Smart Routing for Electric Vehicles During Power Outages Boosts Grid and Traffic Resilience
In an era where electric vehicles (EVs) are rapidly transforming urban mobility, a new study introduces a groundbreaking solution to one of the most pressing challenges facing smart cities: how to maintain transportation efficiency when power grids fail. As EV adoption surges, the interdependence between electricity infrastructure and transportation networks has become increasingly evident. A sudden outage can paralyze charging stations, strand vehicles, and trigger cascading traffic disruptions. Now, researchers at Jiangsu University have developed an intelligent routing framework that not only enhances grid stability during faults but also guides EVs to functional charging points via optimal paths, minimizing delays and costs.
The research, led by Xiao Wen and Wang Yang from the School of Electrical and Information Engineering at Jiangsu University, presents a dual-layer optimization model that integrates power distribution network reconfiguration with dynamic EV route planning. Published in Computer Development & Applications, the study addresses a critical gap in existing EV navigation systems, which typically assume a fully operational grid. By contrast, this new approach proactively responds to grid disturbances, ensuring that EV drivers are not left stranded when power failures occur.
The innovation lies in its two-phase methodology. The first phase focuses on the power grid itself. When a fault is detected—whether due to equipment failure, weather events, or even cyberattacks—the system initiates a network reconfiguration process. This involves selectively opening and closing switches in the distribution network to reroute power flow, thereby isolating the fault and restoring service to as many customers as possible. The goal is to maximize node voltage stability, a key indicator of grid health. The researchers employed an enhanced harmony search algorithm to solve this complex optimization problem. Unlike traditional methods that may converge too quickly to suboptimal solutions, the improved algorithm incorporates adaptive parameters and early-maturity detection to maintain search diversity and avoid local optima.
This reconfiguration step is crucial because it determines which charging stations remain viable. In the model, only stations with node voltages above 0.96 per unit (p.u.) are considered reliable enough to support charging. This threshold ensures that EVs receive stable, high-quality power, reducing the risk of incomplete charging cycles or damage to onboard electronics. While this may reduce the number of available stations during an outage, it prioritizes service quality and system resilience over sheer availability.
Once the grid has been reconfigured and the operational charging infrastructure is identified, the second phase of the model comes into play: route optimization for individual EVs. Using a mixed-integer constrained programming framework, the system calculates the most efficient path for each vehicle from its current location to its destination, with a mandatory detour to a functional charging station if needed. The objective function balances two key factors: minimizing total travel distance and reducing charging costs. Users or fleet operators can adjust the weighting between these factors based on their priorities—whether they value speed, economy, or a balance of both.
What sets this approach apart is its ability to handle real-world complexities such as charging wait times and congestion at popular stations. The model includes constraints that prevent multiple vehicles from being assigned to the same charger simultaneously without accounting for service duration. It also considers the time required to recharge each vehicle’s battery, which depends on the state of charge upon arrival and the charging rate at the selected station. By incorporating these temporal dynamics, the system avoids unrealistic scenarios where dozens of EVs are routed to a single station with only a few charging points.
To validate their model, the researchers conducted a case study based on a simplified version of the road and power network in Shiyan City, Hubei Province. The test network included 27 intersections and 82 road segments, linked to a 33-node distribution grid with five designated charging stations. Simulations were run under two conditions: normal operation and a simulated fault affecting nodes 22 and 23. In the fault scenario, the reconfiguration algorithm successfully rerouted power, maintaining voltages above the 0.96 p.u. threshold at four of the five charging stations. The station at node 29, however, fell below this threshold and was therefore excluded from service.
Six EVs were then routed through the network under both normal and fault conditions. Each vehicle had a different origin and destination, and all required recharging en route. The results demonstrated the model’s effectiveness in reducing both travel time and energy expenditure. Compared to a benchmark solution using the bacterial foraging optimization (BFO) algorithm—a well-known bio-inspired method for path planning—the new approach achieved significant savings. For example, EV No.1 reduced its charging cost by 48%, while EV No.6 saw a 53% decrease. Across the fleet, half of the vehicles experienced lower costs, with average savings exceeding 50% in several cases.
Moreover, the optimized routes led to shorter charging and waiting times. In the BFO-based solution, some vehicles faced extended delays due to congestion at heavily used stations. In contrast, the new model distributed the load more evenly, directing vehicles to less crowded but still accessible charging points. This load-balancing effect not only improved individual trip efficiency but also enhanced the overall resilience of the coupled power-transport system.
One of the most compelling aspects of this research is its holistic view of urban infrastructure. Rather than treating the power grid and transportation network as separate domains, the authors recognize their deep interconnection. A failure in one can rapidly propagate to the other, creating a feedback loop of disruption. By integrating control strategies across both systems, the proposed model breaks this cycle, enabling faster recovery and reduced societal impact.
From a technical standpoint, the use of the harmony search algorithm represents a thoughtful choice. Originally inspired by the improvisation process of musicians, the algorithm excels at exploring complex solution spaces without getting trapped in local optima. The researchers further refined it with dynamic parameter adjustment and a variance-based early termination mechanism, enhancing its ability to adapt to changing conditions. This level of algorithmic sophistication is essential for solving the combinatorial complexity of network reconfiguration, where even small grids can have thousands of possible switch configurations.
The route optimization component, meanwhile, leverages linear weighted sum methods to combine multiple objectives into a single cost function. This approach allows for flexible trade-offs between distance and cost, reflecting the diverse preferences of real-world drivers. The inclusion of binary variables to represent route choices and continuous variables for charging times transforms the problem into a mixed-integer program, which can be solved efficiently using modern optimization solvers.
Beyond its immediate technical contributions, the study raises important questions about the future of smart mobility. As cities deploy more EVs and expand charging infrastructure, they must also invest in the resilience of their power systems. Static charging networks may suffice under normal conditions, but they are vulnerable to cascading failures. Adaptive, intelligent routing systems like the one described here will be essential for maintaining service during emergencies.
The researchers acknowledge limitations in their current work. For instance, the model assumes perfect information about grid conditions and vehicle states, which may not always be available in practice. It also does not account for driver behavior, such as preferences for specific charging brands or willingness to pay premium rates for faster service. Future work could incorporate real-time data from vehicle telematics, smart meters, and traffic sensors to create a fully dynamic system.
Another area for extension is the integration of renewable energy sources and energy storage. In a future grid with widespread solar and battery systems, reconfiguration strategies could prioritize routes that align with local generation patterns, further reducing strain on the main network. Similarly, vehicle-to-grid (V2G) capabilities could allow EVs to support grid stability during outages, turning them from passive consumers into active participants in system recovery.
The implications of this research extend beyond individual drivers and utility operators. Municipal planners, emergency responders, and policy makers can all benefit from tools that enhance the resilience of urban infrastructure. In the aftermath of a natural disaster or cyber incident, the ability to keep transportation flowing—even with limited power—can be critical for evacuation, supply delivery, and public safety.
Moreover, the economic benefits are substantial. By reducing unnecessary travel and energy consumption, the optimized routing system lowers fuel costs (in terms of electricity), extends vehicle battery life, and reduces wear on roadways. It also minimizes the indirect costs of traffic congestion, such as lost productivity and increased emissions. For fleet operators, these savings can translate into improved profitability and service reliability.
From an environmental perspective, the model supports the broader goals of sustainable urban development. By ensuring that EVs can always find a working charger, it reinforces consumer confidence in electric mobility and encourages further adoption. At the same time, by improving grid efficiency during faults, it reduces the need for fossil-fuel-based backup generation, contributing to lower carbon emissions.
The success of this approach also highlights the importance of interdisciplinary research. Solving complex urban challenges requires expertise in power systems, transportation engineering, computer science, and operations research. The collaboration between these fields is evident in the study’s design, which seamlessly blends electrical network analysis with traffic flow optimization.
As cities continue to evolve into interconnected, data-driven ecosystems, solutions like this will become increasingly vital. The days of siloed infrastructure management are ending. The future belongs to integrated, adaptive systems that can respond intelligently to disruptions, whether they originate in the grid, the road network, or beyond.
In conclusion, the work by Xiao Wen and Wang Yang offers a powerful example of how advanced algorithms can be applied to real-world problems with tangible benefits. By rethinking the relationship between power and transportation, they have created a framework that enhances resilience, efficiency, and user satisfaction. As extreme weather events and cyber threats become more common, such innovations will be essential for building cities that are not only smart but also robust and equitable.
The study demonstrates that with the right tools, we can turn potential crises into opportunities for optimization. Instead of viewing power outages as insurmountable obstacles, we can treat them as dynamic conditions that intelligent systems can navigate. This shift in perspective—from reactive to proactive, from isolated to integrated—is what will define the next generation of urban infrastructure.
For EV owners, the promise is clear: even when the lights go out, their journey doesn’t have to stop. With smart routing guided by real-time grid conditions, they can find their way to a working charger and continue on their path. That peace of mind is not just a convenience—it’s a cornerstone of a truly sustainable and resilient transportation future.
Xiao Wen, Wang Yang, School of Electrical and Information Engineering, Jiangsu University. Published in Computer Development & Applications, DOI: 10.3969/j.issn.1000-386x.2024.07.008