Power exchange cabinet scheduling method, system and storage medium based on dynamic programming model

Abstract

The present invention discloses a method, system and storage medium for scheduling a battery swap cabinet based on a dynamic programming model, including collecting user battery swapping trajectories, battery swap cabinet status data and user battery swapping behavior data, and performing preprocessing; calculating the correlation effect between all battery swap cabinets based on the preprocessed data, combining the correlation effect with the geographical location of the battery swap cabinet, and using a spatial clustering algorithm for clustering analysis to obtain clustering results; constructing a dynamic programming model based on the preprocessed data and the clustering results, and solving the dynamic programming model to obtain a solution for the regional layout of the battery swap cabinet with the maximum benefit; planning the optimal cabinet moving strategy for moving the battery swap cabinet from the original position to the latest position according to the solution for the regional layout of the battery swap cabinet, and executing the operation of moving the battery swap cabinet; when the set time limit is reached, continuing to execute the intelligent scheduling of the battery swap cabinet. Reduce the user's battery swapping time, improve the overall service level, and provide users with a more efficient and convenient battery swapping experience.

Description

Technical Field

[0001] The invention relates to an intelligent scheduling method for a power exchange cabinet based on a dynamic programming model. Background Art

[0002] As an important infrastructure for battery replacement, battery swap cabinets have been widely used in the two-wheeled vehicle battery replacement industry. However, with the surge in the number of people swapping batteries in cities and the gradual expansion of the scale, the existing battery swap cabinet system has many deficiencies in terms of regional layout, battery scheduling and user services, which directly affects the user's battery swap experience and the battery swap efficiency of the user during battery swap, and may also increase the time it takes for the user to swap batteries in the battery swap cabinet. Summary of the invention

[0003] Purpose of the invention: In order to overcome the shortcomings of the prior art, the present invention provides a battery swap cabinet scheduling method, system and storage medium based on a dynamic programming model. The intelligent scheduling system optimizes the regional layout of the battery swap cabinets, reduces the user's battery swap time, improves the overall service level, and provides users with a more efficient and convenient battery swap experience, while ensuring the maximum profit of the battery swap cabinet.

[0004] Technical solution: To achieve the above purpose, a method for scheduling a battery swap cabinet based on a dynamic programming model of the present invention comprises the following steps:

[0005] Step 1: Collect user battery swapping trajectories, battery swapping cabinet status data, and user battery swapping behavior data, and perform preprocessing;

[0006] Step 2: Calculate the correlation effect between all the battery swap cabinets based on the preprocessed data, combine the correlation effect with the geographical location of the battery swap cabinet, and use the spatial clustering algorithm to perform clustering analysis to obtain the clustering results;

[0007] Step 3: Build a dynamic programming model based on the preprocessed data and clustering results, and solve the dynamic programming model to obtain a solution for the regional layout of the battery swap cabinet with the maximum benefit;

[0008] Step 4: Plan the optimal cabinet moving strategy for moving the battery swap cabinet from its original location to its latest location based on the battery swap cabinet regional layout plan, and execute the operation of moving the battery swap cabinet;

[0009] Step 5. When the set time limit is reached after moving the battery swap cabinet, continue with steps 1 to 4.

[0010] Furthermore, the correlation effect between all the battery swap cabinets is calculated, and the correlation between the battery swap cabinet i1 and the battery swap cabinet i2 is defined as E i1，i2 ;

[0011]

[0012] Where Pi1，i2 It represents the number of times the user changes batteries in the battery swap cabinet i2 after changing batteries in the battery swap cabinet i1, ∑ k≠i1 P i1，k The total number of times the user switches batteries in all other battery switching cabinets k after switching batteries in the battery switching cabinet i1;

[0013] The correlation effect E between the battery swap cabinet i1 and all other battery swap cabinets i , the calculation process is as follows:

[0014] E i =∑ i2≠i1 E i1，i2 .

[0015] Furthermore, the geographical location and correlation effect of the battery swap cabinet are combined, and the spatial clustering algorithm k-means clustering analysis is used to obtain the clustering results; all the battery swap cabinets are divided into K clusters according to the size of the city and the number of users, and each cluster is represented by the latitude and longitude of a central point battery swap cabinet;

[0016] The objective function is defined as minimizing the sum of the squares of the longitude and latitude of each battery swap cabinet to the central point battery swap cabinet.

[0017]

[0018] J(C) is the objective function, x i is the longitude and latitude of the location of the battery swap cabinet i, μ h is the central longitude and latitude of the hth cluster among the K clusters, δ i，h is the indicator function.

[0019] Furthermore, the silhouette coefficient S i Measure the correlation of battery swap cabinets and the rationality of battery swap cabinet clustering;

[0020]

[0021] In the formula, a i is the average distance from the battery swap cabinet i to other battery swap cabinets in the same cluster, b i is the average distance from the battery swap cabinet i to the nearest battery swap cabinet in the cluster, E i is the correlation effect between the battery swap cabinet i and other battery swap cabinets in the same cluster, is the correlation effect between the battery swap cabinet i and the battery swap cabinet in the nearest cluster.

[0022] Furthermore, the step three includes the following steps:

[0023] S1. Set the revenue function of the battery swap cabinet based on the user experience score and the cost of the battery swap cabinet;

[0024] S2. Set the state transfer equation based on the revenue function of the battery swap cabinet;

[0025] S3. Initialize the state transfer equation, and gradually update the maximum benefit of each battery swap cabinet under different capacities through the state transfer equation;

[0026] S4. By comparing the benefits of all battery swap cabinets under different capacities, select a layout scheme that maximizes the benefits of all battery swap cabinets;

[0027] S5. Combine the layout plan that maximizes the benefits of all battery swap cabinets with the clustering results to obtain a plan for the regional layout of battery swap cabinets with the maximum benefit.

[0028] Furthermore, the revenue function of the battery swap cabinet is set according to the user experience score and the cost of the battery swap cabinet;

[0029] Maxmize: R(i,j)=U(i,j)-C(i,j)

[0030] In the formula, R(i, j) is the total revenue of the i-th battery swap cabinet, and the capacity of the i-th battery swap cabinet is j; U(i, j) is the user experience score, and the calculation process is as follows:

[0031] U(i, j) = α×convenience of battery replacement + β×reduction in waiting time + γ×user coverage

[0032] In the formula, α, β, and γ are all weight coefficients; C(i, j) represents the cost of the i-th battery swap cabinet, and the calculation process is as follows:

[0033] C(i,j)=fixed cost+δ×unit capacity cost×j

[0034] Where, is the weight coefficient of unit capacity cost, and j is the capacity of the i-th battery swap cabinet.

[0035] Furthermore, the state transfer equation f(i, j) is set based on the revenue function of the battery swap cabinet;

[0036] f(i,j)=max(f(i-1,j),f(i-1,j-ω i )+R(i,j))

[0037] Where f(i, j) represents the maximum profit obtained when the i-th battery swap cabinet has a capacity of j; ω i Arrange the capacity required for the i-th battery swap cabinet.

[0038] Furthermore, a device is provided for the battery swap cabinet scheduling method based on the dynamic programming model, comprising:

[0039] Data collection module: used to collect user battery replacement trajectory, battery replacement cabinet status data and user battery replacement behavior data;

[0040] Data processing module: used for preprocessing and calculation of collected data;

[0041] Data execution module: used to formulate the battery swap cabinet relocation strategy and execute the operation of moving the battery swap cabinet.

[0042] Furthermore, a storage medium stores an executable program, and the executable program is executed by a processor to implement the battery swap cabinet scheduling method based on the dynamic programming model.

[0043] Beneficial effects: The present invention provides a battery swap cabinet scheduling method, system and storage medium based on a dynamic programming model, which collects and analyzes data such as user battery swap trajectories, battery swap cabinet usage, battery demand, etc., and constructs a dynamic programming model through cluster analysis, thereby realizing intelligent scheduling and optimization of the battery swap cabinets; the intelligent scheduling system optimizes the regional layout of the battery swap cabinets, reduces the user's battery swap time, improves the overall service level, and provides users with a more efficient and convenient battery swap experience, while ensuring that the battery swap cabinets have the greatest benefits while providing users with a more efficient and convenient battery swap experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 The figure is a flow chart of the intelligent scheduling method for battery swap cabinets based on the dynamic programming model. DETAILED DESCRIPTION

[0045] The present invention will be further described below in conjunction with the accompanying drawings.

[0046] like Figure 1 As shown, a method, system and storage medium for scheduling a power exchange cabinet based on a dynamic programming model include the following steps:

[0047] Step 1: Collect data such as user battery replacement trajectory, battery replacement cabinet status data and user battery replacement behavior data, and perform pre-processing; the system needs to collect data such as user battery replacement trajectory, battery replacement cabinet usage status data, battery demand, user battery replacement behavior data, etc.; these data can be obtained through the user's mobile device, battery replacement cabinet usage records and other IoT devices; then, clean, classify and archive these data to form a complete data set.

[0048] The system collects the user's battery swapping trajectory, that is, the path information that the user passes when swapping batteries. These path data include the user's driving trajectory, road conditions, traffic information, etc.; through the interface of GPS and traffic data providers, the system can obtain and analyze these path data in real time. Each battery swap cabinet is equipped with an IoT device that can monitor the number of batteries in the cabinet, the charging status of each battery, the temperature, and the service life, that is, the status data of the battery swap cabinet; these data are collected through IoT sensors and stored in local devices for subsequent transmission to the cloud platform for processing. The user's battery swapping behavior data is collected through the user's mobile device, that is, the user's behavior data when using the battery swap cabinet; including the time, frequency, location, and type of battery used when the user chooses to swap batteries; these data can be collected through mobile applications, GPS positioning services, and communication between user devices and battery swap cabinets.

[0049] The collected data is first stored in the local device and then transmitted to the cloud platform through a stable network connection; the network connection is for the IoT device to establish a connection with the cloud platform through 2G / 3G / 4G / NB-IoT; data upload uses efficient communication protocols such as MQTT to ensure the real-time and integrity of the data; the system will package the locally stored data and upload it to the cloud to ensure that the data will not be lost when the network fluctuates.

[0050] Data preprocessing includes data parsing and storage, as well as data cleaning. Data parsing and storage: The system first parses the uploaded data and converts the raw data into a structured format; this includes parsing user behavior data, battery cabinet or battery swap cabinet status information, user battery swap trajectory, or path data, into different dimensional data to prepare for subsequent analysis and processing. Data distribution and storage: Use Kafka to distribute the parsed data to the big data platform center for archiving and storage; each data type, such as user behavior, battery status, and path information, will be stored separately in a dedicated database to ensure that the data is organized in an orderly manner for subsequent query and analysis.

[0051] The data cleaning includes error data processing, missing data processing and abnormal data elimination; error data processing means that the system will automatically detect obvious errors in the data, including data where the order date is later than the battery replacement date. These error data will be marked and automatically corrected or deleted; missing data processing means that the system fills in the missing values ​​in the data through interpolation or other algorithms to ensure the integrity of the data; abnormal data elimination means that the system detects abnormal data in the data through statistical analysis, such as extreme battery usage time, and eliminates or adjusts these abnormal data according to preset rules to avoid misleading the model and analysis results.

[0052] Step 2: Calculate the correlation effect between all battery swap cabinets based on the preprocessed data, combine the correlation effect with the geographical location of the battery swap cabinet, and use spatial clustering algorithm for cluster analysis to obtain the clustering results; cluster analysis involves the calculation of correlation effect and clustering of geographical locations to ensure that the optimization strategy can take into account the geographical distribution and correlation requirements of the cabinets.

[0053] The correlation effect is used to quantify the importance of the battery swap cabinet in the network, reflecting the user's battery swapping behavior between different battery swap cabinets. The correlation effect between all battery swap cabinets is calculated, and the correlation between battery swap cabinet i1 and battery swap cabinet i2 is defined as E i1，i2 ;

[0054]

[0055] Where P i1，i2 It represents the number of times the user changes batteries in the battery swap cabinet i2 after changing batteries in the battery swap cabinet i1, ∑ k≠i1 P i1，k The total number of times the user switches batteries in all other battery switching cabinets k after switching batteries in the battery switching cabinet i1;

[0056] The correlation effect E between the battery swap cabinet i1 and all other battery swap cabinets i , the calculation process is as follows:

[0057] E i =∑ i2≠i1 E i1，i2 .

[0058] The geographical location and correlation effect of the battery swap cabinet are combined, and the spatial clustering algorithm k-means clustering analysis is used to obtain the clustering results; all the battery swap cabinets are divided into K clusters according to the size of the city and the number of users, and each cluster is represented by the longitude and latitude of a central point battery swap cabinet; the K value of different cities is determined according to the size and number of users of different cities; the objective function is defined as minimizing the sum of the squares of the longitude and latitude of each battery swap cabinet to the central point battery swap cabinet,

[0059]

[0060] J(C) is the objective function, x i is the longitude and latitude of the location of the battery swap cabinet i, μ h is the central longitude and latitude of the hth cluster among the K clusters, that is, the longitude and latitude of the central point of the hth cluster, the power exchange cabinet; δ i，h is the indicator function, when x i When it belongs to the hth cluster, δ i，h =1, otherwise δ i，h =0.

[0061] To verify the clustering results, the silhouette coefficient S iMeasure the correlation of battery swap cabinets and the rationality of battery swap cabinet clustering;

[0062]

[0063] In the formula, a i is the average distance from the battery swap cabinet i to other battery swap cabinets in the same cluster, b i is the average distance from the battery swap cabinet i to the nearest battery swap cabinet in the cluster, E i is the correlation effect between the battery swap cabinet i and other battery swap cabinets in the same cluster, is the correlation effect between the battery swap cabinet i and the battery swap cabinet in the nearest cluster.

[0064] Silhouette coefficient S i The value range is between [-1, 1]. The larger the silhouette coefficient is, the better the clustering effect is. The clustering results with results less than 0.3 are merged and the calculation is repeated until all silhouette coefficients are greater than 0.3.

[0065] The clustering results include the number of clusters, that is, the K clusters divided; the cluster center is the average value of the coordinates of all objects in the cluster, that is, the sum of the squares of the longitude and latitude of each battery swap cabinet to the center point battery swap cabinet. The cluster center can help describe the location and characteristics of the cluster; the clustering quality index is the quality of the clustering results, which is usually evaluated by some statistical indicators, that is, the silhouette coefficient measures the correlation of battery swap cabinets and the rationality of battery swap cabinet clustering.

[0066] Step 3. Construct a dynamic programming model based on the preprocessed data and clustering results, and solve the dynamic programming model to obtain the solution for the regional layout of the battery swap cabinets with maximum benefit; based on the preprocessed data and clustering results, the system optimizes the placement of the battery swap cabinets by constructing a dynamic programming model of the 0-1 knapsack problem. The goal of the model is to find a regional layout that can balance cost minimization and user experience maximization, so as to maximize the overall battery swap revenue.

[0067] The step three includes the following steps:

[0068] S1. Set the revenue function of the battery swap cabinet based on the user experience score and the cost of the battery swap cabinet;

[0069] S2. Set the state transfer equation based on the revenue function of the battery swap cabinet;

[0070] S3. Initialize the state transfer equation, and gradually update the maximum benefit of each battery swap cabinet under different capacities through the state transfer equation;

[0071] S4. By comparing the benefits of all battery swap cabinets under different capacities, select a layout scheme that maximizes the benefits of all battery swap cabinets;

[0072] S5. Combine the layout plan that maximizes the benefits of all battery swap cabinets with the clustering results to obtain a plan for the regional layout of battery swap cabinets with the maximum benefit.

[0073] In the dynamic programming model, the maximum benefit refers to the lowest system operating cost and the best user experience while meeting the user's battery replacement needs. The battery swap cabinet benefit function is set based on the user experience score and the cost of the battery swap cabinet.

[0074] Maxmize: R(i,j)=U(i,j)-C(i,j)

[0075] In the formula, R(i,j) is the total revenue of the i-th battery swap cabinet, and the capacity of the i-th battery swap cabinet is j; U(i,j) is the user experience score, which is based on factors such as the user's battery swap trajectory, convenience, and battery swap time. The calculation process is as follows:

[0076] U(i,j) = α×convenience of battery replacement + β×reduction in waiting time + γ×user coverage

[0077] In the formula, α, β, and γ are all weight coefficients, which are set according to specific business needs; convenience of battery swapping: measured by the distance from the user to the nearest battery swapping cabinet, the shorter the distance, the higher the score; reduction in waiting time: refers to the reduction in the user's waiting time when swapping batteries, the shorter the waiting time, the higher the score; user coverage rate: the number of covered users and the size of the coverage area.

[0078] C(i, j) represents the cost of the i-th battery swap cabinet, including equipment maintenance, operating costs, and moving costs. The calculation process is as follows:

[0079] C(i,j)=fixed cost+δ×unit capacity cost×j

[0080] In the formula, δ is the weight coefficient of unit capacity cost, j is the capacity of the i-th battery swap cabinet; fixed costs include equipment purchase and installation costs, etc.; unit capacity cost is the marginal cost of increasing the capacity of the battery swap cabinet, such as maintenance and management costs.

[0081] In the dynamic programming model, the state transfer equation f(i, j) is set based on the battery swap cabinet revenue function, and the state transfer equation is used to gradually solve the maximum revenue after each battery swap cabinet is arranged;

[0082] f(i,j)=max(f(i-1,j),f(i-1,j-ω i )+R(i,j))

[0083] Where f(i, j) represents the maximum profit obtained when the i-th battery swap cabinet has a capacity of j; ω iThe capacity required for arranging the battery swap cabinet for the i-th battery swap cabinet is usually a pre-set value determined based on the demand for the battery swap cabinet.

[0084] Initialize the state transfer equation to f(0, j) = 0. The initialized state transfer equation indicates that no battery swap cabinet is deployed and the benefit is zero. For each battery swap cabinet i, calculate the maximum benefit of j under different capacities. Calculate f(i-1, j) and f(i-1, j-ω i )+R(i, j), compare the sizes of the two, and take the maximum value as the maximum benefit that the i-th battery swap cabinet can obtain; based on this algorithm, gradually calculate the maximum benefit that each battery swap cabinet can obtain, compare the benefits of all battery swap cabinets under different capacities, and select the layout plan that maximizes the benefits of all battery swap cabinets; combine the layout plan that maximizes the benefits of all battery swap cabinets with the clustering results to obtain the battery swap cabinet area layout plan with the maximum benefit.

[0085] Step 4. Plan the optimal cabinet moving strategy for moving the battery swap cabinet from its original location to its latest location based on the regional layout plan of the battery swap cabinet, and execute the operation of moving the battery swap cabinet; when the regional layout is optimized, the system will formulate a specific cabinet moving strategy based on current user needs and battery swap cabinet usage; planning the optimal cabinet moving strategy for moving the battery swap cabinet from its original location to its latest location includes determining the shifting time of the battery swap cabinet, the moving path of the battery swap cabinet, and the latest location of the battery swap cabinet, etc., to ensure that users can complete the battery swap operation in the shortest time.

[0086] The moving time of the battery swap cabinet; based on the analysis of historical data and prediction models, determine which cabinets have lower or higher demand in a specific time period; based on historical data, the system can formulate specific cabinet relocation strategies and dynamically adjust the position of the battery swap cabinet so that users can complete the battery swap operation after planning a reasonable path, optimizing the amount of electricity and user experience. Consider the battery swap behavior data of users in different time periods such as holidays, weekdays, morning and evening peaks, and predict the trend of future changes in battery swap demand; finally, choose the low period of battery swap demand to move the battery swap cabinet to reduce the impact on users. The moving path of the battery swap cabinet; based on the user's battery swap trajectory and path planning, intelligently adjust the regional distribution of the battery swap cabinet to reduce the user's detour time. According to the regional layout plan of the battery swap cabinet, combined with map data and real-time traffic information, plan the optimal path for the battery swap cabinet to move from the original location to the latest location; use geographic information system technology GIS, combined with cluster analysis results and user battery swap behavior data, to determine the specific detailed geographical location of the latest location of all battery swap cabinets.

[0087] Evaluate the moved battery swap cabinet to ensure that the new location of the new cabinet can better meet user needs. By simulating user battery swapping behavior, evaluate the impact of the new area layout on user experience. After the battery swap cabinet is moved, use mobile phone user feedback and cabinet usage data to evaluate the actual effect of the mobile battery swap cabinet strategy.

[0088] Step 5. When the mobile battery swap cabinet reaches the set time limit, continue to execute steps 1 to 4; the number of users in the same area may not always remain unchanged, nor can it be guaranteed that the number of times users use the battery swap cabinets within a period of time, nor can it be guaranteed that the trajectory of users' battery swaps remains unchanged; therefore, set a time limit, such as half a year or a year, to perform intelligent scheduling of the battery swap cabinets. Of course, it is also possible to perform intelligent scheduling of the battery swap cabinets based on the changes in the revenue of the battery swap cabinets, and re-execute steps 1 to 4; when the revenue of the battery swap cabinets in one of the various areas of a city decreases, and the decline is large, it means that the regional layout of the battery swap cabinets at this time can no longer provide users with an efficient and convenient use experience, nor can it guarantee the maximum revenue of the battery swap cabinets, so the location of the battery swap cabinets needs to be rescheduled. Of course, the intelligent scheduling of the battery swap cabinet can be carried out according to the number of times the user uses the battery swap operation in the battery swap cabinet. When the frequency of use of the battery swap cabinet in one area of ​​a city is greatly reduced, or the frequency of use of the battery swap cabinet in another area is greatly increased, most of the battery swap cabinets in the area where the frequency of use of the battery swap cabinet is greatly reduced will be idle, and the battery swap cabinets in the area where the frequency of use of the battery swap cabinet is greatly increased will be in a situation of insufficient use; therefore, it is necessary to re-schedule the battery swap cabinet intelligently. The intelligent scheduling of the battery swap cabinet based on the present invention always maximizes the benefits of the battery swap cabinet, ensures the user's efficient and convenient battery swap experience, and at the same time ensures the maximum benefits of the battery swap cabinet.

[0089] A device, which is used in the battery swap cabinet scheduling method based on the dynamic programming model, comprises:

[0090] Data collection module: used to collect user battery replacement trajectory, battery replacement cabinet status data and user battery replacement behavior data;

[0091] Data processing module: used for preprocessing and calculation of collected data;

[0092] Data execution module: used to formulate the battery swap cabinet relocation strategy and execute the operation of moving the battery swap cabinet.

[0093] A storage medium stores an executable program, and the executable program is executed by a processor to implement the battery swap cabinet scheduling method based on the dynamic programming model.

[0094] Example

[0095] In some special cases, such as in a certain area, such as the office area of ​​​​staff, during weekdays, most of the staff in the office area will order takeout, and the staff will go to work. Therefore, the number of times the battery swap cabinet is used in this area and during weekdays will increase; but during holidays and weekends, the staff in the office area will be on holiday, no one will order takeout, and the staff will not go to work. Therefore, the number of times the battery swap cabinet is used in this area and during holidays will decrease; at this time, if intelligent scheduling of battery swap cabinets is adopted, a large number of battery swap cabinets will need to be moved, which will result in a waste of funds.

[0096] During holidays, more people may order takeout in residential areas, or use two-wheeled electric vehicles, or go to shopping malls. Therefore, some short-time mobile battery swap cabinets can be set up separately. During weekdays, the short-time mobile battery swap cabinets can be moved to office areas; during holidays, the short-time mobile battery swap cabinets can be moved to residential areas or shopping malls. In order to meet the needs of users for battery swap cabinets in a short period of time, there is no need to adopt intelligent scheduling of battery swap cabinets based on dynamic programming models. If intelligent scheduling of battery swap cabinets is adopted, it will cause waste of personnel, time and funds. Instead, setting up a short-time mobile battery swap cabinet only requires moving some of the battery swap cabinets to areas with high demand within a short period of time at night between weekdays and holidays, and there is no need to move most of the battery swap cabinets. This satisfies users' efficient and convenient battery swap experience and ensures the maximum benefit of the battery swap cabinets.

[0097] The above is only a description of the preferred embodiments of the present invention. Ordinary technicians in this technical field can make several modifications and optimizations based on the above disclosure without departing from the above basic principles. These improvements and optimizations should be regarded as the understood protection scope of the present invention.

Claims

1. A battery swap cabinet scheduling method based on a dynamic programming model, characterized in that: The following steps are involved: Step 1: Collect user battery swapping trajectories, battery swapping cabinet status data, and user battery swapping behavior data, and perform preprocessing; Step 2: Calculate the correlation effect between all the battery swap cabinets based on the preprocessed data, combine the correlation effect with the geographical location of the battery swap cabinet, and use the spatial clustering algorithm to perform clustering analysis to obtain the clustering results; The correlation effect between all the battery swap cabinets is calculated, and the correlation between the battery swap cabinet i1 and the battery swap cabinet i2 is defined as E i1，i2 ; Where P i1，i2 It represents the number of times the user changes batteries in the battery swap cabinet i2 after changing batteries in the battery swap cabinet i1, ∑ k≠i1 P i1，k The total number of times the user switches batteries in all other battery switching cabinets k after switching batteries in the battery switching cabinet i1; The correlation effect E between the battery swap cabinet i1 and all other battery swap cabinets i , the calculation process is as follows: Step 3: Build a dynamic programming model based on the preprocessed data and clustering results, and solve the dynamic programming model to obtain a solution for the regional layout of the battery swap cabinet with the maximum benefit; Step 4: Plan the optimal cabinet moving strategy for moving the battery swap cabinet from its original location to its latest location based on the battery swap cabinet regional layout plan, and execute the operation of moving the battery swap cabinet; Step 5. When the set time limit is reached after moving the battery swap cabinet, continue with steps 1 to 4.

2. The method for scheduling a battery swap cabinet based on a dynamic programming model according to claim 1 is characterized in that: The geographical location and correlation effect of the battery swap cabinet are combined, and the spatial clustering algorithm k-means clustering analysis is used to obtain the clustering results; all the battery swap cabinets are divided into K clusters according to the size of the city and the number of users, and each cluster is represented by the latitude and longitude of a central point battery swap cabinet; The objective function is defined as minimizing the sum of the squares of the longitude and latitude of each battery swap cabinet to the central point battery swap cabinet. J(C) is the objective function, x i is the longitude and latitude of the location of the battery swap cabinet i, μ h is the central longitude and latitude of the hth cluster among the K clusters, δ i，h is the indicator function.

3. The method for scheduling a battery swap cabinet based on a dynamic programming model according to claim 2 is characterized in that: Using the silhouette coefficient S i Measure the correlation of battery swap cabinets and the rationality of battery swap cabinet clustering; In the formula, a i is the average distance from the battery swap cabinet i to other battery swap cabinets in the same cluster, b i is the average distance from the battery swap cabinet i to the nearest battery swap cabinet in the cluster, E i is the correlation effect between the battery swap cabinet i and other battery swap cabinets in the same cluster, is the correlation effect between the battery swap cabinet i and the battery swap cabinet in the nearest cluster.

4. The method for scheduling a battery swap cabinet based on a dynamic programming model according to claim 1 is characterized in that: The step three includes the following steps: S1. Set the revenue function of the battery swap cabinet based on the user experience score and the cost of the battery swap cabinet; S2. Set the state transfer equation based on the revenue function of the battery swap cabinet; S3. Initialize the state transfer equation, and gradually update the maximum benefit of each battery swap cabinet under different capacities through the state transfer equation; S4. By comparing the benefits of all battery swap cabinets under different capacities, select a layout scheme that maximizes the benefits of all battery swap cabinets; S5. Combine the layout plan that maximizes the benefits of all battery swap cabinets with the clustering results to obtain a plan for the regional layout of battery swap cabinets with the maximum benefit.

5. The method for scheduling a battery swap cabinet based on a dynamic programming model according to claim 4 is characterized in that: Set the revenue function of the battery swap cabinet based on the user experience score and the cost of the battery swap cabinet; Maxmize: R(i,j)=U(i,j)-C(i,j) In the formula, R(i, j) is the total revenue of the i-th battery swap cabinet, and the capacity of the i-th battery swap cabinet is j; U(i, j) is the user experience score, and the calculation process is as follows: U(i, j) = α×convenience of battery replacement + β×reduction in waiting time + γ×user coverage In the formula, α, β, and γ are all weight coefficients; C(i, j) represents the cost of the i-th battery swap cabinet, and the calculation process is as follows: C(i, j) = fixed cost + δ × unit capacity cost × j In the formula, δ is the weight coefficient of unit capacity cost, and j is the capacity of the i-th battery swap cabinet.

6. The method for scheduling a battery swap cabinet based on a dynamic programming model according to claim 4 is characterized in that: Set the state transfer equation f(i, j) based on the revenue function of the battery swap cabinet; f(i,j)=max(f(i-1,j),f(i-1,i-ω i )+R(i,j)) Where f(i, j) represents the maximum profit obtained when the i-th battery swap cabinet has a capacity of j; ω i Arrange the capacity required for the i-th battery swap cabinet.

7. A device for implementing the battery swap cabinet scheduling method based on a dynamic programming model as described in any one of claims 1 to 6, characterized in that: include: Data collection module: used to collect user battery replacement trajectory, battery replacement cabinet status data and user battery replacement behavior data; Data processing module: used for preprocessing and calculation of collected data; Data execution module: used to formulate the battery swap cabinet relocation strategy and execute the operation of moving the battery swap cabinet.

8. A storage medium, characterized in that: An executable program is stored therein, and the executable program can be executed by a processor to implement the battery swap cabinet scheduling method based on a dynamic programming model as described in any one of claims 1 to 6.

Images