An operation and management platform for automobile charging and battery swapping services

Abstract

The present invention belongs to the technical field of energy service scheduling, and specifically discloses an operation management platform for automobile charging and swapping services, including: a data acquisition module for acquiring service data; a queue estimation module for calculating the expected waiting time in a future time period through a constructed prediction model; a pricing and demand mapping module for constructing a relationship curve between the listing price and the time slot; a car owner sensitivity learning module for constructing the car owner's utility function, and pushing and updating the index that affects the car owner's choice of service mode to the car owner's mobile terminal; a dynamic adjustment module for updating service data according to the service mode reported by the car owner, and judging whether the charging queue threshold is triggered; if so, adjusting the listing price and broadcasting the update; an iterative convergence coordination module for determining whether the updated listing price and expected waiting time are in an equilibrium solution in each time slot; it has the following advantages: realizing dynamic coupling of supply and demand, rapid price adjustment, precise flow control, queue smoothing, and synchronous optimization of revenue and load.

Description

Technical Field

[0001] The present invention relates to the field of energy service scheduling technology, and in particular to an operation management platform for automobile charging and battery swapping services. Background Art

[0002] The global number of electric vehicles continues to rise, creating a massive demand for centralized charging and battery swapping along urban expressways and highways. Large numbers of vehicles arrive simultaneously during peak commuting times and holiday return trips, causing queue lengths to grow rapidly, with the average waiting time per vehicle far exceeding the user's psychological threshold. At the same time, equipment utilization is very low during off-peak hours, significantly extending the investment payback period for charging and battery swapping infrastructure. Existing operations management platforms mostly employ single-layer control and passive queuing logic:

[0003] Stations only publish a single electricity price or service fee curve throughout the day, lacking dynamic guidance for peak loads. Prices fail to reflect the station's instantaneous capacity surplus, leaving drivers' decisions about where to go out of touch with actual supply.

[0004] Queue scheduling is primarily based on first-come, first-served or coarse-grained time-slot reservations. The algorithm lacks real-time predictions of future traffic flows, nor does it employ a game model that balances station revenue with driver experience. Consequently, when queues are excessively long, the system must resort to manual intervention to limit traffic flow.

[0005] All the original data on the station side needs to be uploaded to the cloud for calculation. The return cycle often exceeds one minute, and it is impossible to adjust the price in seconds according to the actual queue status.

[0006] To this end, an operation and management platform for automobile charging and battery swapping services is proposed to solve the above-mentioned problems. Summary of the Invention

[0007] The present invention aims to provide an operation and management platform for automobile charging and swapping services to solve or improve the above-mentioned technical problems in the operation and management of existing charging and swapping stations, such as rigid pricing, extensive queuing scheduling, and excessive cloud computing latency, which make it impossible to achieve dynamic guidance and optimized control based on real-time supply and demand status.

[0008] In view of this, a first aspect of the present invention is to provide an operation and management platform for automobile charging and swapping services.

[0009] The first aspect of the present invention provides an operation and management platform for automobile charging and battery swapping services, including: a data acquisition module for acquiring service data of charging stations at the current moment through a station-side gateway; a queue estimation module for calculating the expected waiting time in the future time period based on the service data by constructing a prediction model, taking into account the sudden load generated by the charging station performing battery swapping and the service time fluctuation affected by the listing price; a pricing and demand mapping module for predicting the number of vehicles receiving charging and battery swapping in a time slot under the condition of considering the marginal cost of unit electric energy, and constructing a relationship curve between the listing price and the time slot; a vehicle owner sensitive a degree learning module, configured to construct a vehicle owner utility function taking into account the relationship curve and the expected waiting time, and to push and update an index influencing the vehicle owner's choice of service mode to the vehicle owner's mobile terminal through the vehicle owner utility function during the time period or the time slot; a dynamic adjustment module, configured to update the service data based on the service mode reported by the vehicle owner, and to determine whether a charging queue threshold is triggered; if so, to adjust the listing price and broadcast the update; and an iterative convergence coordination module, configured to determine whether an equilibrium solution is reached in each time slot based on the updated listing price and the expected waiting time; if not, to update the relationship curve and perform iterative calculations until an equilibrium solution is reached.

[0010] In any of the above technical solutions, the service data includes queue length, service rate, arrival rate and listing price collected synchronously from charging piles, battery swap stations or entrance cameras.

[0011] In any of the above technical solutions, the queue estimation module includes: a vehicle arrival rate filtering unit, which is used to continuously correct the arrival rate of vehicles using recursive Kalman filtering; a service time variance evaluation unit, which counts the service time variance caused by the sudden load generated by battery swapping and the charging time within a sliding window of a preset length; and a variance correction prediction unit, which embeds the service time variance into the prediction model to obtain the expected waiting time of the charging station when performing charging and / or battery swapping.

[0012] In any of the above technical solutions, the prediction model includes the following formula: in, is the time period; The expected waiting time for the next time period; is the system load rate; is the service time variance; The number of parallel service stations; is the average service rate for a single workstation; is the predicted arrival rate.

[0013] In any of the above technical solutions, the pricing and demand mapping module includes: an elastic demand mapping unit, which constructs the relationship between the expected waiting time and the elasticity of the listing price and the number of potential vehicles in different time slots; a balanced optimization unit, which is used to balance the benefits affected by the listing price and the congestion penalty affected by the charging queue, and solve to obtain the optimal price vector.

[0014] In any of the above technical solutions, the balance optimization unit adjusts the solution mode for obtaining the optimal price vector according to the time requirements in the time period and the time slot; the output format of the optimal price vector in the time period and the time slot is the same.

[0015] In any of the above technical solutions, the elastic demand mapping unit includes the following formula: in, is the industry elasticity curve; are sensitivity weights respectively; is the expected number of arriving vehicles; is the current reference price; is the price of the current time slot to be optimized; The expected waiting time for the current time period.

[0016] In any of the above technical solutions, the car owner sensitivity learning module includes: a sensitivity updating unit, which updates the waiting sensitivity and price sensitivity based on the individual historical behavior data of the car owner; an index evaluation unit, which obtains the index through the car owner utility function based on the current waiting sensitivity and price sensitivity in each time period.

[0017] In any of the above technical solutions, the index evaluation unit includes the following formula: in, The index is used to characterize the owner In the time slot Internal choice utility; To wait for sensitivity; is price sensitivity; The price locked by the car owner's mobile terminal in the current time slot or time period.

[0018] In any of the above technical solutions, the length of the time period is related to the number of vehicle owners received by the platform, and / or the length of the time slot is related to the charging time of the vehicle.

[0019] Compared with the prior art, the present invention has the following beneficial effects:

[0020] Through the linkage between the data acquisition module and the queue estimation module, the platform can perceive the changes in the instantaneous queue length, service rate and arrival rate of the station in real time based on the synchronized data of charging piles, battery swap stations and entrance cameras, and then dynamically predict the expected waiting time in the future, so that the pricing and guidance strategies are closely coupled with the actual supply and demand status, avoiding queue out of control or equipment idleness due to information lag.

[0021] Through the pricing and demand mapping module, a curve corresponding to the listing price and time slot is dynamically generated according to the relationship between the marginal cost of unit electricity and the elastic demand of vehicles. Combined with the second-level feedback adjustment of the dynamic adjustment module, the listing price can be accurately regulated by the vehicle owner's arrival behavior. This can effectively curb congestion during peak periods, promote equipment utilization during off-peak periods, and significantly improve overall operating income and site load stability.

[0022] Through the iterative convergence coordination module, the arrival volume after the listing price adjustment and the waiting time forecast are continuously checked in each time slot to see whether they are in balance. If not, the relationship curve is continuously updated and iterative calculations are performed until a dynamic equilibrium is reached, ensuring that the price adjustment and service load converge synchronously. The system stability and prediction accuracy are significantly improved, avoiding the decision lag and frequent manual intervention caused by static pricing in traditional systems.

[0023] Additional aspects and advantages of embodiments according to the present invention will become apparent in the following description or may be learned through practice of embodiments according to the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments with reference to the accompanying drawings, in which:

[0025] Figure 1 This is a logic structure diagram of the platform of the present invention. DETAILED DESCRIPTION

[0026] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that, in the absence of conflict, the embodiments of the present application and the features therein can be combined with each other.

[0027] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present invention is not limited to the specific embodiments disclosed below.

[0028] See also Figure 1 , the following describes an operation and management platform for automobile charging and swapping services according to some embodiments of the present invention.

[0029] The embodiment of the first aspect of the present invention provides an operation and management platform for automobile charging and battery swapping services. In some embodiments of the present invention, such as Figure 1 As shown, the platform includes:

[0030] The data acquisition module is used to obtain the service data of charging performed by the charging station at the current moment through the station-side gateway.

[0031] The queue estimation module calculates the expected waiting time in the future time period based on service data by constructing a prediction model, taking into account the sudden load generated by the charging station performing battery swaps and the service time fluctuations affected by the listing price.

[0032] The pricing and demand mapping module is used to predict the number of vehicles that can be charged and swapped within a time slot, taking into account the marginal cost of unit electricity, and to construct a relationship curve between the listing price and the time slot.

[0033] The car owner sensitivity learning module is used to construct the car owner's utility function by considering the relationship curve and expected waiting time, and push and update the index that affects the car owner's choice of service mode to the car owner's mobile terminal through the car owner's utility function within a time period or time slot.

[0034] The dynamic adjustment module updates service data based on the service methods reported by the car owner and determines whether the charging queue threshold has been triggered; if so, it adjusts the listing price and broadcasts the update.

[0035] The iterative convergence coordination module is used to determine whether the updated listing price and expected waiting time are in an equilibrium solution in each time slot. If not, the relationship curve is updated and iterative calculation is performed until an equilibrium solution is reached.

[0036] The present invention provides an operation and management platform for automobile charging and battery swapping services. The data acquisition module completes high-fidelity collection of multi-dimensional real-time information through a three-layer collaborative mechanism of station-side gateway, edge bus, and secure pipeline. In principle, the platform embeds a Gigabit Ethernet interface inside each national standard DC charging gun, each liquid-cooled battery swapping robot arm, and each box transformer measurement and control device, forming a full-duplex star topology with the station-side gateway; the platform deploys the IEEE1588 precision clock protocol on the physical link to unify all sampling frames to a nanosecond time base, thereby ensuring that high-speed quantities such as instantaneous current peak, contact resistance, and battery core temperature can be spatially aligned. A lightweight eBPF filter program is loaded in real time on the gateway side: first, quartile distance group removal is performed in the ring buffer to remove spike spurious pulses generated by connector jitter at the microsecond scale; then, the Savitzky–Golay convolution kernel is applied to perform local polynomial smoothing on the high-frequency noise, which not only retains key dynamic features such as slope and acceleration, but also reduces the variance in the first-order and second-order differentials, thereby improving the separability of subsequent feature learning from the root. All packets that pass time calibration and filtering are packaged into a five-tuple consisting of "service index - millisecond timestamp - physical channel - engineering unit - raw bytes" and enter a ZeroMQ-based publish-subscribe bus for simultaneous delivery to the edge inference engine and the cloud feature lake. To protect the collection chain from malicious injection, the gateway firmware completes an SM2 two-way handshake with the platform root certificate during startup and appends a 256-bit HMAC-SM3 fingerprint to the end of each message frame. Any tampering will be immediately discarded at the edge.

[0037] Unified time base and noise suppression enable strict sub-millisecond alignment of the electrochemical response curves of the same vehicle in multiple scenarios, such as battery swapping, fast charging, and balancing. This significantly improves the convergence speed and accuracy of SoH estimation, abnormal arc flash identification, and instantaneous power prediction models. Furthermore, the eBPF pipeline is implemented in the gateway kernel, with data capture, processing, and forwarding latency under 150 microseconds. Compared with the traditional "full packet direct push" solution that caches first and then reports, this reduces average edge latency by 92%, providing sufficient iteration budget for the upper layer's 30-second rolling Stackelberg game. Furthermore, the five-tuple packet header enables any subsequent functional module to subscribe, parse, and backtrack by field, eliminating the maintenance cost of multiple parsers. Historical data and real-time streams are absolutely consistent in physical dimensions, facilitating digital twin and replay simulation. Finally, full-link national secret signatures and fingerprint verification ensure isolation between the data and control planes, preventing forged pile-side messages from disrupting service capabilities. This has resulted in a cumulative decrease of 0.47 percentage points in engine offline and false shutdown rates, directly improving annual site availability and shortening the payback period.

[0038] When performing battery swapping services at charging stations, the queue estimation module incorporates a dynamic load-surge mechanism into its prediction model, as the battery swapping process typically triggers a higher user arrival density per unit time than traditional charging. By monitoring the number of battery swapping trigger events and queue trends at the battery swapping stations in real time, the module generates a weighted arrival rate adjustment factor based on the burst load. This accurately maps abnormal increases in user arrivals within a short period of time into the future queue model, thereby predicting the potential service congestion risk associated with peak battery swapping activity before it arrives.

[0039] In terms of service time modeling, the queue estimation module not only considers the basic service capabilities of charging or battery swapping equipment, but also introduces the listing price as an influencing factor. Since changes in listing prices will directly affect users' selection behavior, such as when the price is high, users may tend to short-term charging or fast battery swapping, and when the price is low, they may choose slow charging or deep charging. Therefore, the module dynamically adjusts the service time mean and variance parameters when predicting the service time distribution. Specifically, the price-service time elasticity function trained with historical service data maps real-time listing price changes to the fluctuation range of the time required for each single service, allowing the model to accurately capture subtle changes in service time under different pricing strategies, further improving the accuracy of the overall prediction.

[0040] During the model solution phase, the queue estimation module uses a multi-server queuing system as its theoretical foundation and incorporates a correction term for service time fluctuations. Based on the current arrival rate, the adjusted service time distribution characteristics, and the number of service stations at a station, the module calculates the expected wait time over multiple, smaller time periods in the future, such as every five minutes. The model output includes not only the average wait time forecast for the next time period but also, based on a set window accumulation, produces queue trend curves over longer periods, such as the next 30 minutes or one hour. This provides real-time reference for subsequent pricing guidance and driver decision-making.

[0041] Because the model can sensitively capture the sudden load characteristics caused by battery swaps at charging stations, it can effectively prevent the risk of uncontrolled queues caused by peak battery swaps, significantly improving the platform's responsiveness to extreme load fluctuations. Compared with traditional static prediction methods, the system significantly reduces queue errors during battery swap surges, effectively ensuring service continuity during peak periods.

[0042] Dynamically considering the impact of listing prices on service time allows the forecast results to be adjusted in real time as pricing strategies change, avoiding the problem of price changes being out of sync with the actual recharging rhythm. This fundamentally improves the accuracy and effectiveness of price guidance and ensures that pricing measures are synchronized with queue management.

[0043] Since the platform can continuously calculate the expected waiting time in the future based on real-time service data, it can trigger corresponding price adjustments or guidance strategies in advance before the queue volume exceeds the standard, and actively guide car owners to replenish energy in appropriate time periods, thereby achieving preventive peak control and reducing the reliance on passive emergency measures such as manual flow control or forced station closures.

[0044] Expected waiting time serves as a fundamental input parameter for subsequent price optimization, owner decision-making, and queue feedback modules. Through high-accuracy predictions, it significantly enhances the link consistency and closed-loop stability of the entire operations management platform, providing solid data support for achieving dynamic scheduling within seconds and achieving dual optimization of revenue and experience.

[0045] When predicting the number of vehicles receiving service, the Pricing and Demand Mapping module uses the marginal cost per unit of electricity as a pricing floor, incorporating fundamental supply-side economic constraints into the forecasting model. Using real-time service data, current site electricity purchase prices, and operating cost data, the module accurately calculates the lowest acceptable electricity price per unit time required to charge or swap batteries for each vehicle. This marginal cost (in RMB / kWh) serves as the basis for pricing decisions, ensuring that all subsequent listed prices are free of inversion risks and that the platform maintains positive returns within each time slot.

[0046] In the demand forecasting phase, the Pricing and Demand Mapping module models the changing patterns of vehicle price sensitivity. By collecting historical service data, an empirical elasticity function is established between the listing price and the probability of vehicle arrival. This function can infer the changing trend in the number of vehicles receiving charging or battery swapping services within each time slot based on different listing price levels. Specifically, the module uses an exponential mapping method, using the listing price and the predicted waiting time for the current time slot as independent variables, to calculate the probability of a vehicle selecting that time slot and multiply it by the base number of arrivals.

[0047] Since pricing is based on the marginal cost per unit of electricity, the platform can guarantee positive returns on a single transaction under any load condition, avoiding economic losses caused by arbitrary price cuts, and improving the overall operational stability of the site. By incorporating both waiting time and price changes into the demand forecast function, the module can accurately reflect the user's actual energy replenishment selection logic, so that the listing price not only regulates the number of vehicles arriving at the station, but also simultaneously affects the owner's arrival time and service type selection, truly realizing a deep linkage between price and user behavior. Real-time optimization of the relationship curve between listing price and time slot enables the platform to automatically adapt to changes in supply and demand in different time periods, forming an intelligent dynamic scheduling mechanism that actively reduces peak loads during peak hours and actively promotes sales during off-peak hours, significantly alleviating the problems of queues or equipment vacancy caused by traditional fixed prices.

[0048] The Owner Sensitivity Learning Module dynamically constructs a utility function for each owner, building on the platform's established relationship curves between listing price and time slots and expected waiting time, further incorporating each owner's individual characteristics. This module evaluates the owner's preference for charging or battery swapping services within each time slot by comprehensively considering the listing price, expected waiting time, and the owner's price and time sensitivity. Within each time period or time slot, the module pushes the influence index calculated based on the utility function to the owner's mobile device to assist in making recharging decisions. Furthermore, individual sensitivity parameters are continuously updated and optimized based on the owner's actual choices to continuously improve the accuracy and guidance effectiveness of subsequent push notifications.

[0049] The vehicle owner sensitivity learning module first uses the platform's output of the listing price vs. time slot relationship curve within the current rolling cycle, along with the predicted expected waiting time for the corresponding time slot, as its basic input. For each vehicle owner currently online or about to arrive at a station, the module initializes or updates their price sensitivity and waiting time sensitivity parameters based on their historical charging and swapping history, refueling preferences, current battery level, and remaining range requirements. These sensitivity parameters reflect the vehicle owner's tolerance and responsiveness to price and queue time fluctuations and form the core basis for constructing a personalized utility function.

[0050] In specific applications, the driver sensitivity learning module calculates the comprehensive utility value of each future available time slot for the driver, combining the current listing price and expected wait time. Lower utility values ​​indicate a more attractive time slot and a higher likelihood of selection. To facilitate user understanding and interface presentation, the utility values ​​are standardized and converted into an intuitive and easily comparable impact index. A higher impact index indicates a more favorable time slot for the driver, considering factors such as price, queuing, and recharging convenience. The impact index of several future time slots is pushed to the driver via a mobile app or in-vehicle smart terminal, and displayed in lists, graphs, and recommendations, guiding the driver to independently select the optimal time for charging or battery swapping. This push information includes both price information and estimated wait time, enabling drivers to make more rational decisions tailored to their needs, informed by the cost and time commitment of recharging.

[0051] More importantly, the driver sensitivity learning module doesn't use fixed initial sensitivity parameters. Instead, it dynamically optimizes sensitivity modeling by continuously tracking the driver's actual responses. When the driver selects a time slot for recharging after receiving a push notification, the module adjusts the driver's price sensitivity and waiting time sensitivity based on the deviation between the actual selection and the predicted optimal time slot. For example, if the driver frequently prefers lower prices and tolerates longer waits, the system will automatically lower the price sensitivity weight and increase the waiting time weight. If the driver tends to choose time slots with shorter wait times, even if the price is slightly higher, the time sensitivity will be appropriately adjusted. This dynamic update process uses an incremental learning mechanism, which can gradually converge to a sensitivity model that more realistically and accurately reflects the driver's behavioral habits without increasing the operator's operational burden.

[0052] When each time slot opens, the platform first initializes the listing price setting and waiting time estimate for this time slot based on the owner's response behavior output by the owner sensitivity learning module in the previous control cycle, combined with the listing price and demand forecast curve provided by the pricing and demand mapping module, and the expected waiting time prediction results provided by the queue estimation module.

[0053] Based on this, the iterative convergence coordination module first verifies whether the current listing price and the expected wait time prediction are consistent. Specifically, it determines whether the listing price meets the set control objectives, such as the target number of vehicles, the tolerated queue length, and the maximum allowable wait time, in terms of guiding vehicle owners to select time slots, balancing queue loads, and controlling the number of vehicles arriving at the station. If the system detects a deviation between the actual prediction and the control objectives, it indicates that the current listing price and vehicle owner response have not yet reached an ideal equilibrium.

[0054] At this point, the iterative convergence coordination module dynamically adjusts the listing price based on the direction and magnitude of the deviation, and simultaneously corrects the demand mapping curve. For example, if the system predicts that the number of arriving vehicles exceeds the set target, or the expected waiting time exceeds the tolerance threshold, the module will proactively increase the listing price to suppress some potential demand. Conversely, if the predicted number of vehicles is insufficient or the waiting time is too short, the module will appropriately lower the listing price to attract more drivers to recharge during that time slot, thereby optimizing the station's equipment utilization and revenue level.

[0055] Each time the listing price is adjusted, the pricing and demand mapping module is triggered to regenerate a new listing price-arrival volume curve. The queue estimation module is then called to update the expected wait time based on the latest forecast. The new listing price and wait time are then fed back into the iterative convergence coordination module for a new round of balance checks.

[0056] This process repeats itself in a rolling loop, either at a fixed iteration frequency or within a maximum number of iterations. Each iteration strives to minimize the discrepancy between the listing price guidance effect and the actual demand response. When the changes in listing price and expected wait time remain below a preset convergence threshold for two consecutive iterations, the module deems the system to have reached a stable equilibrium. At this point, the listing price and wait time predictions are locked in as the final service policy for this time slot, which is then referenced by the subsequent owner decision-making module and officially pushed to the owner's app.

[0057] If the system fails to reach convergence within the specified maximum number of iterations, the module will automatically trigger the emergency convergence mechanism and push the suboptimal strategy based on the current optimal close-to-equilibrium listing price and waiting time to ensure that the time slot update cycle is not excessively extended, thereby ensuring the continuity and real-time nature of the overall platform operation rhythm.

[0058] In summary, the data acquisition module collects service data in milliseconds, the queue estimation module predicts expected wait times in seconds, and the dynamic adjustment module instantly adjusts the listed price, forming a high-frequency closed loop. The platform consistently ensures that the number of arriving vehicles matches the station's service capacity, significantly reducing uncontrolled queues. The queue estimation module identifies sudden load fluctuations in battery swapping, and the pricing and demand mapping module uses price leverage to proactively suppress peaks or promote off-peaks, smoothing the load curve. This reduces peak charging wait times, improves equipment utilization during off-peak periods, and stabilizes grid demand. The pricing and demand mapping module uses the marginal cost per unit of electricity as a baseline to ensure positive returns on each transaction. The owner sensitivity learning module delivers personalized impact indices, guiding owners to find a balance between waiting costs and monetary costs, thereby simultaneously improving the owner experience and station profits. The dynamic adjustment module immediately broadcasts a new listed price when it detects that the queue threshold has been reached, and the iterative convergence coordination module confirms that the price-wait combination has reached equilibrium before locking it in. Traditional price adjustment lags of minutes or even hours have been compressed to seconds, significantly reducing the need for peak throttling. The iterative convergence coordination module records each round of convergence, while the driver sensitivity learning module records each behavioral feedback. These two data streams feed into the model training pipeline. The longer the platform operates, the smaller the prediction error and price fluctuations, resulting in a continuous decrease in long-term operating costs.

[0059] In any of the above embodiments, the service data includes queue length, service rate, arrival rate and listing price collected synchronously from charging piles, battery swap stations or entrance cameras.

[0060] In this embodiment, the platform embeds high-precision time synchronization modules in three types of on-site equipment: charging piles, battery swap stations, and entrance cameras. The control panels of the charging piles and battery swap stations output the service rate and listing price in real time. The entrance camera uses a vehicle recognition algorithm to capture the arrival rate and aligns the frame numbers through the same timing pulse. All data items - queue length, service rate, arrival rate, and listing price - are packaged under the same millisecond-level time label, and after being aggregated by the station-side gateway, they are immediately sent to the edge bus to realize a service data flow with "spatially distributed sampling, absolute time synchronization, and unified field encapsulation." In this way, any subsequent module can directly retrieve the complete data based on the unified time axis without the need for secondary alignment or splicing.

[0061] In any of the above embodiments, the queue estimation module includes:

[0062] The vehicle arrival rate filtering unit is used to continuously correct the vehicle arrival rate using a recursive Kalman filter.

[0063] The service time variance evaluation unit calculates the service time variance caused by the sudden load generated by battery replacement and the charging time within a sliding window of a preset length.

[0064] The variance correction prediction unit embeds the service time variance into the prediction model to obtain the expected waiting time of the charging station when performing charging and / or battery replacement.

[0065] In this embodiment, the vehicle arrival rate filtering unit receives the vehicle arrival rate curve from the data acquisition module in real time. Using recursive Kalman filtering, it smooths out spike noise, recognition errors, and camera occlusion jitter in the original curve in real time. It also automatically adjusts the gain coefficient based on the previous prediction, balancing tracking agility and noise suppression stability. This filtering results in a continuous, stable arrival rate estimate that is sensitive to sudden traffic flow, providing reliable input for subsequent queue prediction.

[0066] The service time variance assessment unit calculates the actual service time of charging and battery swapping tasks in real time within a sliding window of a set length. This unit incorporates quickly completed single battery swaps, longer deep DC fast charging, and any intervening abnormal retries into the same time series, thereby fully characterizing the discreteness of service time. It also captures phenomena such as the increased variance during peak battery swap times and the reduced variance caused by slow nighttime charging, all on a single timeline, providing accurate fluctuation quantification indicators for subsequent models.

[0067] The variance-corrected prediction unit directly embeds the service time variance metric obtained in the previous step into the wait time prediction model, allowing the level of fluctuation to have an explicit impact on the queue evolution calculation. When the variance increases, the prediction model automatically adjusts the wait time estimate upward; when the variance decreases, the model adjusts the wait time downward, avoiding excessive conservatism and wasting service capacity. The corrected wait time prediction, along with the arrival rate filtering results, is simultaneously pushed to the pricing and demand mapping module and the vehicle owner sensitivity learning module.

[0068] Specifically, the vehicle arrival rate filtering unit is calculated using the following formula:

[0069]

[0070] Where, are the estimated covariance and observation noise respectively; is the Kalman gain at time t, which controls the fusion ratio of the actual observation value and the predicted value; is the prediction for time t+1 under the condition of the observed information at time t; is the prior prediction at time t, that is, the estimate before the filter update; is the actual observed vehicle arrival rate at time t, which is the raw sensor count.

[0071] Specifically, the service time variance evaluation unit is calculated using the following formula:

[0072]

[0073] Where, The time taken for a single replacement / charging; is the sliding window width; is the mean of the service duration in the last w moments, that is, the average service time.

[0074] In any of the above embodiments, the prediction model includes the following formula:

[0075]

[0076] in, is the time period; The expected waiting time for the next time period; is the system load rate; is the service time variance; The number of parallel service stations; is the factorial of c; is the average service rate for a single workstation; is the predicted arrival rate.

[0077] In any of the above embodiments, the pricing and demand mapping module includes:

[0078] The elastic demand mapping unit constructs the relationship between the expected waiting time and the listing price elasticity and the number of potential vehicles in different time slots.

[0079] The equilibrium optimization unit is used to balance the benefits affected by the listing price with the congestion penalty affected by the charging queue, and to solve and obtain the optimal price vector.

[0080] In this embodiment, the elastic demand mapping unit dynamically constructs an elastic mapping relationship between expected waiting time and listing price and the number of vehicle owners arriving, based on real-time service data collected and feedback from vehicle owners' arrival behavior in the previous cycle. This unit divides the entire day into several time slots based on the changing trends in the number of vehicle arrivals in different time periods, and fits the intensity of vehicle owners' reactions to extended waiting times and their sensitivity to changes in listing prices within each time slot. Through this time-slot modeling, the platform can accurately depict the specific impact of different listing price and waiting time combinations on the number of potential vehicle owners under different operating conditions, thereby forming a detailed, multi-time slot, multi-dimensional potential traffic flow prediction relationship diagram.

[0081] After the elasticity relationship is established, the equilibrium optimization unit takes over the elasticity prediction results and begins to perform a multi-objective trade-off optimization, comprehensively considering the revenue increase from the list price increase and the congestion penalty caused by the lengthening charging queue. First, the theoretical revenue corresponding to the current list price level is calculated for each time slot. The degradation in customer experience and the loss of queue resource utilization caused by the increased vehicle waiting time are simultaneously estimated, and the two are combined into a unified evaluation metric. Subsequently, within the set price upper and lower limits, the equilibrium optimization unit iteratively adjusts the list price for each time slot, following the principle of maximizing revenue and maintaining acceptable queue costs, until a set of optimal price vectors is obtained. Furthermore, to prevent excessive price fluctuations in different time periods from affecting vehicle owners' expectations, a price change smoothing constraint is introduced during the optimization process to ensure that the resulting list price curve remains continuous, stable, and easy to understand over time.

[0082] In any of the above embodiments, the balance optimization unit adjusts the solution mode for obtaining the optimal price vector according to the time efficiency requirements in the time period and time slot; the output format of the optimal price vector in the time period and time slot is the same.

[0083] In this embodiment, the equilibrium optimization unit first determines, based on the current platform operating rhythm, whether it is in a wide-time pricing mode, primarily at the hourly level, or in a time-slot-level, fine-grained pricing mode, primarily at the minute level or even finer. In this time-slot regulation mode, the platform focuses on the medium- to long-term supply and demand balance and overall revenue trends. Therefore, the module adopts an optimization strategy with a lower update frequency and a higher weighting of comprehensive revenue. By solving the listing price trajectory covering the entire time period, it prioritizes smoothing price fluctuations within large cycles, ensuring the stability and consistency of price adjustments and preventing frequent price adjustments from disrupting the decision-making rhythm of car owners.

[0084] In time-slot control mode, the platform must respond quickly with finer temporal resolution, especially in dynamic scenarios such as sudden battery swap peaks and temporary traffic concentrations. In these situations, the balancing optimization unit, based on the rapid fluctuations in real-time arrival rates and expected wait times, employs an optimization model with higher update frequencies and stronger queue penalty weights. This fine-tunes the listing price for each individual time slot, striving to suppress sudden queue backlogs within a time window of seconds to minutes, dynamically distribute vehicle arrival times, and achieve rapid load peak reduction.

[0085] Regardless of whether time-period or time-slot optimization is used, the optimal price vector output by the balanced optimization unit remains consistent: a time-indexed list of prices. Each record corresponds to a specific future service window and includes the corresponding listing price, validity period, and version number. This facilitates the subsequent reception and processing of listing price pushes by the owner sensitivity learning module, dynamic adjustment module, and iterative convergence coordination module, eliminating the need to distinguish between sources. Simply parsing by time index allows for guiding strategy updates and behavioral feedback collection.

[0086] Specifically, taking time slots as an example, a discrete MIP mode is adopted for medium concurrency scenarios, and the model construction includes:

[0087] Leader objective function:

[0088]

[0089] Where, The marginal cost per unit of electricity, i.e., the lowest internal energy supply cost per kilowatt-hour of electricity provided by the platform; is the queue penalty weight coefficient, which measures the negative impact of excessive queues on the overall efficiency; A price fluctuation penalty weight coefficient controls the severity of the penalty for changes in the listing price relative to the benchmark price, avoiding drastic price fluctuations and improving the owner experience; is the square of the queue length.

[0090] Follow the layer KKT constraints:

[0091]

[0092] Where, is the partial derivative of the owner’s utility function with respect to the recharging time, indicating the rate of change of the comprehensive utility felt by the owner when the recharging time changes slightly; The Lagrange multiplier associated with the lower bound of the recharging time is used to detect whether the boundary condition of the earliest recharging time is reached; The recharging time actually selected by the car owner; The earliest recharging time allowed for the car owner is the lower bound; A Lagrange multiplier related to the upper bound of the recharging time is selected to detect whether the boundary condition of the latest recharging time is reached; is the latest recharging time allowed for car owner i, i.e., the upper bound.

[0093] Discretization and solution are: the owner's decision The problem is represented by mutually exclusive 0-1 variables and an SOS1 constraint is added. The master-slave layer is merged into a single mixed complementary integer program (MPEC-MIP). Sparse rearrangement is performed on the CPU before uploading to the GPU. GPU-HiGHS uses a biphase simplex, aiming to obtain a feasible optimal solution within 25 seconds. If a feasible solution is not obtained within 25 seconds, the controller immediately degrades to continuous mean-field mode, retaining 5 seconds to complete the analytical solution and broadcast.

[0094] Specifically, taking time slots as an example, the continuous Mean-Field mode is adopted for high-concurrency scenarios, and the model construction includes:

[0095] Probability density of car owner strategy adoption :

[0096]

[0097] Where, The average waiting time and price sensitivity of the car owner group; is the temperature factor.

[0098] Leadership Function:

[0099]

[0100] Where, is the comprehensive benefit of the platform during the entire planning period, which is a function of the listing price strategy p. The goal is to maximize; It is the total number of potential car owners or the total demand for arrivals in the current period, which is used to normalize the demand density.

[0101] Analytical first-order conditions:

[0102]

[0103] With box constraint correction:

[0104] Where, is the predicted expected waiting time; It is the service capacity per unit time of time slot τ, which is the product of the number of parallel workstations and the rate of a single workstation.

[0105] If you need to meet , entering L-BFGS-B with the analytical value as the initial point, it only takes 5–8 iterations to converge, which takes about 3 seconds.

[0106] Furthermore, the rolling iteration solution is:

[0107]

[0108] Where, represents the set of all time slots, i.e., all time periods or time periods defined in the optimization process; represents the predicted queue length under time slot in the kth iteration; represents the listing price under the time slot in the kth iteration; is the convergence threshold of queue length, indicating the number of vehicles within which the queue prediction change can be considered converged; The convergence threshold of the listing price, indicating the price change within a certain range to be considered converged; Both conditions must be met at the same time to determine the overall convergence.

[0109] In any of the above embodiments, the elastic demand mapping unit includes the following formula:

[0110]

[0111] in, is the industry elasticity curve; are sensitivity weights respectively; is the expected number of arriving vehicles; is the current reference price; is the price of the current time slot to be optimized; The expected waiting time for the current time period.

[0112] In any of the above embodiments, the vehicle owner sensitivity learning module includes:

[0113] The sensitivity update unit updates the waiting sensitivity and price sensitivity based on the individual historical behavior data of the car owner.

[0114] The index evaluation unit obtains the index through the owner's utility function according to the current waiting sensitivity and price sensitivity in each time period.

[0115] In this embodiment, the sensitivity update unit uses each car owner's historical behavioral data as a foundational input to dynamically update their sensitivity to changes in wait times and prices. The platform records each car owner's behavior in real time when faced with different listing prices and expected wait times, extracting key features such as arrival time slot, actual selected time slot, price level, wait time, and actual completed charging and swapping behaviors. After collecting a certain amount of sample data, the sensitivity update unit continuously estimates each car owner's behavioral elasticity to price and wait time changes through incremental learning. If a car owner demonstrates a strong willingness to switch when faced with a price increase, the system automatically increases their price sensitivity. If a car owner prefers to endure a longer wait but insists on recharging, the system automatically adjusts their wait sensitivity. This update mechanism not only adjusts based on a single behavior, but also takes a weighted average of long-term behavioral trajectories to prevent short-term anomalies from causing significant fluctuations in sensitivity parameters, ensuring stable model convergence and gradually adapting to the car owner's true preferences.

[0116] The index evaluation unit calculates a comprehensive index for each time period based on each car owner's current wait sensitivity and price sensitivity, combined with the platform's real-time listing price and expected wait time information, using the car owner's utility function. Using the utility function as the core calculation, the index evaluation unit converts the price and wait costs into a uniform disutility value for the car owner. Furthermore, through index mapping, the comprehensive utility values ​​for different time periods are converted into a standardized impact index. The impact index is presented on the car owner's mobile device as an intuitive chart or score to guide the car owner in selecting the optimal refueling time or service method. The difference in index values ​​for different time periods reflects the overall attractiveness of each time period, helping car owners quickly understand and compare the refueling costs of different options.

[0117] Specifically, the sensitivity update unit is calculated using the following formula:

[0118]

[0119] Where, is the waiting sensitivity for updates, is the price sensitivity to be updated; Waiting memory factor; is the cost memory factor; is the response indication of the owner in the rth record.

[0120] In any of the above embodiments, the index evaluation unit includes the following formula:

[0121]

[0122] in, is an index used to characterize the owner In the time slot Internal choice utility; To wait for sensitivity; is price sensitivity; The price locked by the car owner's mobile terminal in the current time slot or time period.

[0123] In any of the above embodiments, the length of the time period is related to the number of vehicle owners received by the platform, and / or the length of the time slot is related to the charging time of the vehicle.

[0124] In this embodiment, the platform uses longer time periods for operational forecasting and price updates when the number of vehicle owners is low, improving the overall stability and computational efficiency of system scheduling. When the number of vehicle owners is relatively small, such as at night or during off-peak hours, when vehicle arrival frequency is low and changes gradually, the platform can set price and load balancing strategies based on hourly or even longer time periods. The rationale behind this design is that, when traffic flow remains relatively stable, overly frequent price adjustments can easily cause unnecessary fluctuations and disrupt the vehicle owner experience. Smooth regulation over longer time periods is more suitable for low-load scenarios. However, when the number of vehicle owners the platform receives increases rapidly, particularly during peak hours or during sudden battery swap surges, when vehicle arrival frequency increases dramatically and queue status fluctuates rapidly, the platform automatically shortens the time period, adopting a more frequent price and queue control rhythm, dynamically adapting to changes in congestion with refresh cycles ranging from seconds to minutes. This linkage between time period length and vehicle owner number enables the platform to flexibly adjust scheduling granularity based on varying traffic flow sizes, maintaining operational continuity while avoiding unnecessary resource consumption caused by overly fine-grained granularity.

[0125] In the setting of time slots, the platform dynamically determines the length of the time slot in strict accordance with the service time required for the actual charging or battery replacement of the vehicle. When the vehicle is mainly fast-charging or battery replacement, the service time of a single vehicle is generally short, and the platform sets the time slot length relatively compactly, usually between five and ten minutes. In this way, the dynamic changes of vehicle queuing, charging, and leaving the station in each time slot can be captured in a timely manner and fed back to the listing price and queue estimation model to ensure that the service status and price guidance strategy are adjusted synchronously. On the contrary, when the charging mode within the station is mainly slow charging and deep charging, and the charging time of a single vehicle is longer, for example, thirty minutes to one hour, the platform automatically extends the time slot length, so that the charging behavior covered in each time slot has stronger continuity and predictability, avoiding excessive adjustment of prices or service strategies due to short-term fluctuations.

[0126] Furthermore, the length of the time slots and time periods may be different for each vehicle.

[0127] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0128] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this disclosure.

[0129] In the embodiments provided in the present disclosure, it should be understood that the disclosed devices / electronic devices and methods can be implemented in other ways. For example, the device / electronic device embodiments described above are merely schematic. For example, the division of modules or units is merely a logical function division. In actual implementation, there may be other division methods. Multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, indirect coupling or communication connection of devices or units, which may be electrical, mechanical or other forms.

[0130] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present disclosure implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by the processor, it can implement the steps of the above-mentioned various method embodiments. The computer program may include computer program code, which may be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0131] The above embodiments are only used to illustrate the technical solutions of the present disclosure, rather than to limit them. Although the present disclosure has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present disclosure, and should all be included in the scope of protection of the present disclosure.

Claims

1. An operation and management platform for automobile charging and battery swapping services, characterized in that: include: A data acquisition module is used to obtain the service data of charging performed by the charging station at the current moment through the station-side gateway; A queue estimation module calculates the expected waiting time in a future time period based on the service data using a built prediction model, taking into account the sudden load generated by the charging station performing battery swaps and the service time fluctuations affected by the listing price; The pricing and demand mapping module is used to predict the number of vehicles that can be charged and swapped within a time slot, taking into account the marginal cost of unit electricity, and to construct a curve showing the relationship between the listing price and the time slot. an owner sensitivity learning module, configured to construct an owner utility function by considering the relationship curve and the expected waiting time, and push and update an index influencing the owner's choice of service mode to the owner's mobile terminal based on the owner utility function during the time period or the time slot; a dynamic adjustment module that updates the service data based on the service mode reported by the vehicle owner and determines whether a charging queue threshold is triggered; If so, adjust the listed price and broadcast the update; An iterative convergence coordination module is used to determine whether the updated listing price and expected waiting time are in an equilibrium solution in each time slot, and if not, to update the relationship curve and iteratively calculate until an equilibrium solution is reached; The prediction model includes the following formula: ； in, is the time period; The expected waiting time for the next time period; is the system load rate; is the service time variance; The number of parallel service stations; is the average service rate for a single workstation; is the predicted arrival rate; The vehicle owner sensitivity learning module includes: Sensitivity update unit, which updates waiting sensitivity and price sensitivity based on the car owner's individual historical behavior data; An index evaluation unit, which obtains the index through the vehicle owner utility function according to the current waiting sensitivity and price sensitivity in each time period; The index evaluation unit includes the following formula: ； in, The index is used to characterize the owner In the time slot Internal choice utility; To wait for sensitivity; is price sensitivity; The price locked by the car owner's mobile terminal in the current time slot or time period.

2. The operation and management platform for automobile charging and battery swapping services according to claim 1, characterized in that: The service data includes queue length, service rate, arrival rate and listing price collected synchronously from charging piles, battery swap stations or entrance cameras.

3. The operation and management platform for automobile charging and battery swapping services according to claim 2, characterized in that: The queue estimation module includes: A vehicle arrival rate filtering unit, configured to continuously correct the vehicle arrival rate using a recursive Kalman filter; The service time variance evaluation unit calculates the service time variance caused by the sudden load caused by battery swapping and the charging time within a sliding window of a preset length; The variance correction prediction unit embeds the service time variance into the prediction model to obtain the expected waiting time of the charging station when performing charging and / or battery replacement.

4. The operation and management platform for automobile charging and battery swapping services according to claim 1, characterized in that: The pricing and demand mapping module includes: Elastic demand mapping unit, which constructs the relationship between expected waiting time and listing price elasticity and the number of potential vehicles in different time slots; The equilibrium optimization unit is used to balance the benefits affected by the listing price with the congestion penalty affected by the charging queue, and to solve and obtain the optimal price vector.

5. The operation and management platform for automobile charging and battery swapping services according to claim 4, characterized in that: The balance optimization unit adjusts the solution mode for obtaining the optimal price vector according to the time requirements in the time period and the time slot; the output format of the optimal price vector in the time period and the time slot is the same.

6. The operation and management platform for automobile charging and battery swapping services according to claim 4, characterized in that: The elastic demand mapping unit includes the following formula: ； in, is the industry elasticity curve; are sensitivity weights respectively; is the expected number of arriving vehicles; is the current reference price; is the price of the current time slot to be optimized; The expected waiting time for the current time period.

7. The operation and management platform for automobile charging and battery swapping services according to claim 1, characterized in that: The length of the time period is related to the number of vehicle owners received by the platform, and / or the length of the time slot is related to the charging time of the vehicle.

Images