A robust optimization scheduling method and system for idle mobile charging stations based on different scenarios

Abstract

The present invention relates to a method and system for robust optimization scheduling of idle mobile charging stations that distinguish between scenarios, and belongs to the technical field of electric vehicle Internet of Things. The method comprises: constructing a scheduling model for idle mobile charging stations; constructing a fuzzy set based on historical charging demand data and weather data, classifying the historical charging demand data into different scenarios, optimizing the scheduling model based on the fuzzy set and the status information of the mobile charging station, and obtaining the objective function of the robust optimization scheduling model; solving the objective function of the robust optimization scheduling model, and scheduling the idle mobile charging stations according to the solution results. This method makes the scheduling plan of the charging station more reasonable, can meet the charging needs of more electric vehicles, reduce operating costs, and improve the scheduling efficiency of the charging station.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric vehicle Internet of Things, and in particular relates to a scenario-differentiated idle mobile charging station robust optimization scheduling method and system. Background Art

[0002] In the electric vehicle Internet of Things, the common charging method for electric vehicles is to charge at fixed charging stations. However, the charging requirements of electric vehicles during peak charging periods at fixed charging stations and in areas with insufficient charging facilities such as highways or rural areas are often not met by fixed charging stations. Therefore, the concept of mobile charging stations is proposed as a supplementary charging form for fixed charging stations. Under normal circumstances, idle mobile charging stations will not actively move in the direction of electric vehicles before receiving charging requests from them. However, if cloud servers are allowed to schedule idle mobile charging stations in advance according to these potential charging needs and actively track electric vehicles that may issue charging requests, more electric vehicles will be able to be charged in a timely manner, thereby improving the charging efficiency and charging benefits of charging stations.

[0003] However, when scheduling idle mobile charging stations, there are often some uncertain factors in the surrounding environment, which may make the optimal solution obtained by the original deterministic model meaningless. For the electric vehicle Internet of Things, the more significant uncertainty factor is the charging demand of electric vehicles. This demand is constantly fluctuating due to factors such as weather environment, road traffic flow, and driver travel patterns. At the same time, the size of the charging demand is closely related to multiple life-related factors such as rush hour, holidays, and passenger flow. Therefore, its fluctuation directly affects the results of the mobile charging station scheduling problem. The main challenge faced in making scheduling decisions is that the joint distribution of the uncertain variable charging demand is unknown. Due to the non-stationarity and spatial correlation of driver travel patterns, it is very difficult to obtain accurate distribution estimates from historical data, which is a common feature of the urban transportation industry. Summary of the invention

[0004] The main purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art and to provide a method and system for robust optimization scheduling of idle mobile charging stations that distinguish between scenarios, so that the scheduling plan of the charging station is more reasonable, can meet the charging needs of more electric vehicles, reduce operating costs, and improve the scheduling efficiency of the charging station.

[0005] According to one aspect of the present invention, the present invention provides a scenario-differentiated idle mobile charging station robust optimization scheduling method, the method comprising the following steps:

[0006] S1: Build a scheduling model for idle mobile charging stations;

[0007] S2: construct a fuzzy set based on historical charging demand data and weather data, classify the historical charging demand data into different scenarios, optimize the scheduling model based on the fuzzy set and the status information of the mobile charging station, and obtain the objective function of the robust optimization scheduling model;

[0008] S3: Solving the objective function of the robust optimization scheduling model, and scheduling idle mobile charging stations according to the solution results.

[0009] Preferably, the constructing of the scheduling model of the idle mobile charging station includes: determining the objective function of the scheduling model of the idle mobile charging station, which is as follows:

[0010]

[0011] in, is the total charging demand generated in area j during time period k to k+1, is the scheduling decision matrix, is the number of idle mobile charging stations dispatched from area i to area j at time k, W ij is the driving distance from area i to area j; is the number of idle mobile charging stations in area i at time k before making a scheduling decision, β is the weight coefficient, N k Supplying the total number of mobile charging stations in the entire city,

[0012] Preferably, the objective function of the robust optimization scheduling model obtained by constructing a fuzzy set based on historical charging demand data and weather data includes:

[0013] Using a multivariate regression tree, the weather data is used as a covariate to classify the historical data of charging demand into different scenarios. Assuming there are T historical charging demand samples and covariate observation sample Covariates The set of is represented by Ω, T is divided into L subsets Ω l , l∈[L], and for all l, k∈[L], l≠k, ∪ l∈[L] Ω l =Ω, exists p l is the probability of each scenario occurring, constructed as follows from the probability distribution The fuzzy sets are:

[0014]

[0015] The objective function in constructing the robust optimization model is as follows:

[0016]

[0017] Where M is the number of regions, [M] is a set from 1 to M, I is the number of covariate features, By charging demand and covariates The probability distribution that the combination obeys, is the probability distribution set of M+I dimensional variables, is The expected value of a variable under a probability distribution, P is the probability of an event; μ l is the mean value of charging demand in scenario l, σ jl is the standard deviation of charging demand in region j under scenario l, p l is the probability of occurrence of scenario l, is the set of upper and lower bounds in the charging demand historical data.

[0018] Preferably, the classifying the historical data of charging demand into different scenarios includes:

[0019] Solve the following equation:

[0020]

[0021] Among them, the j-th characteristic variable in the covariate and its value s as the segmentation variable and segmentation point, dividing the two regions into and and After finding the optimal (j, s), the input space is divided into two regions in turn, and the above division process is repeated for each region until the stopping condition is met, and a regression decision tree is generated. According to the regression decision tree, the data set is classified into a scene set with L subsets.

[0022] Preferably, scheduling the idle mobile charging stations according to the solution results includes:

[0023] Solve the objective function of the robust optimization model for distinguishing scenes to obtain the final decision matrix For the decision matrix X k Each element in At time k, the The idle mobile charging station sends a scheduling task to drive to area j to respond to the charging request of the electric vehicle.

[0024] According to another aspect of the present invention, the present invention also provides a scenario-differentiated idle mobile charging station robust optimization scheduling system, the system comprising:

[0025] A first building module is used to build a scheduling model for idle mobile charging stations;

[0026] The second construction module is used to construct a fuzzy set according to the historical charging demand data and the weather data, classify the historical charging demand data into different scenarios, optimize the scheduling model based on the fuzzy set and the status information of the mobile charging station, and obtain the objective function of the robust optimization scheduling model;

[0027] The solution scheduling module is used to solve the objective function of the robust optimization scheduling model and schedule the idle mobile charging stations according to the solution results.

[0028] Preferably, the first building module constructs the scheduling model of the idle mobile charging station, including: determining the objective function of the scheduling model of the idle mobile charging station, specifically as follows:

[0029]

[0030] in, is the total charging demand generated in area j during time period k to k+1, is the scheduling decision matrix, is the number of idle mobile charging stations dispatched from area i to area j at time k, W ij is the driving distance from area i to area j; is the number of idle mobile charging stations in area i at time k before making a scheduling decision, β is the weight coefficient, N k Supplying the total number of mobile charging stations in the entire city,

[0031] Preferably, the second building module builds a fuzzy set according to historical charging demand data and weather data, and obtains the objective function of the robust optimization scheduling model including:

[0032] Using a multivariate regression tree, the weather data is used as a covariate to classify the historical data of charging demand into different scenarios. Assuming there are T historical charging demand samples and covariate observation sample Covariates The set of is represented by Ω, T is divided into L subsets Ω l , l∈[L], and for all l, k∈[L], l≠k, ∪ l∈[L] Ω l =Ω, exists p l is the probability of each scenario occurring, constructed as follows from the probability distribution The fuzzy sets are:

[0033]

[0034] The objective function in constructing the robust optimization model is as follows:

[0035]

[0036] Where M is the number of regions, [M] is a set from 1 to M, I is the number of covariate features, By charging demand and covariates The probability distribution that the combination obeys, is the probability distribution set of M+I dimensional variables, is The expected value of a variable under a probability distribution, P is the probability of an event; μ l is the mean value of charging demand in scenario l, σ jl is the standard deviation of charging demand in region j under scenario l, p l is the probability of occurrence of scenario l, is the set of upper and lower bounds in the charging demand historical data.

[0037] Preferably, the classifying the historical data of charging demand into different scenarios includes:

[0038] Solve the following equation:

[0039]

[0040] Among them, the j-th characteristic variable in the covariate and its value s as the segmentation variable and segmentation point, dividing the two regions into and and After finding the optimal (j, s), the input space is divided into two regions in turn, and the above division process is repeated for each region until the stopping condition is met, and a regression decision tree is generated. According to the regression decision tree, the data set is classified into a scene set with L subsets.

[0041] Preferably, the solution scheduling module schedules idle mobile charging stations according to the solution result, including:

[0042] Solve the objective function of the robust optimization model for distinguishing scenes to obtain the final decision matrix For the decision matrix X k Each element in At time k, the The idle mobile charging station sends a scheduling task to drive to area j to respond to the charging request of the electric vehicle.

[0043] Beneficial effects: The present invention obtains fuzzy sets of uncertain variables in multiple scenarios by training and analyzing data such as historical charging demand and weather conditions in various areas of the city. Then, the objective function and constraint conditions of the robust optimization model are constructed according to the scenario set, and the decision matrix is ​​obtained after solving the model. The scheduling task is sent to the mobile charging station, making the scheduling plan of the charging station more reasonable, able to meet the charging needs of more electric vehicles, reduce operating costs, and improve the scheduling efficiency of the charging station.

[0044] The features and advantages of the present invention will become clear through reference to the following drawings and detailed description of specific embodiments of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a flow chart of a robust optimization scheduling method for idle mobile charging stations that distinguishes scenarios;

[0046] Figure 2 It is a schematic diagram of a scenario-differentiated robust optimization scheduling system for idle mobile charging stations. DETAILED DESCRIPTION

[0047] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] Example 1

[0049] Figure 1 This is a flowchart of a robust optimization scheduling method for idle mobile charging stations that distinguishes scenarios. Figure 1 As shown, the present invention provides a scenario-differentiated idle mobile charging station robust optimization scheduling method, the method comprising the following steps:

[0050] S1: Build a scheduling model for idle mobile charging stations;

[0051] S2: construct a fuzzy set based on historical charging demand data and weather data, classify the historical charging demand data into different scenarios, optimize the scheduling model based on the fuzzy set and the status information of the mobile charging station, and obtain the objective function of the robust optimization scheduling model;

[0052] In this step, the status information of the mobile charging station includes whether the charging station is in an idle state or in an occupied state for charging electric vehicles.

[0053] S3: Solving the objective function of the robust optimization scheduling model, and scheduling idle mobile charging stations according to the solution results.

[0054] In this method, the working status of the charging station and the fuzzy set constructed by historical charging demand data are considered to construct the objective function of the optimization scheduling model, so that the scheduling plan of the charging station is more reasonable, which can meet the charging needs of more electric vehicles, reduce operating costs, and improve the scheduling efficiency of the charging station.

[0055] Preferably, the constructing of the scheduling model of the idle mobile charging station includes: determining the objective function of the scheduling model of the idle mobile charging station, which is as follows:

[0056]

[0057] in, is the total charging demand generated in area j during time period k to k+1, is the scheduling decision matrix, is the number of idle mobile charging stations dispatched from area i to area j at time k, W ij is the driving distance from area i to area j; is the number of idle mobile charging stations in area i at time k before making a scheduling decision, β is the weight coefficient, N k Supplying the total number of mobile charging stations in the entire city,

[0058] Specifically, the scheduling of mobile charging stations has two goals. One is to meet as many charging requests of electric vehicles as possible and keep the supply-demand ratio of each area in the city as close as possible. The supply-demand ratio refers to the ratio of the number of idle mobile charging stations dispatched by the server to the area and the charging demand generated by electric vehicles in the area, so as to reduce the mismatch between supply and demand in all areas of the city; the second is to reduce the total idle driving distance of mobile charging stations to reduce operating costs. When making scheduling decisions, it is necessary to fully consider the charging demand in the future time period of each area of ​​the city, so that the mobile charging station can charge more electric vehicles to increase the total profit. In order to improve the calculation efficiency, the present invention assumes that the entire city is divided into n areas, and the time of a day is discretized and represented by indexes t=1, 2,..., K. When the server makes a scheduling decision at time t, it considers the future charging demand of the city in the time period (t, t+1), and sends the robust optimization model result, that is, the scheduling decision, to the mobile charging station to perform the scheduling task. The server collects the status information of the mobile charging station after scheduling at time t+1, and performs a new round of modeling and scheduling.

[0059] represents the empty driving distance after all mobile charging stations execute the decision, N is the number of idle mobile charging stations in area i from time period k to k+1 after making the scheduling decision. k The total number of mobile charging stations in the city. The difference between the supply-demand ratio of idle mobile charging stations in each area and the supply-demand ratio of the entire city. The smaller the value, the more evenly distributed the mobile charging stations are in each area. Therefore, when modeling, it is necessary to compromise between the two objectives and adjust the weights of the two by changing the β value.

[0060] Preferably, the objective function of the robust optimization scheduling model obtained by constructing a fuzzy set based on historical charging demand data and weather data includes:

[0061] Using a multivariate regression tree, the weather data is used as a covariate to classify the historical data of charging demand into different scenarios. Assuming there are T historical charging demand samples and covariate observation sample Covariates The set of is represented by Ω, T is divided into L subsets Ω l , l∈[L], and for all l, k∈[L], l≠k, ∪ l∈[L] Ω l =Ω, exists p l is the probability of each scenario occurring, constructed as follows from the probability distribution The fuzzy sets are:

[0062]

[0063] The objective function in constructing the robust optimization model is as follows:

[0064]

[0065] Where M is the number of regions, [M] is a set from 1 to M, I is the number of covariate features, By charging demand and covariates The probability distribution that the combination obeys, The probability distribution set of M+I-dimensional variables, is The expected value of a variable under a probability distribution, P is the probability of an event; μ l is the mean value of charging demand in scenario l, σ jl is the standard deviation of charging demand in region j under scenario l, p l is the probability of occurrence of scenario l, is the set of upper and lower bounds in the charging demand historical data.

[0066] Specifically, since a small disturbance in the uncertain parameters may make the optimization result infeasible, robust optimization uses fuzzy sets to evaluate uncertain variables in order to overcome this uncertainty and find a solution that must satisfy the constraints and is as optimal as possible. The charging demand determination value of each area within a specific time is often difficult to predict, so the present invention assumes that the probability distribution of the demand uncertainty belongs to the fuzzy set. For example, for the objective function min x f(x, ε), where ε is an uncertain parameter. Robust optimization takes ε as a fuzzy set μ. The optimization goal is to make the objective function have the best result value under the worst case of uncertain parameters, that is:

[0067] min x max ε∈μ f(x, ε)

[0068] Furthermore, in order to solve the problem of limited charging demand information, the present invention adopts a robust optimization framework that distinguishes scenarios. It is assumed that the uncertain demand is distributed in a specific fuzzy set, in which the support set of the uncertain variable is classified into multiple scenarios, and the moment information of the variable in each scenario is obtained. The classification method uses a multivariate regression tree to classify the historical data of charging demand into different scenarios based on weather data as a covariate. Assume that there are T historical charging demand and covariate observation samples, such as Covariates The set of is represented by Ω, T is divided into L subsets Ω l , l∈[L], and for all l, k∈[L], l≠k, ∪ l∈[L] Ω l =Ω, exists p l is the probability of each scenario occurring. Combining all scenarios and the moment information of uncertain variables in the corresponding scenarios, the following probability distribution is constructed: composed of fuzzy sets.

[0069] Among them, in this fuzzy set, The constraints specify the moment information of the demand variables in each scenario; The constraints specify the upper bound of the variance of the demand variable in each scenario; The constraints specify the probability of each scenario occurring; Constraints specify the upper and lower bounds on the support of the demand variables.

[0070] Preferably, the classifying the historical data of charging demand into different scenarios includes:

[0071] Solve the following equation:

[0072]

[0073] Among them, the j-th characteristic variable in the covariate and its value s as the segmentation variable and segmentation point, dividing the two regions into and and After finding the optimal (j, s), the input space is divided into two regions in turn, and the above division process is repeated for each region until the stopping condition is met, and a regression decision tree is generated. According to the regression decision tree, the data set is classified into a scene set with L subsets.

[0074] Specifically, the uncertain parameter of charging demand is expressed in the model as follows: The total amount of charging demand issued by electric vehicles in each area at all scheduling times is represented in the model as a K×n two-dimensional matrix. Furthermore, in order to construct the fuzzy set required for the robust optimization model, the historical data set of the parameter needs to be converted into a set containing L scenarios.

[0075] This embodiment uses a decision tree model to construct a scene set. The tree structure is specified by a set of splitting rules that can divide the covariate space. Specifically, each recursive split divides the set Ω into two subsets. A decision tree with L leaf nodes correspondingly divides the set Ω into L subsets, and the leaf nodes correspond to the subsets one by one. At the same time, the data set T also falls into the L scene sets according to the decision tree rules. The splitting rule is the core of the decision tree, and it uses a heuristic method to divide the feature space. For the covariate n is the covariate The number of features. Each time the decision tree is divided, all the values ​​of each feature in the current covariate set are examined one by one, and the best one is selected as the split point according to the square error minimization criterion. The jth feature variable in and its value s as the segmentation variable and segmentation point, and assume that the two regions and And find the optimal j and s, and solve the following equation:

[0076]

[0077] That is to find j and s with the smallest sum of square errors in the two regions after division. Among them, c1 and c2 are fixed output values ​​in the two regions after division. The two min in square brackets mean that the optimal c1 and c2 are used, that is, c1 and c2 that minimize the square error in each region. It is easy to know that these two optimal output values ​​are the values ​​in the corresponding regions of each group. The mean of , so the above formula can be written as:

[0078]

[0079] in and After finding the optimal split point (j, s), the input space is divided into two regions in turn, and then the above division process is repeated for each region until the stopping condition is met. In this way, a regression decision tree is generated, according to which the data set can be classified into a scene set with L subsets.

[0080] Preferably, scheduling the idle mobile charging stations according to the solution results includes:

[0081] Solve the objective function of the robust optimization model for distinguishing scenes to obtain the final decision matrix For the decision matrix X k Each element in At time k, the The idle mobile charging station sends a scheduling task to drive to area j to respond to the charging request of the electric vehicle.

[0082] Specifically, the RSOME toolkit is imported into Matlab, the objective function and constraints are input, and the historical data of charging demand are added. The robust optimization model for differentiating scenarios is solved using Gurobi or CPLEX solver to obtain the final decision matrix. Observe the decision matrix, generally and At least one value in is 0, which means that the two regions will not dispatch idle mobile charging stations to each other at the same time. k Each element in The server sends data to the i region at time k. The idle mobile charging station sends a scheduling task to drive to area j to respond to the charging request of the electric vehicle.

[0083] The method of this embodiment can be executed by a server. At time t, the server constructs a fuzzy set based on the historical charging demand and weather data at the current time, determines the status information of all mobile charging stations at the current time, and performs modeling based on the fuzzy set and the status information. The server sends a scheduling task to the mobile charging station based on the solution result. After one unit of time, the next round of modeling and solving is performed at time t+1 until the last time index.

[0084] This embodiment obtains fuzzy sets of uncertain variables in multiple scenarios by training and analyzing data such as historical charging demand and weather conditions in various areas of the city. Then, the objective function and constraints of the robust optimization model are constructed based on the scenario set, and the decision matrix is ​​obtained after solving the model. The scheduling task is sent to the mobile charging station, making the scheduling plan of the charging station more reasonable, meeting the charging needs of more electric vehicles, reducing operating costs, and improving the scheduling efficiency of the charging station.

[0085] Example 2

[0086] Figure 2 Figure 1 is a schematic diagram of a robust optimization scheduling system for idle mobile charging stations that distinguishes scenarios. Figure 2 As shown, this embodiment provides a scenario-differentiated idle mobile charging station robust optimization scheduling system, the system comprising:

[0087] The first construction module 201 is used to construct a scheduling model for idle mobile charging stations;

[0088] A second construction module 202 is used to construct a fuzzy set based on historical charging demand data and weather data, classify the historical charging demand data into different scenarios, optimize the scheduling model based on the fuzzy set and the status information of the mobile charging station, and obtain the objective function of the robust optimization scheduling model;

[0089] The scheduling solution module 203 is used to solve the objective function of the robust optimization scheduling model and schedule the idle mobile charging stations according to the solution result.

[0090] Preferably, the first construction module 201 constructs the scheduling model of the idle mobile charging station, including: determining the objective function of the scheduling model of the idle mobile charging station, specifically as follows:

[0091]

[0092] in, is the total charging demand generated in area j during time period k to k+1, is the scheduling decision matrix, is the number of idle mobile charging stations dispatched from area i to area j at time k, W ij is the driving distance from area i to area j; is the number of idle mobile charging stations in area i at time k before making a scheduling decision, β is the weight coefficient, N k Supplying the total number of mobile charging stations in the entire city,

[0093] Preferably, the second construction module 202 constructs a fuzzy set according to the historical charging demand data and weather data, and obtains the objective function of the robust optimization scheduling model including:

[0094] Using a multivariate regression tree, the weather data is used as a covariate to classify the historical data of charging demand into different scenarios. Assuming there are T historical charging demand samples and covariate observation sample Covariates The average set is represented by Ω, and T is divided into L subsets Ω l , l∈[L], and for all l, k∈[L], l≠k, ∪ l∈[L] Ω l =Ω, exists p l is the probability of each scenario occurring, constructed as follows from the probability distribution The fuzzy sets are:

[0095]

[0096] The objective function in constructing the robust optimization model is as follows:

[0097]

[0098] Where M is the number of regions, [M] is a set from 1 to M, I is the number of covariate features, By charging demand and covariates The probability distribution that the combination obeys, is the probability distribution set of M+I dimensional variables, is The expected value of a variable under a probability distribution, P is the probability of an event; μ l is the mean value of charging demand in scenario l, σ jl is the standard deviation of charging demand in region j under scenario l, p l is the probability of occurrence of scenario l, z l is the set of upper and lower bounds in the charging demand historical data.

[0099] Preferably, the classifying the historical data of charging demand into different scenarios includes:

[0100] Solve the following equation:

[0101]

[0102] Among them, the j-th characteristic variable in the covariate and its value s as the segmentation variable and segmentation point, dividing the two regions into and and After finding the optimal (j, s), the input space is divided into two regions in turn, and the above division process is repeated for each region until the stopping condition is met, and a regression decision tree is generated. According to the regression decision tree, the data set is classified into a scene set with L subsets.

[0103] Preferably, the solution scheduling module 203 schedules idle mobile charging stations according to the solution result, including:

[0104] Solve the objective function of the robust optimization model for distinguishing scenes to obtain the final decision matrix For each element in the decision matrix Xk At time k, the The idle mobile charging station sends a scheduling task to drive to area j to respond to the charging request of the electric vehicle.

[0105] The specific implementation process of the functions realized by each module in this embodiment 2 is the same as the implementation process in embodiment 1, and will not be repeated here.

[0106] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. All equivalent structural changes made by using the contents of the present invention specification and drawings under the concept of the present invention, or directly / indirectly applied in other related technical fields are included in the patent protection scope of the present invention.

Claims

1. A scenario-based robust optimization scheduling method for idle mobile charging stations, characterized in that: The method comprises the following steps: S1: Build a scheduling model for idle mobile charging stations; S2: construct a fuzzy set based on historical charging demand data and weather data, classify the historical charging demand data into different scenarios, optimize the scheduling model based on the fuzzy set and the status information of the mobile charging station, and obtain the objective function of the robust optimization scheduling model; S3: solving the objective function of the robust optimization scheduling model, and scheduling the idle mobile charging stations according to the solution result; The construction of the scheduling model of the idle mobile charging station includes: determining the objective function of the scheduling model of the idle mobile charging station, which is as follows: in, For the period to ,area The total charging demand generated within is the scheduling decision matrix, For at the moment From area Dispatch to area The number of idle mobile charging stations, ; From the area To area driving distance; To make a scheduling decision before ,area The number of idle mobile charging stations, is the weight coefficient, Supplying the total number of mobile charging stations in the entire city, ； The fuzzy set is constructed according to the historical charging demand data and weather data, the historical charging demand data is classified into different scenarios, and the scheduling model is optimized based on the fuzzy set and the status information of the mobile charging station. The objective function of the robust optimization scheduling model includes: Using a multivariate regression tree, the weather data is used as a covariate to classify the historical data of charging demand into different scenarios; assuming that Historical charging demand samples and covariate observation sample , , covariate The collection of express, Divided into Subset , and for all , ,exist ; is the probability of each scenario occurring, constructed as follows from the probability distribution The fuzzy set composed of: The objective function in constructing the robust optimization scheduling model is as follows: in, is the number of regions, is a set from 1 to M, is the number of covariate features, By charging demand and covariates The probability distribution that the combination obeys, yes The set of probability distributions of dimensional variables, is The expected value of a variable under a probability distribution, P is the probability of an event occurring; is The average charging demand in the scenario, is In the scenario, Standard deviation of charging demand in the region, yes The probability of the scenario occurring, is the set of upper and lower bounds in the charging demand historical data.

2. The method according to claim 1, characterized in that The classification of the historical data of charging demand into different scenarios includes: solving the following equation: Among them, the covariate Feature variables and its value As the dividing variable and dividing point, the two regions are divided into and , ,and , find the optimal Then, the input space is divided into two regions in turn, and the above division process is repeated for each region until the stop condition is met, and a regression decision tree is generated. According to the regression decision tree, the data set is classified into A subset of scenes.

3. The method according to claim 2, characterized in that The scheduling of idle mobile charging stations according to the solution results includes: Solve the objective function of the robust optimization model for distinguishing scenes to obtain the final decision matrix , for the decision matrix Each element in ,exist Always In the region Send a dispatch task to an idle mobile charging station and ask it to drive to Regions respond to charging requests from electric vehicles.

4. A scenario-differentiated idle mobile charging station robust optimization scheduling system, characterized in that: The system comprises: A first building module is used to build a scheduling model for idle mobile charging stations; The second construction module is used to construct a fuzzy set according to the historical charging demand data and the weather data, classify the historical charging demand data into different scenarios, optimize the scheduling model based on the fuzzy set and the status information of the mobile charging station, and obtain the objective function of the robust optimization scheduling model; A solution scheduling module, used to solve the objective function of the robust optimization scheduling model and schedule idle mobile charging stations according to the solution result; The first construction module constructs the scheduling model of the idle mobile charging station, including: determining the objective function of the scheduling model of the idle mobile charging station, which is as follows: in, For the period to ,area The total charging demand generated within is the scheduling decision matrix, For at the moment From area Dispatch to area The number of idle mobile charging stations, ; From the area To area driving distance; To make a scheduling decision before ,area The number of idle mobile charging stations, is the weight coefficient, Supplying the total number of mobile charging stations in the entire city, The second construction module constructs a fuzzy set based on historical charging demand data and weather data, classifies the historical charging demand data into different scenarios, optimizes the scheduling model based on the fuzzy set and the status information of the mobile charging station, and obtains the objective function of the robust optimization scheduling model including: Using a multivariate regression tree, the weather data is used as a covariate to classify the historical data of charging demand into different scenarios; assuming that Historical charging demand samples and covariate observation sample , , covariate The collection of express, Divided into Subset , and for all , ,exist ; is the probability of each scenario occurring, constructed as follows from the probability distribution The fuzzy set composed of: The objective function in constructing the robust optimization scheduling model is as follows: in, is the number of regions, is a set from 1 to M, is the number of covariate features, By charging demand and covariates The probability distribution that the combination obeys, yes The set of probability distributions of dimensional variables, is The expected value of a variable under a probability distribution, P is the probability of an event occurring; is The average charging demand in the scenario, is In the scenario, Standard deviation of charging demand in the region, yes The probability of the scenario occurring, is the set of upper and lower bounds in the charging demand historical data.

5. The system according to claim 4, characterized in that The classification of the historical data of charging demand into different scenarios includes: Solve the following equation: Among them, the covariate Feature variables and its value As the dividing variable and dividing point, the two regions are divided into and , ,and , find the optimal Then, the input space is divided into two regions in turn, and the above division process is repeated for each region until the stop condition is met, and a regression decision tree is generated. According to the regression decision tree, the data set is classified into A subset of scenes.

6. The system according to claim 5, characterized in that The solution scheduling module schedules the idle mobile charging stations according to the solution result, including: Solve the objective function of the robust optimization model for distinguishing scenes to obtain the final decision matrix , for the decision matrix Each element in ,exist Always In the region Send a dispatch task to an idle mobile charging station and ask it to drive to Regions respond to charging requests from electric vehicles.

Images