⚡ electrical engineering

Braess' Paradoxes in Coupled Power and Transportation Systems

This paper generalizes Braess' paradox to coupled power and transportation systems by demonstrating how capacity expansions in either network can degrade overall performance due to cross-system coupling, while also deriving conditions for these paradoxes and proposing pricing policies to mitigate them.

Original authors: Minghao Mou, Junjie Qin

Published 2026-04-20

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Minghao Mou, Junjie Qin

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are managing a city with two massive, interconnected systems: the Roads (where people drive) and the Power Grid (where electricity flows to charge their electric cars).

For a long time, city planners treated these two systems separately. They thought, "If we build a new road, traffic will get better. If we build a stronger power line, electricity will be cheaper."

But this paper reveals a surprising truth: Sometimes, making one system stronger actually makes the whole city worse off.

This is a phenomenon known as Braess' Paradox. The authors of this paper show that when you couple electric vehicles (EVs) with the power grid, this paradox gets even stranger and more complex.

Here is the breakdown of their discovery using simple analogies.

1. The Classic "Braess' Paradox" (The Shortcut Trap)

Imagine a city with two routes to get to work:

Route A: A wide highway that gets very crowded.
Route B: A narrow, quiet backroad.

Everyone naturally chooses the route that gets them there fastest. If the highway is too slow, some people switch to the backroad until both routes take the same amount of time. This is the "equilibrium."

Now, imagine the city builds a super-fast shortcut connecting the two routes. You'd think this would help everyone, right?

The Paradox: Because everyone is selfish and wants the fastest route, they all rush to use the new shortcut. This clogs up the shortcut and the roads leading to it. Suddenly, everyone is slower than they were before the shortcut existed.

The Lesson: Adding capacity (a new road) doesn't always improve performance; it can sometimes create a traffic jam that didn't exist before.

2. The New Twist: The "Electric Car" Effect

In this paper, the authors add a second layer: Electric Vehicles.
When you drive an EV, you don't just care about traffic; you also care about where you charge and how much electricity costs.

The Power Grid: Think of the power grid like a water system. If too many people try to fill their buckets (charge cars) at the same tap (charging station), the water pressure drops, and the cost of water (electricity price) goes up.
The Feedback Loop:
1. Drivers choose a route to avoid traffic.
2. That route leads to a specific charging station.
3. If too many cars go there, the local electricity price spikes.
4. Drivers see the high price and switch to a different route with a cheaper charger.
5. This changes the traffic pattern, which changes the electricity price again.

3. The Four Types of "Paradoxes"

The authors discovered that because these two systems are linked, a change in one can hurt the other in four specific ways. They call these Type T-T, T-P, P-T, and P-P.

Let's use a metaphor of a Restaurant and a Highway:

The Highway = Transportation System.
The Restaurant = Power System (where you get your "fuel").

Here are the four weird scenarios they found:

A. Type T-T (Roads hurting Roads)

The Scenario: You build a new road to help traffic.
The Result: Traffic gets worse on the roads.
Why? The new road leads to a charging station that is slightly cheaper. Everyone rushes there, clogging the road and the charger. It's the classic paradox, but now the "trap" is the electricity price.

B. Type T-P (Roads hurting the Power Grid)

The Scenario: You build a new road to help traffic.
The Result: The Power Grid gets more expensive and inefficient, even though the roads might be fine.
Why? The new road lures drivers to a charging station connected to a "weak" part of the power grid. The grid has to work overtime to supply that area, driving up the cost of electricity for everyone, even though the drivers are happy with their route.
Analogy: You built a new highway to a popular restaurant. The restaurant is so overwhelmed it has to buy expensive ingredients from far away, raising the price of food for the whole city, even though the drive was easy.

C. Type P-T (Power Grid hurting Roads)

The Scenario: You upgrade a power line to make electricity cheaper/better.
The Result: Traffic gets worse.
Why? The cheaper electricity attracts a massive crowd of EVs to a specific area. That area was already a bit congested, but now it's a parking lot. The "cheap power" incentive caused a traffic jam.
Analogy: You lowered the price of gas at a specific gas station. Everyone drove there to fill up, causing a massive traffic jam that made everyone late, even though the gas was cheaper.

D. Type P-P (Power Grid hurting the Power Grid)

The Scenario: You upgrade a power line.
The Result: The Power Grid becomes less efficient and more expensive.
Why? This is the most counter-intuitive one. The new power line changes the "price map" in a way that tricks drivers into charging at the most expensive, inefficient power plants. The grid operator ends up paying more to generate the same amount of power.
Analogy: You widened a pipe to a water tower. But because the water pressure changed, the town's pumps started pulling water from a dirty, expensive well instead of the clean, cheap lake. The town now pays more for water, even though the pipe is bigger.

4. How Do We Fix It? (The "Traffic Cop" Solution)

If building things can make things worse, what do we do? The authors suggest Smart Pricing.

Imagine a smart traffic cop who can change the tolls on the road or the price of electricity in real-time.

The Goal: Instead of letting drivers choose whatever is cheapest for them (which leads to the paradox), the system sets prices so that what is best for the individual is also best for the whole city.
The Solution: The paper proposes mathematical formulas to set these prices. If done right, you can prevent these paradoxes. You can expand a road or a power line, and the system will adjust the prices just enough to keep traffic flowing and costs low.

Summary

This paper is a warning to city planners: You cannot fix roads without thinking about electricity, and you cannot fix electricity without thinking about roads.

If you just add more capacity (more lanes, more wires) without understanding how electric cars will react, you might accidentally create a mess where everyone is slower and paying more. But, if you use smart pricing strategies, you can unlock the true potential of both systems.

1. Problem Statement

The rapid electrification of transportation creates a strong coupling between power and transportation systems. Electric Vehicle (EV) drivers make decentralized routing and charging decisions based on travel time and electricity prices. These decisions determine the spatial distribution of charging loads, which affects power system dispatch and Locational Marginal Prices (LMPs). In turn, LMPs influence driver decisions, creating a feedback loop.

The central problem addressed is whether capacity expansions (adding roads or upgrading power lines) in such coupled systems can paradoxically deteriorate the performance of one or both systems. This phenomenon, known as Braess' Paradox (BP), is well-studied in isolated networks but its manifestation in coupled power-transportation systems remains unexplored. The paper asks:

Can new types of BP arise due to cross-system coupling?
What are the necessary and sufficient conditions for these paradoxes to occur?
How can they be mitigated through pricing mechanisms?

2. Methodology

The authors develop a mathematical framework to model the interaction between the two networks:

System Modeling:
- Transportation Network: Modeled as a graph where EV drivers (a nonatomic population) select routes and charging points to minimize individual costs (travel time + charging cost). This is modeled as a Transportation User Equilibrium (UE).
- Power Network: Modeled as a graph with economic dispatch optimization. The operator minimizes generation costs subject to power flow constraints, taking the induced charging loads as inputs and returning LMPs.
- Coupling: The systems are linked via charging points connected to power buses. The equilibrium of the coupled system is defined as a Generalized User Equilibrium (GUE), where traffic flows and LMPs are mutually consistent (a fixed point).
Mathematical Formulation:
- The GUE is characterized as the optimal solution to a convex quadratic program. This allows for the application of sensitivity analysis and convex optimization theory.
- Social Cost Metrics: The authors define three metrics: Transportation Social Cost ( $\Phi_T$ ), Power System Social Cost ( $\Phi_P$ ), and Coupled System Social Cost ( $\Phi_C$ ).
Paradox Definition:
- A Generalized Braess' Paradox occurs if increasing a capacity parameter (road capacity $\alpha$ or line capacity $\bar{f}$ ) leads to an increase in a social cost metric ( $\partial \Phi / \partial \text{capacity} < 0$ for roads, $>0$ for lines).
- The paper categorizes paradoxes into four types based on the source of expansion and the affected system:
  - T-T: Road expansion $\to$ Increased Transport Cost.
  - T-P: Road expansion $\to$ Increased Power Cost.
  - P-T: Power line expansion $\to$ Increased Transport Cost.
  - P-P: Power line expansion $\to$ Increased Power Cost.

3. Key Contributions

Generalization of Braess' Paradox: The paper formally defines and identifies four distinct types of BP in coupled systems, demonstrating that paradoxes can occur not only within a single network but also propagate across the coupling (e.g., expanding a power line worsening traffic congestion).
Mechanism Identification: Through simple analytical examples (2-Route/2-Bus and 2-Route/3-Bus systems), the authors uncover the mechanisms:
- T-P/P-T Mechanism: Capacity changes alter LMP differentials, incentivizing traffic shifts that increase load imbalance or congestion in the other system.
- T-T/P-P Mechanism: Coupling introduces new externalities where traffic flow relocation, driven by price changes, exacerbates congestion or generation costs even if the direct system would not exhibit BP in isolation.
Necessary and Sufficient Conditions:
- For uncongested power networks, only classical T-T BP is possible; cross-system paradoxes do not occur.
- For radial (tree-structured) power networks, the authors derive analytical conditions for all BP types. They introduce a network reduction technique, aggregating uncongested subnetworks and route bundles to simplify the analysis.
- For general (non-radial) networks, they propose a computational framework using derivative computation to screen for BP occurrences.
Mitigation via Pricing:
- System-Optimal Pricing: The authors prove that adaptive pricing policies can align individual incentives with system optima, eliminating specific types of BP.
- Static Pricing: They formulate a convex optimization problem to find static charging prices that preclude all BP types, provided a feasible solution exists.

4. Key Results

Existence of All BP Types: Numerical studies on a coupled IEEE 9-bus system and a California Bay Area transportation network confirm that all four types of BP (T-T, T-P, P-T, P-P) can occur in realistic settings.
- Example: Expanding a transmission line (P-P) can shift load to expensive generators, increasing total power cost.
- Example: Expanding a road (T-P) can attract drivers to a route with a cheaper charger, causing load imbalance that increases power generation costs.
Sensitivity Analysis: The severity of BP is highly sensitive to parameters such as EV charging demand ( $\rho$ ) and generation cost curvature ( $Q$ ). Higher generation cost curvature generally strengthens the paradox.
Mitigation Effectiveness:
- Adaptive Pricing: Successfully eliminates specific BP types (e.g., $\Pi^*_T$ eliminates T-T, P-T, P-P, P-C).
- Static Pricing: The proposed convex formulation can identify static prices that eliminate all BP types simultaneously in tested scenarios, breaking the feedback loop that causes the paradox.

5. Significance

Infrastructure Planning: The findings challenge the conventional wisdom that "more capacity is always better." Planners must account for cross-system feedback loops; a road expansion might save time but increase electricity costs, or a grid upgrade might worsen traffic congestion.
Operational Strategy: The paper provides a theoretical basis for charging pricing policies as a tool to manage infrastructure inefficiencies. It suggests that dynamic or carefully designed static pricing can prevent paradoxical outcomes without requiring costly physical infrastructure changes.
Theoretical Advancement: By establishing the GUE as a convex optimization problem, the paper bridges game theory, power systems, and transportation engineering, offering rigorous tools (necessary/sufficient conditions) for analyzing complex coupled networks.

In conclusion, this paper demonstrates that the coupling between power and transportation systems introduces new, complex forms of Braess' Paradox. It provides the analytical tools to detect these paradoxes and the pricing strategies to mitigate them, offering critical insights for the strategic planning of electrified infrastructure.