Urban Congestion Patterns under High Electric Vehicle Penetration: A Case Study of 10 U.S. Cities

With the global energy transition and the rapid penetration of electric vehicles (EVs), the widening travel cost gap between EVs and gasoline vehicles (GVs) increasingly affects commuters’ route choices and may reshape urban congestion patterns. Existing research remains in its preliminary exploratory phase. On the one hand, multi-class models do not account for fixed user class scenarios, which may not align with actual commuters; on the other hand, there is a lack of systematic quantitative analysis based on real-world complex road networks across multiple cities. As a result, the congestion effects induced by heterogeneous GV-EV cost structures may be mischaracterized or substantially underestimated. To address these limitations, this paper proposes a multi-user equilibrium (MUE) assignment model for mixed GV-EV traffic, constructs a dual algorithm with convergence guarantees, and designs multi-dimensional evaluation metrics for congestion patterns. Using 10 representative U.S. cities as a case study, this research explores the evolution trends of traffic congestion under different EV penetration scenarios based on real city-level road networks and block-level commuter origin-destination (OD) demand. The results show that full EV penetration reduces average system travel time by 2.27%–10.78% across the 10 cities, with New Orleans achieving the largest reduction (10.78%) and San Francisco the smallest (2.27%), but the effectiveness of alleviating congestion exhibits urban heterogeneity. Moreover, for cities with sufficient network redundancy, benefits are primarily concentrated during the low to medium EV penetration stage (0-0.5), though cities with topological constraints (e.g., San Francisco) show more limited improvements throughout all penetration levels. This paper can provide a foundation for formulating differentiated urban planning and congestion management policies.

💡 Research Summary

This paper investigates how the rapid penetration of electric vehicles (EVs) reshapes urban traffic congestion patterns by exploiting the cost differential between EVs and gasoline vehicles (GVs). Recognizing that existing multi‑class traffic assignment models typically allow users to switch classes—a premise that does not reflect the reality of fixed vehicle ownership—the authors develop a fixed‑class Multi‑User Equilibrium (MUE) model tailored to mixed GV‑EV traffic. In this formulation, the vehicle class of each traveler is exogenously fixed, while route choice follows the classic user‑equilibrium condition: every traveler selects a path whose generalized cost (including travel time, fuel/energy cost, and charging considerations) is minimal among all feasible routes. Under standard convexity and continuity assumptions, the authors prove existence and uniqueness of the equilibrium solution.

To solve the resulting variational inequality, they propose a dual‑based iterative algorithm. By introducing Lagrange multipliers for the flow conservation constraints, the algorithm alternates between solving a linearized primal subproblem (via a Frank‑Wolfe line search) and updating the dual variables. A bi‑conjugate Frank‑Wolfe scheme ensures that each iteration reduces the duality gap, and the authors provide a formal convergence proof showing an O(1/k) reduction in the primal‑dual objective difference after k iterations.

Beyond the methodological contribution, the paper designs a six‑dimensional congestion‑pattern evaluation framework: (1) average travel time (AT), (2) potential savings, (3) volume‑over‑capacity (VOC), (4) road utilization rate (RUR), (5) link congested time (LCT), and (6) difference in delay factor (ΔDelay). These metrics capture system‑wide efficiency, link‑level saturation, and spatial distribution of congestion, enabling a nuanced comparison across scenarios.

Empirically, the authors apply the model to ten heterogeneous U.S. metropolitan areas (including New Orleans, San Francisco, Boston, Atlanta, and Chicago). Realistic road networks are extracted from OpenStreetMap, and block‑level commuter origin‑destination (OD) matrices are derived from U.S. Census data. EV penetration rates are varied from 0 % to 100 % in 10 % increments, and for each scenario the MUE is recomputed.

Key findings are:

1. Full EV penetration reduces system‑wide average travel time by 2.27 %–10.78 % across the ten cities. New Orleans exhibits the largest reduction (10.78 %), while San Francisco shows the smallest (2.27 %).
2. Benefits are not linear. In cities with ample network redundancy (e.g., New Orleans, Atlanta), the bulk of travel‑time savings occurs during the low‑to‑medium penetration window (30 %–50 % EV share). Beyond this range, marginal gains diminish as the network approaches capacity equilibrium.
3. In topologically constrained cities (San Francisco, Boston), improvements are modest throughout all penetration levels. Bottlenecks such as bridges, tunnels, and limited alternative routes limit the ability of lower‑cost EVs to divert traffic away from congested links.
4. VOC and RUR analyses reveal a “re‑distribution effect”: while major arterials experience reduced overload, low‑capacity local streets may see increased congestion as EV drivers exploit cheaper routes. This underscores the need for spatially targeted policy measures.
5. Link‑level metrics (LCT, ΔDelay) highlight that EV adoption does not automatically alleviate congestion at critical choke points; without complementary infrastructure (e.g., dedicated EV lanes or strategically placed fast‑charging stations), the cost advantage of EVs may be insufficient to shift traffic away from these nodes.

Policy implications drawn from the study include: (a) cities with high network redundancy should prioritize early‑stage EV incentives and charging infrastructure to capture the steep part of the benefit curve; (b) constrained cities require supplemental interventions such as congestion pricing, EV‑only lanes, or targeted capacity upgrades; (c) multi‑dimensional congestion metrics should guide the placement of charging stations to avoid exacerbating local bottlenecks.

In conclusion, the paper makes four major contributions: (1) a theoretically sound fixed‑class MUE model that respects real vehicle ownership patterns; (2) a convergent dual algorithm with provable performance guarantees; (3) a comprehensive set of congestion evaluation metrics; and (4) a large‑scale empirical assessment across ten U.S. cities that reveals pronounced heterogeneity in EV‑induced congestion relief. The authors suggest future work to integrate vehicle‑to‑grid interactions, autonomous EV behavior, and dynamic, real‑time demand modeling to further enrich the analytical framework.