Historically, urban infrastructure has been designed and operated in silos. Transportation engineers optimized roads, transit timetables, and traffic signals to minimize travel latency, while power systems engineers managed substations, transformers, and transmission lines to ensure grid stability and voltage regulation.
The mass adoption of electric vehicles (EVs) shatters this separation. An EV is not merely a transportation asset; it is a massive, mobile electrical load. When thousands of vehicles move through a city and plug into fast-charging stations, they dynamically shift demand across the electrical grid. If a traffic management system routes vehicles purely to minimize travel time, it may inadvertently steer hundreds of EVs to the same charging hub, causing localized grid overloads or voltage drops. Conversely, if a grid operator raises electricity prices at a critical node to protect a vulnerable transformer, they may push EVs onto alternative routes, triggering traffic gridlock.
To solve this, a growing body of research focuses on the joint optimization of coupled transportation and power distribution networks. By treating these two distinct systems as a single, interdependent network, operators can prevent bottlenecks in both domains simultaneously.
Modeling the Interdependence
The fundamental challenge of joint optimization is bridging two entirely different mathematical worlds: the continuous power flow equations of electrical grids and the discrete routing or queuing models of transportation networks. As noted by Serdar Çolak et al. in Planning for electric vehicle needs by coupling charging profiles with urban mobility, failing to couple urban mobility data with energy consumption patterns prevents the development of holistic strategies for managing large-scale electrification. Indeed, Xing Lu et al. demonstrated in An Open Source Modeling Framework for Interdependent Energy-Transportation-Communication Infrastructure in Smart and Connected Communities that modeling these systems independently introduces substantial errors—up to 10.5% in traffic speed predictions and 7% in power grid demand.
To couple these systems, researchers often construct multi-layer or extended network graphs. For example, Mahnoosh Alizadeh et al. introduced an "Extended Transportation Graph" in Optimal Pricing to Manage Electric Vehicles in Coupled Power and Transportation Networks to model joint routing and charging decisions, allowing operators to see exactly how a vehicle's physical path on the road dictates its charging location on the grid.
Similarly, LiangSong Guo et al. developed a two-layer optimization framework in A multi-objective spatio-temporal pricing method for fast-charging stations oriented to transformer load balancing. The first layer simulates user routing and preferences on the road network, while the second layer calculates grid stability metrics like voltage and load variance. This structural coupling allows operators to visualize and manage the complex interdependencies in real time, a capability also emphasized by Avishan Bagherinezhad et al. in Spatio-Temporal Visualization of Interdependent Battery Bus Transit and Power Distribution Systems.
Operational Co-Optimization: Routing, Scheduling, and Pricing
Once the networks are coupled, how do we influence driver behavior or fleet schedules to protect both the grid and the road? The primary mechanism is spatio-temporal coordination, often driven by dynamic pricing or decentralized game theory.
Instead of flat electricity rates, prices are varied based on both when and where a vehicle plugs in. Mahnoosh Alizadeh et al. proposed collaborative pricing schemes that allow power and traffic operators to reach a socially optimal equilibrium without compromising their private operational data. From the driver's perspective, S. Rasoul Etesami et al. modeled this in Smart Routing of Electric Vehicles for Load Balancing in Smart Grids as a decentralized routing game. Selfish players navigate a complex trade-off between traffic latency, charging wait times, and energy costs. The authors proved that as the number of vehicles grows, the system naturally converges to a highly efficient state that balances grid loads.
For managed fleets, such as public transit networks or autonomous mobility-on-demand (AMoD) systems, the coordination must be even tighter. Traditional transit management treats vehicle scheduling and charging infrastructure as independent tasks. However, Maximilian Schiffer et al. argued in On the Integration of Battery Electric Buses into Urban Bus Networks that this creates a circular dependency: charging station placement requires knowledge of schedules, while schedules depend on charger availability.
[Siloed Planning]
Vehicle Schedules ---> Fixed Constraints ---> Charging Infrastructure (Suboptimal)
[Co-Optimized Planning]
Vehicle Schedules <=========================> Charging Infrastructure (Globally Optimal)
To break this loop, researchers co-optimize schedules and energy constraints simultaneously. Amutheezan Sivagnanam et al. demonstrated in Minimizing Energy Use of Mixed-Fleet Public Transit for Fixed-Route Service that integrating electric vehicle range and limited charging windows directly into fixed transit schedules prevents service disruptions while minimizing energy costs.
Taking this a step further, Rémi Lacombe et al. developed a hierarchical control framework in Integrated Charging Scheduling and Operational Control for an Electric Bus Network that jointly optimizes bus charging schedules alongside real-time operational controls like speed adjustments and stop holding times. This deep coupling ensures that a delay on the road does not cascade into a battery depletion crisis or a grid peak-load violation. This need for multi-modal infrastructure coordination is echoed by Ramón Iglesias et al. in Data-Driven Model Predictive Control of Autonomous Mobility-on-Demand Systems, who highlighted that autonomous fleet rebalancing must be integrated with broader urban transit and grid constraints to maximize city-wide efficiency.
Strategic Co-Optimization: Infrastructure Planning
Beyond daily operations, long-term capital investments—such as where to build charging stations and where to upgrade power lines—must also be co-optimized. As reviewed by Jinhao Li et al. in Investigating state-of-the-art planning strategies for electric vehicle charging infrastructures in coupled transport and power networks: a comprehensive review, placing charging infrastructure without considering the existing electrical grid's capabilities leads to low utilization, inadequate service, or localized grid failures.
To solve this, researchers deploy bi-level optimization models. In these frameworks, the upper-level problem typically determines the strategic placement and capacity of charging stations, while the lower-level problem simulates how drivers or fleets will respond operationally to that infrastructure.
For instance, Linwei Li proposed a bi-level model in Electric vehicle charging station planning based on the development of distribution networks and coupled charging demand that simultaneously plans power distribution network expansions and charging station placements. By incorporating real-world traffic congestion into demand forecasting, this coordinated approach ensures that grid infrastructure grows alongside mobile demand, keeping voltage deviations within safe operating limits.
[Bi-Level Optimization Framework]
┌────────────────────────────────────────────────────────┐
│ Upper Level: Strategic Planning │
│ - Siting of Charging Stations │
│ - Power Grid Distribution Network Expansion │
└───────────────────────────┬────────────────────────────┘
│ Infrastructure Constraints
▼
┌────────────────────────────────────────────────────────┐
│ Lower Level: Operational Simulation │
│ - Driver Routing Decisions & Traffic Congestion │
│ - Real-Time Grid Load & Voltage Stability │
└────────────────────────────────────────────────────────┘
Similarly, Xian Zhang et al. developed a framework in Yen’s Algorithm-Based Charging Facility Planning Considering Congestion in Coupled Transportation and Power Systems that simulates realistic driver detours for charging. Their model minimizes total system costs by reducing electrical grid losses and traffic congestion delays simultaneously.
Finally, Justin Luke et al. proved the tangible benefits of this approach in Joint Optimization of Autonomous Electric Vehicle Fleet Operations and Charging Station Siting. By co-designing infrastructure and day-to-day fleet routing, they demonstrated that operators can reduce total costs, empty-vehicle travel, and peak grid loads by up to 10% compared to traditional, decoupled planning methods.
Why It Works
The core intuition behind the joint optimization of these networks is simple: it transforms a blind, reactive system into an anticipatory, coordinated one.
When systems are optimized in isolation, they treat the constraints of the neighboring network as static, unpredictable disruptions. A traffic jam at a charging station or a voltage drop at a substation is handled only after it occurs. By mathematically coupling the networks, the optimization algorithms can exploit the flexibility of one system to relieve stress on the other.
As summarized by M. Hadi Amini et al. in Distributed Holistic Framework for Smart City Infrastructures: Tale of Interdependent Electrified Transportation Network and Power Grid, coordinating these infrastructures through shared agents allows routing decisions and charging strategies to be optimized simultaneously. Vehicles are guided to chargers that are not only geographically convenient but also electrically stable, ensuring that the clean energy transition does not come at the cost of urban gridlock or blackouts.
Key Foundational Work
Optimal Pricing to Manage Electric Vehicles in Coupled Power and Transportation Networks (2017) — contributed 1 concept elements to this direction.
Impact of Interdisciplinary Research on Planning, Running, and Managing Electromobility as a Smart Grid Extension (2015) — contributed 1 concept elements to this direction.
Marginal charging station pricing in an intelligent electric transportation system (2017) — contributed 1 concept elements to this direction.