits claims of explanatory power beyond classical models.

major comments (3)
  1. [Abstract / Results] Abstract and results: The central claim that |b|^2_HDV ≈ 0.52 'accurately reproduces the observed mixed equilibrium' is achieved by direct calibration of the entanglement parameter to the target 42% cooperation statistic derived from the same Waymo dataset (after QRE payoff estimation). This makes the QGT match tautological by construction, as the Marinatto-Weber scheme adds exactly the degree of freedom needed to fit the discrepancy that classical EGT lacks; the paper must demonstrate that the parameter value emerges from independent data or out-of-sample prediction rather than post-hoc fitting.
  2. [Methods] Methods section on quantization: The assumption that the single Marinatto-Weber entanglement parameter suffices to embed all relevant latent correlations in human lane-changing decisions is load-bearing for both the equilibrium reproduction and the subsequent AV simulations. No robustness checks (e.g., alternative quantization schemes, multi-parameter extensions, or cross-dataset validation) are described to rule out that the parameter merely compensates for unmodeled classical factors such as incomplete information or heterogeneous driver types.
  3. [Simulations] AV simulation results: The reported differences in system-wide cooperation under classical vs. inverted AV strategies at varying penetration rates inherit the fitted |b|^2_HDV value without reported sensitivity analysis. If the entanglement parameter is varied by ±0.1 around 0.52, do the qualitative rankings of the three AV strategies remain stable, or do they reverse?
minor comments (2)
  1. [Abstract] Notation: The subscript in |b|^2_HDV is written inconsistently (sometimes with space as HD V); standardize and define the symbol explicitly on first use.
  2. [Methods] Clarity: The transition from QRE-derived payoffs to the quantized game matrix should include an explicit equation showing how the entanglement operator modifies the classical payoff entries.

Simulated Author's Rebuttal

3 responses · 1 unresolved

We thank the referee for the constructive and detailed comments, which identify key issues in the calibration, assumptions, and robustness of our quantum game-theoretic model. We address each major comment point by point below, indicating revisions where we agree changes are needed.

read point-by-point responses

  1. Referee: [Abstract / Results] Abstract and results: The central claim that |b|^2_HDV ≈ 0.52 'accurately reproduces the observed mixed equilibrium' is achieved by direct calibration of the entanglement parameter to the target 42% cooperation statistic derived from the same Waymo dataset (after QRE payoff estimation). This makes the QGT match tautological by construction, as the Marinatto-Weber scheme adds exactly the degree of freedom needed to fit the discrepancy that classical EGT lacks; the paper must demonstrate that the parameter value emerges from independent data or out-of-sample prediction rather than post-hoc fitting.

    Authors: We agree that the entanglement parameter is calibrated directly to the observed 42% cooperation rate from the Waymo data after QRE payoff estimation. This is not entirely tautological, however, because the classical EGT model with the same empirically derived payoffs converges to full cooperation, failing to match the data; the MW scheme supplies a structured mechanism for incorporating decision correlations absent from classical models. The resulting value of 0.52 offers an interpretable quantification of effective entanglement in human lane changes. We lack independent datasets to derive the parameter a priori or perform out-of-sample tests at present. In revision we will qualify the abstract and results language to describe the parameter as calibrated rather than claiming it 'accurately reproduces' without qualification, and we will add discussion of the need for future validation on other traffic datasets. revision: partial

  2. Referee: [Methods] Methods section on quantization: The assumption that the single Marinatto-Weber entanglement parameter suffices to embed all relevant latent correlations in human lane-changing decisions is load-bearing for both the equilibrium reproduction and the subsequent AV simulations. No robustness checks (e.g., alternative quantization schemes, multi-parameter extensions, or cross-dataset validation) are described to rule out that the parameter merely compensates for unmodeled classical factors such as incomplete information or heterogeneous driver types.

    Authors: We concur that the single-parameter MW assumption is central and that the absence of robustness checks is a limitation. The MW scheme was selected for its parsimonious extension of classical games. In the revised manuscript we will expand the methods section to explicitly discuss this modeling choice, its potential to proxy unmodeled classical effects, and possible alternatives such as the Eisert-Wilkens-Lewenstein scheme or multi-parameter extensions. We will also note that cross-dataset validation is desirable but outside the scope of the current WOMD-focused study. revision: yes

  3. Referee: [Simulations] AV simulation results: The reported differences in system-wide cooperation under classical vs. inverted AV strategies at varying penetration rates inherit the fitted |b|^2_HDV value without reported sensitivity analysis. If the entanglement parameter is varied by ±0.1 around 0.52, do the qualitative rankings of the three AV strategies remain stable, or do they reverse?

    Authors: We welcome the request for sensitivity analysis. We will add this analysis to the simulations section by varying |b|^2_HDV by ±0.1 around the fitted value of 0.52, recomputing the system-wide cooperation levels for the three AV strategies at different penetration rates, and reporting whether the qualitative rankings (classical AVs maximizing cooperation at high penetration, inverted AVs at low penetration) remain stable. Any changes in ordering or effect sizes will be discussed. revision: yes

standing simulated objections not resolved
  • Demonstrating that the entanglement parameter value emerges from independent data or out-of-sample prediction rather than post-hoc fitting to the Waymo dataset, as no additional independent lane-change datasets are available in the current study.

Circularity Check

1 steps flagged

Fitting |b|^2_HD V ≈ 0.52 to match the 42% cooperation rate makes the QGT 'reproduction' of the mixed equilibrium tautological by construction.

specific steps
  1. fitted input called prediction [Abstract (results claim)]
    "Our results identify a human entanglement parameter of |b|^2_HD V ≈ 0.52 that accurately reproduces the observed mixed equilibrium."

    The parameter |b|^2_HD V is chosen to match the empirically observed 42% cooperation rate from the Waymo dataset. The 'accurate reproduction' of the mixed equilibrium is therefore achieved by construction once this fitted value is inserted into the MW-quantized payoff structure; it is not a prediction independent of the calibration target.

full rationale

The derivation chain proceeds from empirical payoffs (via QRE on Waymo data) to the MW quantization scheme, where a single entanglement parameter is introduced to embed latent correlations. The paper then states that this parameter value is identified such that the quantized game reproduces the observed 42% cooperation rate (the stable mixed equilibrium that classical EGT cannot match). Because the parameter is calibrated directly to that target statistic, the claimed accurate reproduction reduces to the fitted input by the model's own equations; the match is forced rather than independently verified. This is the central load-bearing step for all subsequent AV-strategy simulations. No other circular patterns (self-citation chains, ansatz smuggling, or renaming) are present in the provided text, and the MW scheme itself is an external reference. The circularity is therefore partial but significant for the strongest claim.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 1 invented entities

The claim rests on empirical payoff derivation via QRE, the applicability of the MW scheme, and a single fitted entanglement value; no independent evidence for the entanglement construct is supplied beyond matching this dataset.

free parameters (1)
  • entanglement parameter |b|^2_HD V = 0.52
    Chosen to reproduce the 42% cooperation rate observed in WOMD lane-change data.
axioms (2)
  • domain assumption Quantal Response Equilibrium accurately captures human payoff perception in lane changes
    Used to convert observed trajectories into payoff matrices.
  • ad hoc to paper Marinatto-Weber quantization embeds latent correlations via a single parameter
    Introduced in the paper to resolve the independence assumption of classical EGT.
invented entities (1)
  • entanglement parameter representing latent driver correlations no independent evidence
    purpose: To allow non-classical mixed equilibria in single-interaction lane-change games
    No external falsifiable prediction or independent dataset validation is provided.

pith-pipeline@v0.9.0 · 5599 in / 1440 out tokens · 36174 ms · 2026-05-10T08:04:27.577359+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

21 extracted references · 21 canonical work pages

  1. [1]

    An, G., J. H. Bae, and A. Talebpour, An optimized car-following behavior in response to a lane- changing vehicle: A Bézier curve-based approach.IEEE Open Journal of Intelligent Transportation Systems, V ol. 4, 2023, pp. 682–689

    work page 2023

  2. [2]

    An, G. and A. Talebpour, Vehicle platooning for merge coordination in a connected driving envi- ronment: A hybrid ACC-DMPC approach.IEEE Transactions on Intelligent Transportation Systems, V ol. 24, No. 5, 2023, pp. 5239–5248

    work page 2023

  3. [3]

    Rahmati, Y ., M. K. Hosseini, and A. Talebpour, Helping automated vehicles with left-turn maneuvers: A game theory-based decision framework for conflicting maneuvers at intersections.IEEE Transac- tions on Intelligent Transportation Systems, V ol. 23, No. 8, 2021, pp. 11877–11890

    work page 2021

  4. [4]

    Chung, S., D. Ka, Y . Kim, and C. Lee, Gap setting control strategy for connected and automated vehicles in freeway lane-drop bottlenecks.IET Intelligent Transport Systems, V ol. 18, No. 12, 2024, pp. 2641–2659

    work page 2024

  5. [5]

    Talebpour, and S

    Chung, S., A. Talebpour, and S. H. Hamdar, Characterizing Lane-Changing Behavior in Mixed Traffic. arXiv preprint arXiv:2512.07219, 2025

    work page arXiv 2025

  6. [6]

    Cheng, B

    Ettinger, S., S. Cheng, B. Caine, C. Liu, H. Zhao, S. Pradhan, Y . Chai, B. Sapp, C. R. Qi, Y . Zhou, Z. Yang, A. Chouard, P. Sun, J. Ngiam, V . Vasudevan, A. McCauley, J. Shlens, and D. Anguelov, Large Scale Interactive Motion Forecasting for Autonomous Driving: The Waymo Open Motion Dataset. In Proceedings of the IEEE/CVF International Conference on Com...

    work page 2021

  7. [7]

    Khan, F. S., N. Solmeyer, R. Balu, and T. S. Humble, Quantum games: a review of the history, current state, and interpretation.arXiv preprint arXiv:1803.07919, 2018

    work page arXiv 2018

  8. [8]

    Marinatto, L. and T. Weber, A quantum approach to static games of complete information.Physics Letters A, V ol. 272, No. 5-6, 2000, pp. 291–303

    work page 2000

  9. [9]

    Iqbal, A. and A. H. Toor, Evolutionarily stable strategies in quantum games.Physics Letters A, V ol. 280, No. 5-6, 2001, pp. 249–256

    work page 2001

  10. [10]

    Transportation Research Part A: Policy and Practice, V ol

    Kita, H., A merging–giveway interaction model of cars in a merging section: a game theoretic analysis. Transportation Research Part A: Policy and Practice, V ol. 33, No. 3-4, 1999, pp. 305–312

    work page 1999

  11. [11]

    Tanimoto, and K

    Kita, H., K. Tanimoto, and K. Fukuyama, A game theoretic analysis of merging-giveway interaction: a joint estimation model.Transportation and Traffic Theory in the 21st Century, 2002, pp. 503–518

    work page 2002

  12. [12]

    Talebpour, A., H. S. Mahmassani, and S. H. Hamdar, Modeling lane-changing behavior in a connected environment: A game theory approach.Transportation Research Procedia, V ol. 7, 2015, pp. 420–440. 17

    work page 2015

  13. [13]

    Kukida, and A

    Tanimoto, J., S. Kukida, and A. Hagishima, Social dilemma structures hidden behind traffic flow with lane changes.Journal of Statistical Mechanics: Theory and Experiment, V ol. 2014, No. 7, 2014, p. P07019

    work page 2014

  14. [14]

    O., An introduction to quantum game theory.arXiv preprint quant-ph/0506219, 2005

    Grabbe, J. O., An introduction to quantum game theory.arXiv preprint quant-ph/0506219, 2005

    work page arXiv 2005

  15. [15]

    Eisert, J. and M. Wilkens, Quantum games.Journal of Modern Optics, V ol. 47, No. 14-15, 2000, pp. 2543–2556

    work page 2000

  16. [16]

    A., Quantum strategies.Physical Review Letters, V ol

    Meyer, D. A., Quantum strategies.Physical Review Letters, V ol. 82, No. 5, 1999, p. 1052

    work page 1999

  17. [17]

    Debreu, G., A social equilibrium existence theorem.Proceedings of the National Academy of Sciences, V ol. 38, No. 10, 1952, pp. 886–893

    work page 1952

  18. [18]

    J., Subjectivity and correlation in randomized strategies.Journal of Mathematical Eco- nomics, V ol

    Aumann, R. J., Subjectivity and correlation in randomized strategies.Journal of Mathematical Eco- nomics, V ol. 1, No. 1, 1974, pp. 67–96

    work page 1974

  19. [19]

    F., Equilibrium points in n-person games.Proceedings of the National Academy of Sciences, V ol

    Nash, J. F., Equilibrium points in n-person games.Proceedings of the National Academy of Sciences, V ol. 36, No. 1, 1950, pp. 48–49

    work page 1950

  20. [20]

    McKelvey, R. D. and T. R. Palfrey, Quantal Response Equilibria for Normal Form Games.Games and Economic Behavior, V ol. 10, No. 1, 1995, pp. 6–38

    work page 1995

  21. [21]

    Nowak, M. A. and R. M. May, Evolutionary games and spatial chaos.Nature, V ol. 359, No. 6398, 1992, pp. 826–829. 18

    work page 1992