A new simulation-based study focused on electric vehicle charging sharing platforms has concluded that coordinated pricing and service-level decisions can significantly improve operator profitability while balancing station utilization across an urban network. The research examines how a platform integrating multiple charging operators can guide electric vehicle users through dynamic prices, differentiated service levels, and multidimensional recommendations. Using real road-network and traffic-flow data from Beijing’s Fifth Ring Road area, the study builds a bilevel programming model in which operators act at the upper level and electric vehicle users respond at the lower level. The results indicate that optimized platform operations can raise total operator profit and reduce utilization imbalance when compared with uncoordinated pricing, while also improving user utility through a comprehensive recommendation strategy.
The study places electric vehicle charging sharing platforms at the center of urban energy and mobility management. As electric vehicle adoption expands, public charging infrastructure has grown rapidly, but multiple operators often run independent platforms with limited interoperability. This fragmentation makes it difficult for electric vehicle users to identify available charging facilities, while operators face low utilization at some stations and congestion at others. The research argues that a sharing platform can integrate multi-operator resources, improve information transparency, and support more efficient matching between electric vehicle charging demand and station supply. The central question is how such a platform should set prices and service levels while accounting for heterogeneous electric vehicle user preferences.
1. Why Electric Vehicle Charging Sharing Platforms Matter
The study begins from a familiar challenge in electric vehicle infrastructure: charging stations are not uniformly used. Some public charging stations remain underused, while others experience queues and overload. For electric vehicle users, the problem is not simply the number of charging piles but the availability of suitable charging services at the right time and place. For operators, low utilization and uneven demand limit revenue and increase operating pressure. For platforms, the challenge is to coordinate multiple operators and guide electric vehicle drivers toward stations that can serve them efficiently without creating new imbalances.
According to the background presented in the study, China’s electric vehicle and new energy vehicle market has expanded quickly. By the end of 2023, the country’s new energy vehicle fleet reached 20.41 million units, and cumulative charging infrastructure reached 8.596 million units, including 2.726 million public charging piles. The research notes that despite this growth, charging markets still face resource allocation problems. In Beijing, for example, utilization rates in central urban areas have been reported at less than 20 percent in some contexts, while some suburban stations face overload. Such contrasts highlight the need for platform-level coordination rather than independent operator decision-making.

The study also identifies information asymmetry as a critical issue. Electric vehicle users may not know which charging station is available, how long the queue is, what service level is offered, or how price compares across stations. Operators may not fully observe user preferences or the effects of their pricing on network-level utilization. A charging sharing platform can reduce this asymmetry by collecting and distributing information on station location, real-time price, available piles, and service conditions. However, the research stresses that information alone is not enough. The platform also needs operational strategies that shape electric vehicle charging demand in a coordinated way.
In this context, the study proposes a bilevel programming framework. The upper level represents charging operators or a platform coordinating multiple operators. The lower level represents electric vehicle users who choose charging stations based on price, service level, time, and distance. The upper-level model seeks to maximize total operator profit and minimize the variance of station utilization. The lower-level model uses discrete choice theory to characterize electric vehicle user charging preferences. The two levels interact: operators set prices and service levels, electric vehicle users respond with station choices and charging demand, and the resulting demand feeds back into operator objectives.
2. A Bilevel Model for Electric Vehicle Charging Operations
The research designs a bilevel optimization model with clear decision roles. At the upper level, operators decide charging prices and service levels for each station and time period. The upper-level objectives are twofold: maximize total profit and minimize utilization variance. The first objective captures economic performance, while the second captures resource balance. The study argues that focusing only on profit can lead to excessive concentration of electric vehicle charging demand at certain stations, while focusing only on balance can weaken operator incentives. The bilevel structure allows the platform to evaluate trade-offs between these goals.
At the lower level, electric vehicle users choose among charging stations. The utility function includes total time, distance, service level, and price. Total time combines charging time, queue time, and travel time. Distance is calculated using road-network shortest paths. Service level reflects the fast-charging and slow-charging mix at a station. Price is the charging price set by the operator. The study uses a discrete choice model, specifically a mixed logit approach, to quantify user sensitivity to these factors. Survey data from 526 valid respondents and 14,202 choice observations are used to estimate preference parameters. The reported sensitivity coefficients in the main analysis are -0.21 for time, -0.4 for distance, 0.43 for service level, and -0.62 for price. These values indicate that electric vehicle users generally prefer shorter time, shorter distance, higher service levels, and lower prices.
The lower-level model also includes charging demand prediction and queueing. Charging demand depends on the state of charge of electric vehicle batteries. The study models minimum and maximum charging amounts and uses a quadratic demand function to determine the optimal charging amount for each electric vehicle user. A queueing model based on M/M/c is adopted to estimate average waiting time at each charging station. The service rate is linked to the station’s service level, so that stations with a higher share of fast charging can serve electric vehicle users more quickly. Queue time then feeds back into the user utility function, influencing station choice probabilities and total charging demand.
| Model Component | Role in the Electric Vehicle Charging Platform | Key Variables or Parameters |
|---|---|---|
| Upper-level operator model | Coordinates charging prices and service levels to maximize total profit and reduce utilization variance | Station price, service level, operating cost, service cost, utilization, charging demand |
| Lower-level electric vehicle user model | Represents station choice behavior and charging demand based on user preferences | Total time, distance, service level, price, state of charge, battery capacity |
| Discrete choice model | Quantifies electric vehicle user sensitivity and choice probabilities | Mixed logit estimation, survey data, preference coefficients |
| Queueing model | Estimates waiting time and station congestion for electric vehicle charging | Arrival rate, service rate, number of chargers, service level |
| Multi-objective algorithm | Generates trade-off solutions between profit and utilization balance | NSGA-II, Pareto frontier, crowding distance, non-dominated sorting |
| Decision selection method | Selects a preferred strategy from the Pareto set under time-of-use conditions | TOPSIS, dynamic weights, peak and trough periods |
3. Simulation Design Using Real Road Network and Electric Vehicle Traffic Flow
The simulation is built on MATLAB 2022b and uses real road-network data from within Beijing’s Fifth Ring Road. The study extracts road attributes and node information and assigns unique identifiers to roads. The full road network contains 51,434 nodes and 68,206 road segments. To improve computational efficiency and focus on a manageable urban area, the analysis limits the road network to a region between longitude 116.345 degrees east and 116.367 degrees east and between latitude 40.000 degrees north and 40.030 degrees north. This area is approximately 8 to 10 square kilometers and lies in the Haidian District of Beijing. Within this area, three charging stations are fixed at specified locations, and 50 electric vehicles are randomly placed on road nodes.
The simulation uses time-varying traffic conditions. Electric vehicle travel speed is derived from road-segment speed data, which reflects average speed and congestion. The study sets the speed range from 10 kilometers per hour to 70 kilometers per hour. The upper bound follows urban main-road speed limits, while the lower bound reflects severe congestion conditions reported for Beijing. Charging power is set at 10 kilowatts. Each charging station has an initial charging capacity of 500 kilowatt-hours. Fixed operating cost is 10 yuan per hour per charging pile, and the service cost coefficient is 50. The price range for charging stations is set from 0 to 3 yuan per kilowatt-hour. The three stations are assigned different service-level ranges: station 1 from 0.01 to 0.3, station 2 from 0.3 to 0.6, and station 3 from 0.6 to 1.0. These ranges position station 1 as a low-service station with a high share of slow charging, station 2 as a medium-service station with mixed fast and slow charging, and station 3 as a high-service station with a high share of fast charging.
The time-of-use cost electricity price follows Beijing’s peak, flat, and valley structure. The peak price is 0.9864 yuan per kilowatt-hour, the flat price is 0.6770 yuan per kilowatt-hour, and the valley price is 0.3766 yuan per kilowatt-hour. The peak periods include 10:00 to 15:00 and 18:00 to 21:00. The flat periods include 7:00 to 9:00, 15:00 to 18:00, and 21:00 to 23:00. The valley period runs from 23:00 to 7:00. These time-of-use prices shape operator costs and influence the optimized charging prices for electric vehicle users.
| Simulation Parameter | Value Used in the Study | Relevance for Electric Vehicle Charging |
|---|---|---|
| Road network size | 51,434 nodes and 68,206 road segments | Provides realistic routing for electric vehicle users |
| Focused study area | Longitude 116.345E to 116.367E, latitude 40.000N to 40.030N | Defines the urban electric vehicle charging service region |
| Electric vehicle fleet | 50 vehicles in the base scenario | Represents charging demand in the simulation |
| Charging stations | 3 fixed stations with different service positioning | Supports differentiated electric vehicle charging services |
| Charging price range | 0 to 3 yuan per kilowatt-hour | Bounds operator pricing decisions |
| Service-level ranges | 0.01 to 0.3, 0.3 to 0.6, and 0.6 to 1.0 | Captures low, medium, and high electric vehicle charging service levels |
| Charging capacity | 500 kilowatt-hours per station | Defines station utilization and capacity pressure |
| Charging power | 10 kilowatts | Used to calculate electric vehicle charging time |
| Vehicle speed range | 10 to 70 kilometers per hour | Reflects urban traffic conditions for electric vehicle travel |
| Time-of-use cost electricity | 0.9864, 0.6770, and 0.3766 yuan per kilowatt-hour | Drives cost differences across peak, flat, and valley periods |
The multi-objective algorithm uses NSGA-II with a generation count of 100, population size of 200, crossover probability of 0.7, and mutation probability of 0.5. The algorithm generates a Pareto frontier representing trade-offs between total operator profit and utilization variance. To select a preferred solution from the Pareto set, the study applies TOPSIS with dynamic weights. During peak electricity price periods, utilization variance receives a weight of 0.6 and total profit receives 0.4, prioritizing load balance. During valley periods, total profit receives 0.6 and utilization variance receives 0.4, prioritizing economic performance. During flat periods, the weights are 0.5 and 0.5. This dynamic weighting allows the electric vehicle charging sharing platform to adapt to changing system conditions across the day.
4. Differentiated Electric Vehicle Charging Strategies Emerge from Optimization
The optimized results show that the three charging stations adopt differentiated operating strategies. Station 1 tends toward low price and low service level, with prices between 1.0 and 1.5 yuan per kilowatt-hour and service levels between 0.08 and 0.27. Station 2 occupies a middle position, with prices between 1.5 and 2.0 yuan per kilowatt-hour and service levels between 0.38 and 0.62. Station 3 follows a high-price and high-service strategy, with prices between 2.0 and 3.0 yuan per kilowatt-hour and service levels between 0.62 and 0.95. The results indicate that the platform does not push all stations toward the same strategy. Instead, it allows each station to serve different segments of electric vehicle users.
Over a 24-hour cycle, the average prices of stations 1, 2, and 3 are 1.28, 1.76, and 2.46 yuan per kilowatt-hour, respectively. Their standard deviations are 0.09, 0.23, and 0.21, suggesting relatively stable pricing under the optimization strategy. Utilization also differs by station. Station 1 has utilization ranging from 0.2044 to 0.5034, with a mean of 0.3449. Station 2 has utilization from 0.1596 to 0.5148, with a mean of 0.2776. Station 3 has utilization from 0.0768 to 0.3345, with a mean of 0.2219. The lower-price station attracts more electric vehicle charging demand and therefore records higher utilization, while the higher-price station relies on higher unit margins and higher service levels to maintain profitability.
| Charging Station | Price Range in the Optimized Strategy | Service-Level Range | Average Price | Average Utilization | Strategic Position for Electric Vehicle Users |
|---|---|---|---|---|---|
| Station 1 | 1.0 to 1.5 yuan per kilowatt-hour | 0.08 to 0.27 | 1.28 yuan per kilowatt-hour | 0.3449 | Low-price, low-service electric vehicle charging option |
| Station 2 | 1.5 to 2.0 yuan per kilowatt-hour | 0.38 to 0.62 | 1.76 yuan per kilowatt-hour | 0.2776 | Balanced electric vehicle charging option |
| Station 3 | 2.0 to 3.0 yuan per kilowatt-hour | 0.62 to 0.95 | 2.46 yuan per kilowatt-hour | 0.2219 | High-price, high-service electric vehicle charging option |
The study interprets this outcome as evidence that electric vehicle users are heterogeneous. Some electric vehicle users prioritize low charging prices and are willing to accept lower service levels. Others value faster charging and better service and are willing to pay more. A sharing platform that coordinates multiple stations can accommodate both groups without forcing every station into a single operating model. This differentiation also helps balance utilization because demand is distributed across stations with different price and service profiles rather than concentrating at a single low-price station.
5. Optimized Pricing Outperforms Uncoordinated Pricing for Electric Vehicle Charging
To evaluate the value of coordination, the study compares optimized pricing with an uncoordinated pricing benchmark. In the uncoordinated case, each operator sets prices based only on electricity cost and its own operating cost, without considering user choice behavior or other operators’ strategies. The price ranges in the uncoordinated case are set as 1 to 2, 1.5 to 2.5, and 2 to 3 yuan per kilowatt-hour for the three stations, while service-level ranges remain the same as in the optimized case. The comparison is run over 24 hours with 100 simulations to assess the distribution of total profit and utilization variance.
The results show a substantial difference. Under optimized pricing, the median total profit is 243.75 yuan, and the maximum reaches 485.84 yuan. Under uncoordinated pricing, the median total profit is 131.47 yuan, and the maximum is 364.63 yuan. The median total profit improvement is 85.4 percent. For utilization variance, the optimized case has a median of 0.017, while the uncoordinated case has 0.0233, a reduction of 27.0 percent. The optimized case also has a more concentrated distribution and fewer outliers, indicating greater stability. Over the full 24-hour period, the optimized pricing strategy increases total profit by 59.1 percent and reduces average utilization difference by 60.9 percent.
| Performance Indicator | Optimized Electric Vehicle Charging Pricing | Uncoordinated Pricing | Reported Change |
|---|---|---|---|
| Median total profit | 243.75 yuan | 131.47 yuan | 85.4 percent higher median |
| Maximum total profit in simulation distribution | 485.84 yuan | 364.63 yuan | Higher upper range under optimization |
| Median utilization variance | 0.017 | 0.0233 | 27.0 percent lower variance |
| 24-hour total profit change | Optimized platform outcome | Uncoordinated outcome | 59.1 percent increase |
| Average utilization difference | Optimized platform outcome | Uncoordinated outcome | 60.9 percent reduction |
The study further examines station-level utilization and profit. Under uncoordinated pricing, utilization within a day fluctuates widely, from as low as 0.05 to as high as 0.90, indicating significant resource imbalance. Under optimized pricing, all stations show more stable utilization distributions, and station 2 utilization improves notably. Station 1 and station 3 also move toward more balanced utilization, avoiding the overload and idleness observed in the uncoordinated case. Profit results show similar patterns. In the uncoordinated case, station 3 profit ranges from -37.13 yuan to 207.65 yuan within a day, showing high instability. Under optimized pricing, all stations display more consistent positive profits.
Total charging volume also becomes more stable under optimization. In the uncoordinated case, total charging volume peaks at 578.62 kilowatt-hours and falls to 288.68 kilowatt-hours, with demand concentrated in high-price periods such as 9:00 to 11:00 and 18:00 to 22:00. Under optimized pricing, total charging volume peaks at 546.01 kilowatt-hours and falls to 291.52 kilowatt-hours. The standard deviation of 24-hour total charging volume is approximately 74.40 in the uncoordinated case and 72.64 in the optimized case. The study concludes that optimized pricing is better suited to long-term electric vehicle charging station operations because it smooths demand and reduces extreme fluctuations.
6. Cost Electricity and Electric Vehicle Fleet Sensitivity
The study analyzes how changes in cost electricity prices affect electric vehicle charging operators. Three scenarios are considered: the existing cost electricity level, a 15 percent reduction, and a 30 percent reduction. Under the existing price, total profit for the three charging stations is 6,032.71 yuan. When cost electricity falls by 15 percent, total profit rises to 6,484.93 yuan. When cost electricity falls by 30 percent, total profit rises further to 6,973.73 yuan. The 30 percent reduction scenario produces a profit increase of approximately 15.6 percent relative to the base case.
The study explains that profit growth comes from two channels. First, lower cost electricity expands the margin between charging price and cost, improving profitability per kilowatt-hour of electric vehicle charging. Second, lower prices stimulate electric vehicle user demand, increasing total charging volume. When cost electricity is 15 percent lower, average total charging volume rises from 394.7229 kilowatt-hours to 410.2083 kilowatt-hours, and the average price spread rises from 1.1507 to 1.4229. When cost electricity is 30 percent lower, average total charging volume rises to 428.6623 kilowatt-hours, and the average price spread rises to 1.7494. The results highlight the sensitivity of electric vehicle charging demand to price and the importance of cost-side changes for platform profitability.
| Cost Electricity Scenario | Total Profit for the Three Stations | Average Total Charging Volume | Average Price Spread |
|---|---|---|---|
| Base cost electricity | 6,032.71 yuan | 394.7229 kilowatt-hours | 1.1507 |
| 15 percent reduction | 6,484.93 yuan | 410.2083 kilowatt-hours | 1.4229 |
| 30 percent reduction | 6,973.73 yuan | 428.6623 kilowatt-hours | 1.7494 |
The study also tests changes in electric vehicle fleet size. Four scenarios are simulated with 25, 50, 75, and 100 electric vehicles, each run 100 times. Using data for 9:00 as an example, the median total profit rises from 341.13 yuan with 25 electric vehicles to 1,347.40 yuan with 100 electric vehicles. The relative growth rates are 43.14 percent from 25 to 50 electric vehicles, 99.59 percent from 50 to 75 electric vehicles, and 38.21 percent from 75 to 100 electric vehicles. The slowing growth at the highest fleet size suggests that station capacity or market saturation begins to constrain profit growth. The study advises operators to consider expanding charging facilities or adding stations when electric vehicle demand approaches capacity limits.
Utilization variance follows a different pattern. At 9:00, the median utilization variance is 0.047 with 25 electric vehicles, 0.0724 with 50 electric vehicles, 0.0436 with 75 electric vehicles, and 0.0346 with 100 electric vehicles. The relative changes are 54.04 percent from 25 to 50 electric vehicles, -39.78 percent from 50 to 75 electric vehicles, and -20.64 percent from 75 to 100 electric vehicles. The study interprets this as an initial increase in imbalance followed by greater stability as the electric vehicle fleet grows and station capacity constraints shape demand distribution. For platform operators, smaller electric vehicle fleets require careful allocation strategies to avoid localized overload, while larger fleets require capacity expansion and network planning.
7. Comprehensive Recommendations Improve Electric Vehicle User Utility
Beyond operator pricing, the study evaluates user guidance. The platform can recommend charging stations to electric vehicle users based on price, service level, time, and distance. The comprehensive recommendation strategy is compared with a time-only recommendation method in which electric vehicle users choose stations solely on time. The simulation covers 50 electric vehicles over 24 hours and calculates total time, total distance, total price, total service level, and total user utility.
Under the time-only method, total user utility is -1118.11. Total time is 1171.49 minutes, total distance is 167.32 kilometers, total price is 1419.20 yuan, and total service level is 173.80. Under the comprehensive recommendation method, total user utility is -1042.58. Total time is 1421.75 minutes, total distance is 147.92 kilometers, total price is 1218.40 yuan, and total service level is 164.10. The comprehensive method reduces total distance by 11.62 percent and total price by 14.14 percent, while increasing total user utility by 6.75 percent. Total time increases, indicating that the method does not optimize a single factor but balances multiple dimensions. Service level is slightly lower than in the time-only method, but the difference is small, suggesting that the comprehensive strategy maintains service stability while improving overall electric vehicle user utility.
| Metric | Time-Only Recommendation | Comprehensive Recommendation | Reported Change |
|---|---|---|---|
| Total user utility | -1118.11 | -1042.58 | 6.75 percent improvement |
| Total time | 1171.49 minutes | 1421.75 minutes | Higher total time under comprehensive method |
| Total distance | 167.32 kilometers | 147.92 kilometers | 11.62 percent reduction |
| Total price | 1419.20 yuan | 1218.40 yuan | 14.14 percent reduction |
| Total service level | 173.80 | 164.10 | Slightly lower but relatively stable |
The study argues that the comprehensive recommendation method is more suitable for electric vehicle charging sharing platforms because it reflects the multidimensional nature of electric vehicle user decisions. Electric vehicle drivers do not care only about time. They also consider distance, price, and service level. A platform that ignores these dimensions may recommend stations that are fast to reach but expensive or poorly serviced. By incorporating estimated user sensitivities, the platform can improve satisfaction and encourage more efficient station utilization.
8. User Preference Heterogeneity Shapes Electric Vehicle Charging Outcomes
The study further analyzes how changes in electric vehicle user sensitivity affect operator decisions. Two scenarios are constructed using income groups. The low-income group is treated as more price-sensitive, while the high-income group is treated as more service-sensitive. In the price-sensitive scenario, the average price of the three charging stations falls to 1.8 yuan per kilowatt-hour from 2.2 yuan per kilowatt-hour, a relative decrease of 18.2 percent. Electric vehicle users shift toward the low-price station. Station 1’s utilization share rises from 38.5 percent to 52.3 percent, while station 3’s share falls from 32.4 percent to 19.8 percent. Station 1 accounts for 47.0 percent of profit, while station 3 profit falls by 28.4 percent. Total profit for the three stations declines by 19.1 percent, and utilization variance increases by 78.0 percent, indicating greater imbalance.
In the service-sensitive scenario, the average service level rises from 0.5 to 0.65, an increase of 30.0 percent. Electric vehicle users shift toward the high-service station. Station 3 service level rises from 0.7 to 0.85, and its price rises from 2.2 to 2.5 yuan per kilowatt-hour. Station 1 service level falls from 0.4 to 0.35, and its price falls from 1.8 to 1.4 yuan per kilowatt-hour. Total profit increases by 8.5 percent, and utilization variance decreases by 21.3 percent, indicating more balanced resource allocation. The study concludes that user preference heterogeneity has a direct effect on electric vehicle charging station pricing, utilization, and profitability. Platforms should adjust recommendation and pricing strategies as user sensitivity changes.
| User Sensitivity Scenario | Key Price or Service Change | Utilization Shift | Profit and Balance Outcome |
|---|---|---|---|
| Price-sensitive electric vehicle users | Average price falls from 2.2 to 1.8 yuan per kilowatt-hour | Station 1 share rises from 38.5 to 52.3 percent; station 3 falls from 32.4 to 19.8 percent | Total profit falls 19.1 percent; utilization variance rises 78.0 percent |
| Service-sensitive electric vehicle users | Average service level rises from 0.5 to 0.65 | Demand shifts toward high-service station 3 | Total profit rises 8.5 percent; utilization variance falls 21.3 percent |
9. Dynamic TOPSIS Weighting Supports Peak and Valley Electric Vehicle Charging Operations
The study also assesses the role of TOPSIS in selecting strategies from the Pareto frontier. During peak periods from 10:00 to 13:00, TOPSIS assigns a higher weight to utilization variance, aiming to reduce differences among charging stations. In this period, the total profit with TOPSIS is 3014.4 yuan, compared with 3133.47 yuan without TOPSIS. The average utilization variance falls from 0.115 to 0.086. The platform sacrifices some profit to achieve better load balance during peak electric vehicle charging demand.
During valley periods from 0:00 to 3:00, TOPSIS assigns a higher weight to total profit. The total profit with TOPSIS is 3137.66 yuan, compared with 3016.29 yuan without TOPSIS. The average utilization variance rises from 0.029 to 0.050. In this period, the platform prioritizes economic performance while accepting a somewhat higher variance. The study argues that dynamic weighting allows the electric vehicle charging sharing platform to coordinate profit and balance according to time-of-use conditions, improving practical adaptability.
| Period | TOPSIS Weighting Focus | Total Profit With TOPSIS | Total Profit Without TOPSIS | Utilization Variance With TOPSIS | Utilization Variance Without TOPSIS |
|---|---|---|---|---|---|
| Peak period, 10:00 to 13:00 | Higher weight on utilization variance | 3014.4 yuan | 3133.47 yuan | 0.086 | 0.115 |
| Valley period, 0:00 to 3:00 | Higher weight on total profit | 3137.66 yuan | 3016.29 yuan | 0.050 | 0.029 |
10. Implications for Electric Vehicle Charging Platform Operators and Policymakers
The study draws several operational implications. First, electric vehicle charging operators should not rely on a single pricing model across all stations. Differentiated strategies can serve different electric vehicle user groups. Low-price and low-service stations can attract price-sensitive electric vehicle users, while high-price and high-service stations can serve users who value speed and convenience. Medium-price and medium-service stations can provide balanced options. A platform that coordinates these strategies can improve total profit and reduce utilization imbalance at the same time.
Second, platforms should monitor utilization and user flow in real time. The study recommends dynamic adjustment of prices and service levels based on station-level demand and network-level balance. Stations with low utilization may need targeted promotions or service improvements, while stations with high utilization may need capacity expansion or demand diversion. Real-time monitoring can help avoid the coexistence of idle charging piles and long electric vehicle queues.
Third, when cost electricity falls due to renewable energy expansion or market reform, operators can pass part of the cost reduction to electric vehicle users through lower charging prices. The study shows that lower cost electricity increases profit through both higher margins and higher charging volume. However, the platform should also maintain service quality and use time-of-use pricing to guide electric vehicle charging demand toward valley periods. This can further balance station utilization and reduce pressure during peak periods.
Fourth, user segmentation matters. When electric vehicle users become more price-sensitive, low-price stations may face concentrated demand, while high-service stations may lose utilization. In such cases, the platform should strengthen service levels at low-price stations and consider profit-sharing mechanisms to balance station-level outcomes. When electric vehicle users become more service-sensitive, high-service stations gain demand and can support higher prices. The platform should then optimize pricing to preserve profitability and ensure that service quality remains consistent across the network.
The study also notes limitations and future research directions. The current model considers cost electricity and user demand but does not fully incorporate policy subsidies or changes in the competitive landscape. It also does not fully capture regional heterogeneity in electric vehicle charging behavior. Future work could introduce policy factors, regional differences, and more detailed demand forecasting. Data mining and machine learning methods, such as clustering and time-series models, could be used to predict electric vehicle user behavior more accurately and support dynamic pricing and resource allocation.
11. Conclusion: Coordinated Electric Vehicle Charging Strategies Deliver Dual Benefits
The research concludes that a bilevel programming approach can support electric vehicle charging sharing platforms in balancing operator profit and resource utilization. By treating operators as upper-level decision-makers and electric vehicle users as lower-level responders, the model captures the interaction between pricing, service levels, station choice, and charging demand. The NSGA-II algorithm generates Pareto-optimal strategies, and TOPSIS selects preferred solutions under different time-of-use conditions. The simulation using real road-network and traffic-flow data from Beijing shows that optimized pricing and service decisions outperform uncoordinated pricing.
The main findings are consistent across the study. Differentiated station strategies emerge, with low-price low-service, medium-price medium-service, and high-price high-service electric vehicle charging options. Optimized pricing increases total operator profit and reduces utilization variance compared with uncoordinated pricing. Comprehensive recommendations improve electric vehicle user utility by considering price, service level, time, and distance rather than time alone. Sensitivity analysis shows that cost electricity reductions and electric vehicle fleet growth have significant effects on profitability and utilization balance. User preference heterogeneity further shapes station-level outcomes, underscoring the need for dynamic platform strategies.
For electric vehicle users, the platform can reduce search costs and improve matching with suitable charging stations. For operators, coordination can increase revenue and stabilize utilization. For urban energy systems, better electric vehicle charging management can support load balance and infrastructure efficiency. The study suggests that electric vehicle charging sharing platforms should combine dynamic pricing, service differentiation, real-time monitoring, and behavior-aware recommendations to achieve sustainable operations. As electric vehicle adoption continues to rise, such platform-level coordination is likely to become increasingly important for efficient charging infrastructure use and for the broader transition toward low-carbon urban mobility.
