Method and device for determining charging quota, computer device and storage medium

Abstract

The application relates to a charging quota determination method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: obtaining historical charging amounts of a plurality of charging objects in each unit period in a plurality of historical statistical periods; determining, according to the historical charging amounts, a total charging quota of each charging object in a target statistical period and an optimization weight value corresponding to each unit period; determining a total loss based on the optimization weight values; meeting the condition that the total loss is minimum; and performing quota allocation on the total charging quota of each charging object in the target statistical period according to the optimization weight value corresponding to each unit period, to obtain a unit charging quota of each charging object in each unit period in the target statistical period. The method can save energy resources. The application can be applied to the transportation field.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, computer device, and storage medium for determining charging quotas. Background Technology

[0002] Currently, electric vehicles are receiving increasing attention from various countries for their energy conservation, emission reduction, and environmental improvement benefits, and are being actively promoted and applied in multiple fields. Electric vehicles refer to vehicles that use onboard power sources to drive their wheels with electric motors and comply with all road traffic and safety regulations.

[0003] With the continuous increase in the number of electric vehicles, the disorderly charging of a large number of electric vehicles will increase the peak-valley difference of the power grid load, resulting in increased network losses, affecting the power quality of the distribution network system, and thus wasting energy resources such as electricity. Therefore, it is urgent to establish a charging quota mechanism for electric vehicles to reduce the waste of energy resources such as electricity. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for determining charging quotas that can save energy resources, in order to address the above-mentioned technical problems.

[0005] In a first aspect, this application provides a method for determining charging quotas, the method comprising:

[0006] Get the historical charging amount of multiple charging objects in each unit time period across multiple historical statistical periods;

[0007] Based on the historical charging volume, the total charging quota for each charging object within the target statistical period and the corresponding optimized weight value for each unit period are determined. The total loss determined based on the optimized weight values ​​satisfies the condition of minimum total loss. The total loss is determined based on the sum of multiple losses. Each loss represents the difference between the statistical total charging volume and the predicted total charging volume of the corresponding charging object. The statistical total charging volume represents the charging volume of the corresponding charging object within the unit statistical period, obtained from statistics. The predicted total charging volume represents the charging volume of the corresponding charging object within the unit statistical period, determined based on the optimized weight values.

[0008] Referring to the optimization weight value corresponding to each of the aforementioned unit time periods, the total charging quota for each of the aforementioned charging objects within the target statistical period is allocated, thereby obtaining the unit charging quota for each of the aforementioned unit time periods within the target statistical period for each of the aforementioned charging objects.

[0009] In one embodiment, the optimized weight value is determined based on the value of the Gaussian weight parameter in the Gaussian mixture model. The value of the Gaussian weight parameter in the Gaussian mixture model is obtained after multiple rounds of parameter value adjustment. The step of determining the value of the Gaussian weight parameter in the current round includes:

[0010] The first total loss for the current round is determined by the total loss model and based on the values ​​of the Gaussian weight parameters corresponding to each of the time periods in the current round and the average historical total charging amount of each of the charging objects.

[0011] Determine the values ​​of the loss weight parameters corresponding to each unit time period in the current round, as output by the loss weight parameter model in the total loss model.

[0012] The second total loss for the current round is determined using the total loss model and based on the values ​​of the loss weight parameters corresponding to each unit time period in the current round.

[0013] Based on the first total loss and the second total loss of the current round, determine the value of the Gaussian weight parameter corresponding to each unit time period of the current round after adjustment.

[0014] In one embodiment, determining the adjusted Gaussian weight parameter value for each unit time period of the current round based on the first total loss and the second total loss of the current round includes:

[0015] When the first total loss of the current round is less than the second total loss of the current round, the value of the Gaussian weight parameter corresponding to each unit time period of the current round is used as the adjusted value of the Gaussian weight parameter corresponding to each unit time period of the current round.

[0016] When the first total loss of the current round is greater than the second total loss of the current round, the value of the loss weight parameter corresponding to each unit time period of the current round is used as the value of the Gaussian weight parameter corresponding to each unit time period of the current round after adjustment.

[0017] In one embodiment, each of the plurality of charging objects is taken as a target charging object, and the total charging quota of the target charging object in the target statistical period includes a total charging quota upper limit and a total charging quota lower limit.

[0018] The steps for determining the unit charging quota for the target charging object in each unit time period within the target statistical period include:

[0019] For each of the multiple optimization weight values, the current optimization weight value is multiplied by the total charging quota limit of the target charging object to obtain the unit charging quota limit of the type per unit time period corresponding to the current optimization weight value;

[0020] Multiply the current optimization weight value by the total charging quota lower limit of the target charging object to obtain the unit charging quota lower limit of the type per unit time period corresponding to the current optimization weight value;

[0021] By combining the upper limit and the lower limit of the unit charging quota, the unit charging quota of the target charging object within the unit time period corresponding to the current optimization weight value is obtained.

[0022] In one embodiment, the unit statistical time period includes multiple unit time periods, with each of the multiple charging objects serving as a target charging object, and each of the multiple unit time periods serving as a target unit time period. The method further includes:

[0023] Determine the actual charging amount of the target charging object within a target unit time period of the target statistical period;

[0024] When the actual charging amount is less than the unit charging quota of the target charging object in the target unit time period of the target statistical period, the difference between the unit charging quota of the target charging object in the target unit time period of the target statistical period and the actual charging amount is stored in the cloud.

[0025] When the actual charging amount is greater than the unit charging quota of the target charging object in the target unit time period of the target statistical period, apply for additional charging quota from the cloud.

[0026] Secondly, this application also provides a charging quota determination device, the device comprising:

[0027] The historical charging amount determination module is used to obtain the historical charging amount of multiple charging objects in each unit time period in multiple historical statistical periods.

[0028] The information determination module is used to determine, based on the historical charging amounts, the total charging quota for each charging object within the target statistical period and the optimized weight value corresponding to each of the unit periods; the total loss determined based on the optimized weight values ​​satisfies the condition of minimum total loss; the total loss is determined based on the sum of multiple losses; each loss represents the difference between the statistical total charging amount and the predicted total charging amount of the corresponding charging object; the statistical total charging amount represents the charging amount of the corresponding charging object within the unit statistical period obtained from statistics; the predicted total charging amount represents the charging amount of the corresponding charging object within the unit statistical period determined based on the optimized weight values.

[0029] The quota allocation module is used to allocate the total charging quota of each charging object in the target statistical period by referring to the optimization weight value corresponding to each of the unit time periods, so as to obtain the unit charging quota of each charging object in each of the unit time periods in the target statistical period.

[0030] In one embodiment, the information determination module further includes a total charging quota determination module, which is used to take each of the plurality of charging objects as a target charging object, and for each of the plurality of historical statistical periods, to sum up the historical charging amount of the target charging object in each unit time period in the current historical statistical period to obtain the historical total charging amount of the target charging object in the current historical statistical period; to average the historical total charging amount of the target charging object in each of the historical statistical periods to obtain the average historical total charging amount of the target charging object; to determine the historical total power variance based on the historical total charging average and the historical total charging amount of the target charging object in each of the historical statistical periods; and to determine the total charging quota to be allocated to the target charging object in the target statistical period under a preset confidence level based on the historical total charging average and the historical total power variance of the target charging object.

[0031] In one embodiment, the information determination module further includes an optimization weight value determination module for obtaining a Gaussian mixture model; the Gaussian mixture model represents a combination of probability distributions of the charging object's charging amount in each unit time period, and includes Gaussian weight parameters corresponding to each unit time period; based on each of the historical charging amounts, the average historical charging amount of each charging object in each of the unit time periods is determined, and based on the total loss model and the average historical charging amounts, the values ​​of the Gaussian weight parameters in the Gaussian mixture model corresponding to each of the unit time periods are adjusted to obtain the optimization weight values ​​corresponding to each of the unit time periods.

[0032] In one embodiment, the optimization weight value determination module is further configured to take each of the plurality of charging objects as a target charging object, extract the target historical charging amount of the target charging object in each unit time period in the plurality of historical statistical periods from the acquired historical charging amounts, and determine the average historical charging amount of the target charging object in each unit time period based on the extracted target historical charging amounts.

[0033] In one embodiment, the unit statistical time period includes multiple unit time periods. The optimization weight value determination module is further configured to take each of the multiple unit time periods as a target unit time period, and to superimpose the target historical charging amount of the target charging object in the target unit time period of each of the historical statistical time periods to obtain the unit historical total charging amount in the target unit time period; and to divide the unit historical total charging amount by the total number of the multiple historical statistical time periods to obtain the average historical charging amount of the target charging object in the target unit time period.

[0034] In one embodiment, the optimization weight value determination module is further configured to obtain an initialized model parameter vector set; the model parameter vector set includes multiple model parameter vectors; the model parameter vectors are vectors determined based on the values ​​of multiple types of model parameters in the Gaussian mixture model; based on the initialized model parameter vector set and the historical average charging amount of each charging object in each unit time period, a target model parameter vector set for the first round is determined; from the second round after the first round, the initial model parameter vector set for the current round is determined based on the target model parameter vector set of the previous round. Based on the total loss model, the initial model parameter vector set of the current round is adjusted to obtain the target model parameter vector set of the current round; the next round is taken as the new current round, and the process of determining the initial model parameter vector set of the current round based on the target model parameter vector set of the previous round is returned to the current round starting from the second round after the first round and continues until a preset stopping condition is met, thus obtaining the target model parameter vector set of the final round. The values ​​of the Gaussian weight parameters corresponding to each unit time period in the target model parameter vector set of the final round are all used as optimization weight values.

[0035] In one embodiment, the Gaussian mixture model includes Gaussian weight parameters, mean parameters, and standard deviation parameters. The Gaussian weight parameters are determined based on probability density parameters. The optimized weight value determination module is further configured to obtain a probability density parameter model corresponding to the probability density parameters. Using the probability density parameter model, and based on the initialized model parameter vector group and the historical average charging amount of each charging object within each unit time period, multiple initial probability density parameter values ​​are obtained. The module also obtains the Gaussian weight parameter model corresponding to the Gaussian weight parameters, the mean parameter model corresponding to the mean parameters, and the standard deviation parameter model corresponding to the standard deviation parameters. The Gaussian weight parameters are then used to determine the optimal weight values. The parameter model, the mean parameter model, and the standard deviation parameter model are used. Based on the values ​​of the multiple initial probability density parameters and the average historical charging amount, the values ​​of the Gaussian weight parameters, the mean parameters, and the standard deviation parameters for each unit time period in the first round are obtained. The values ​​of the Gaussian weight parameters for each unit time period in the first round are adjusted based on the total loss model to obtain the adjusted values ​​of the multiple Gaussian weight parameters for the first round. The target model parameter vector group for the first round is obtained by combining the values ​​of the multiple standard deviation parameters, the multiple mean parameters, and the adjusted values ​​of the multiple Gaussian weight parameters for the first round.

[0036] In one embodiment, the Gaussian mixture model includes Gaussian weight parameters, mean parameters, and standard deviation parameters. The Gaussian weight parameters are determined based on probability density parameters. The optimization weight value determination module is further configured to determine the values ​​of multiple probability density parameters for the current round using a probability density parameter model corresponding to the probability density parameters, and based on the target model parameter vector group of the previous round and the historical average charging amount of each charging object in each unit time period. The module also obtains an initial model parameter vector group for the current round using the Gaussian weight parameter model corresponding to the Gaussian weight parameters, the mean parameter model corresponding to the mean parameters, and the standard deviation parameter model corresponding to the standard deviation parameters, and based on the values ​​of multiple probability density parameters for the current round and the historical average charging amount of each charging object in each unit time period.

[0037] In one embodiment, the optimization weight value determination module is further configured to: determine the value of the Gaussian weight parameter corresponding to each unit time period of the current round using a Gaussian weight parameter model corresponding to the Gaussian weight parameter and based on the values ​​of multiple probability density parameters of the current round; determine the value of the mean parameter corresponding to each unit time period of the current round using a mean parameter model corresponding to the mean parameter and based on the values ​​of multiple probability density parameters of the current round and the historical average charging amount of each charging object in each unit time period; determine the value of the standard deviation parameter corresponding to each unit time period of the current round using a standard deviation parameter model corresponding to the standard deviation parameter and based on the values ​​of multiple probability density parameters of the current round, the historical average charging amount of each charging object in each unit time period, and the value of the mean parameter corresponding to each unit time period of the current round; and obtain an initial model parameter vector group for the current round by combining the values ​​of the multiple Gaussian weight parameters, the multiple mean parameters, and the multiple standard deviation parameters of the current round.

[0038] In one embodiment, the optimization weight value determination module is further configured to adjust the values ​​of the Gaussian weight parameters corresponding to each unit time period in the initial model parameter vector group of the current round using the total loss model, to obtain the adjusted values ​​of the Gaussian weight parameters corresponding to each unit time period in the current round; and update the initial model parameter vector group of the current round according to the adjusted values ​​of the Gaussian weight parameters corresponding to each unit time period in the current round, to obtain the target model parameter vector group of the current round.

[0039] In one embodiment, the optimization weight value determination module is further configured to: determine the first total loss of the current round using the total loss model and based on the values ​​of the Gaussian weight parameters corresponding to each unit time period of the current round and the historical average total charging amount corresponding to each of the charging objects; determine the values ​​of the loss weight parameters output by the loss weight parameter model corresponding to the loss weight parameters in the total loss model and corresponding to each unit time period of the current round; determine the second total loss of the current round using the total loss model and based on the values ​​of the loss weight parameters corresponding to each unit time period of the current round; and determine the adjusted values ​​of the Gaussian weight parameters corresponding to each unit time period of the current round based on the first total loss and the second total loss of the current round.

[0040] In one embodiment, the optimization weight value determination module is further configured to: when the first total loss of the current round is less than the second total loss of the current round, use the value of the Gaussian weight parameter corresponding to each unit time period of the current round as the adjusted value of the Gaussian weight parameter corresponding to each unit time period of the current round; when the first total loss of the current round is greater than the second total loss of the current round, use the value of the loss weight parameter corresponding to each unit time period of the current round as the adjusted value of the Gaussian weight parameter corresponding to each unit time period of the current round.

[0041] In one embodiment, each of the plurality of charging objects is taken as a target charging object. The total charging quota of the target charging object in the target statistical period includes a total charging quota upper limit and a total charging quota lower limit. The quota allocation module is further configured to, for each of the plurality of optimization weight values, multiply the current optimization weight value by the total charging quota upper limit of the target charging object to obtain the unit charging quota upper limit of the unit time period corresponding to the type of the current optimization weight value; multiply the current optimization weight value by the total charging quota lower limit of the target charging object to obtain the unit charging quota lower limit of the unit time period corresponding to the type of the current optimization weight value; and combine the unit charging quota upper limit and the unit charging quota lower limit to obtain the unit charging quota of the target charging object in the unit time period corresponding to the type of the current optimization weight value.

[0042] In one embodiment, the unit statistical period includes multiple unit time periods, with each of the multiple charging objects serving as a target charging object and each of the multiple unit time periods serving as a target unit time period. The charging quota determination device further includes a cloud processing module, used to determine the actual charging amount of the target charging object within the target unit time period of the target statistical period; when the actual charging amount is less than the unit charging quota of the target charging object within the target unit time period of the target statistical period, the difference between the unit charging quota of the target charging object within the target unit time period of the target statistical period and the actual charging amount is stored in the cloud; when the actual charging amount is greater than the unit charging quota of the target charging object within the target unit time period of the target statistical period, an additional charging quota is requested from the cloud.

[0043] Thirdly, this application also provides a computer device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in any of the charging quota determination methods provided in the embodiments of this application.

[0044] Fourthly, this application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps in any of the charging quota determination methods provided in the embodiments of this application.

[0045] Fifthly, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the charging quota determination methods provided in the embodiments of this application.

[0046] The aforementioned method, apparatus, computer equipment, storage medium, and computer program product for determining charging quotas, by acquiring the historical charging volume of multiple charging objects within each unit time period in multiple historical statistical periods, can determine the total charging quota for each charging object in the target statistical period and the corresponding optimized weight value for each unit time period based on the historical charging volume. By determining the total charging quota and optimized weight values, the total charging quota can be allocated based on each optimized weight value, thereby obtaining the unit charging quota for each charging object within each unit time period in the target statistical period. Since the total charging quota is determined based on historical charging volume, the determined total charging quota conforms to the historical charging patterns of the charging objects, thus making the predicted total charging quota more accurate. This, in turn, reduces the waste of energy resources such as electricity based on the accurately determined total charging quota.

[0047] Furthermore, since loss refers to the difference between the total charging amount actually used within the statistical period and the total charging amount predicted based on multiple optimized weight values, when the total loss determined based on each optimized weight value satisfies the condition of minimizing the total loss, it indicates that the total charging amount predicted based on each optimized weight value within the statistical period is closer to the total charging amount actually needed by the charging object within the statistical period. Since the optimized weight values ​​correspond to the corresponding unit time period, when the total charging amount predicted based on each optimized weight value within the statistical period matches the actual usage, the unit charging amount predicted based on a single optimized weight value within a unit time period will also match the actual usage of the charging object within that unit time period. Therefore, the unit charging quota obtained by allocating the total charging quota through optimized weight values ​​will also be closer to the actual charging amount needed by the charging object within that unit time period, thus achieving the goal of accurately determining the unit charging quota. Because the unit charging quota for each unit time period is accurately determined, the arbitrary abuse of electricity and other energy resources by the charging object can be reduced, thereby saving electricity and other energy resources. Attached Figure Description

[0048] Figure 1This is an application environment diagram of a method for determining charging quotas in one embodiment;

[0049] Figure 2 This is a flowchart illustrating a method for determining charging quotas in one embodiment;

[0050] Figure 3 This is a schematic diagram illustrating the allocation of total charging quota in one embodiment;

[0051] Figure 4 A schematic diagram of charging an object in one embodiment;

[0052] Figure 5 This is a schematic diagram illustrating the iterative adjustment of parameter values ​​in one embodiment;

[0053] Figure 6 This is a schematic diagram illustrating the adjustment of the Gaussian weight parameter values ​​in one embodiment;

[0054] Figure 7 This is a schematic diagram of the overall process for determining multiple optimization weight values ​​in one embodiment;

[0055] Figure 8 This is a schematic diagram of a charging quota determination system in one embodiment;

[0056] Figure 9 This is a schematic diagram illustrating the storage and application of charging quotas in one embodiment;

[0057] Figure 10 This is a flowchart illustrating a method for determining charging quotas in one embodiment;

[0058] Figure 11 This is a structural block diagram of a charging quota determination device in one embodiment;

[0059] Figure 12 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0060] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0061] The method for determining charging quotas provided in this application embodiment can be applied to, for example, Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated on server 104 or placed in the cloud or on other servers. Both terminal 102 and server 104 can be used independently to execute the charging quota determination method provided in this embodiment. Terminal 102 and server 104 can also work together to execute the charging quota determination method provided in this embodiment. Taking the example of terminal 102 and server 104 working together to execute the charging quota determination method provided in this embodiment, terminal 102 can run a charging quota determination platform, through which a charging quota determination request can be initiated. Terminal 102 can send the charging quota determination request to server 104, so that server 104 responds to the charging quota determination request, determines the identifiers of each charging object carried in the charging quota determination request, and determines the unit charging quota of each charging object corresponding to each charging object identifier. The terminal 102 can be, but is not limited to, various desktop computers, laptops, smartphones, tablets, aircraft, electric vehicles, IoT devices, and portable wearable devices. IoT devices can include smart voice interaction devices, smart home appliances, and smart in-vehicle devices. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted devices. The server 104 can be implemented using a standalone server or a server cluster consisting of multiple servers.

[0062] This invention can be applied to the transportation sector. For example, the charging target in this application can be an electric vehicle. By determining the charging quota for each electric vehicle, an energy-saving intelligent transportation system can be obtained. An Intelligent Traffic System (ITS), also known as an Intelligent Transportation System, effectively integrates advanced technologies (information technology, computer technology, data communication technology, sensor technology, electronic control technology, automatic control theory, operations research, artificial intelligence, etc.) into transportation, service control, and vehicle manufacturing. It strengthens the connection between vehicles, roads, and users, thereby forming a comprehensive transportation system that ensures safety, improves efficiency, improves the environment, and saves energy.

[0063] This application also relates to the field of artificial intelligence. For example, this application can obtain the optimization weight value corresponding to each unit time period through an artificial intelligence model. Artificial intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain optimal results. In other words, artificial intelligence is a comprehensive technology of computer science that attempts to understand the essence of intelligence and produce a new kind of intelligent machine that can react in a way similar to human intelligence. Artificial intelligence is also the study of the design principles and implementation methods of various intelligent machines, enabling machines to have the functions of perception, reasoning, and decision-making.

[0064] In one embodiment, such as Figure 2 As shown, a method for determining charging quotas is provided. Taking the application of this method to computer equipment as an example, the computer equipment can provide... Figure 1 For terminals or servers within the system, determining charging quotas includes the following steps:

[0065] Step 202: Obtain the historical charging amount of multiple charging objects within each unit time period in multiple historical statistical time periods.

[0066] The charging object refers to any object powered by electrical resources, such as electric vehicles or electric aircraft. Electric vehicles can be further categorized into manned and unmanned electric vehicles. The historical statistical period refers to a single historical statistical period. For example, each year can be considered a statistical period, and the corresponding historical statistical period could be the past year. A statistical period can include multiple unit periods; for instance, a year can include 12 months, with each different month constituting a unit period. For example, January could be one unit period, February another, and so on. It's easy to understand that a statistical period can also be January, with each day of that month constituting a unit period.

[0067] Specifically, when it is necessary to determine the charging quota for multiple charging objects, the computer equipment can identify each charging object for which the charging quota is to be obtained, and acquire the historical charging amount for each charging object within each unit time period in multiple historical charging statistical periods. For example, when it is necessary to determine the charging quota for charging objects A to D, for each charging object A to D, the computer equipment can acquire the historical charging amount for each of the corresponding charging objects for 12 months in each of the past N years. Here, historical charging amount refers to the amount of electricity replenished by the charging object within a unit time period in the historical statistical period. For example, when the charging object is an electric vehicle, the historical charging amount can be the amount of electricity replenished by the electric vehicle from the charging station in month B of year A in the past.

[0068] In one embodiment, the historical charging amount of multiple charging objects within each unit time period in multiple historical statistical time periods can be denoted as: {x ijk |i=1,...,m;j=1,...,g;k=1,...,N}. Where i represents the charging object i; m represents the total number of charging objects; j represents the unit time period j, for example, unit time period j could be January, February, March, etc.; g represents the total number of categories of various unit time periods, for example, when the statistical period is "year", the total number of categories of various unit time periods is 12; k represents the historical statistical period k; N represents the total number of multiple historical statistical periods; x ijk This represents the historical charging amount of charging object i within a unit time period j in the historical statistical time period k.

[0069] In one embodiment, when the charging target is an electric vehicle, the computer device can connect to the on-board terminal of each electric vehicle in the target area, thereby obtaining multiple historical charging amounts recorded in each on-board terminal with the permission, and then allocating charging quotas to each electric vehicle in the target area based on the obtained historical charging amounts, thus achieving the purpose of unified management of electric vehicles in the target area.

[0070] In one embodiment, each charging object may also proactively report its historical charging amount in each unit time period of multiple historical statistical periods to the computer device.

[0071] Step 204: Based on the historical charging volume, determine the total charging quota for each charging object within the target statistical period and the corresponding optimized weight value for each unit period; the total loss determined based on each optimized weight value satisfies the condition of minimum total loss; the total loss is determined based on the sum of multiple losses; each loss represents the difference between the statistical total charging volume and the predicted total charging volume of the corresponding charging object; the statistical total charging volume represents the charging volume of the corresponding charging object within the unit statistical period obtained from statistics; the predicted total charging volume represents the charging volume of the corresponding charging object within the unit statistical period predicted based on each optimized weight value.

[0072] The target statistical period refers to a future statistical period, such as the next year. The optimized weight value refers to the value obtained after optimizing the weight values ​​corresponding to each unit period; the total loss based on each optimized weight value must satisfy the condition of minimum total loss. The total loss can be represented as the sum of multiple losses. Specifically, the loss refers to the difference between the actual total charging amount within the statistical period obtained from statistics and the predicted total charging amount within the statistical period based on multiple optimized weight values. In simpler terms, the smaller the total loss determined based on the optimized weight values ​​corresponding to each unit period, the more accurate the representation of each optimized weight value, allowing the charging quota predicted based on each optimized weight value to be closer to the actual charging amount needed by the charging target.

[0073] The condition for minimizing total loss can be set according to requirements. For example, the total loss condition could be: substituting each optimized weight value into the preset total loss model to minimize the output value of the total loss model. Or, the total loss condition could be: substituting each optimized weight value into the preset total loss model to make the output value of the total loss model less than or equal to the preset total loss threshold.

[0074] Specifically, when the historical charging volume of multiple charging objects within each unit time period in multiple historical statistical periods is obtained, the computer equipment can determine the total charging quota for each charging object within the target statistical period based on the historical charging volume, and determine the corresponding optimization weight value for each unit time period. The total charging quota refers to the total charging volume that should be allocated to charging objects within a statistical period; for example, the total charging quota could be the total charging volume to be allocated to electric vehicles within the next year.

[0075] In one embodiment, for each of the multiple charging objects, the computer device determines the total historical charging amount of the current charging object in each historical statistical period based on the historical charging amount of the current charging object in each unit time period of multiple historical statistical periods. The computer device averages the total historical charging amounts to obtain the average total historical charging amount, and then uses the average total historical charging amount as the total charging quota of the current charging object in the target statistical period.

[0076] In one embodiment, for each of the multiple charging objects, the computer device determines the total historical charging amount of the current charging object in each historical statistical period based on the historical charging amount of the current charging object in each unit time period of multiple historical statistical periods. The computer device sorts the historical total charging amounts in descending order to obtain a historical total charging amount sequence, and uses the median of the historical total charging amount sequence as the total charging quota of the current charging object in the target statistical period.

[0077] In one embodiment, the computer device determines the average historical charging amount of each charging object in each unit time period based on the acquired historical charging amounts, and determines the corresponding optimization weight parameters for each unit time period based on the average historical charging amount of each charging object in each unit time period.

[0078] In one embodiment, for each of the multiple charging objects, the computer device extracts the target historical charging amount of the current charging object within each unit time period in the multiple historical statistical periods from the acquired historical charging amounts. For each unit time period, the computer device sums the target historical charging amounts of the current charging object within the unit time period of the current type in each historical statistical period to obtain the total historical charging amount within the unit time period of the current type. The computer device divides the total historical charging amount by the total number of the multiple historical statistical periods to obtain the average historical charging amount of the current charging object within the unit time period of the current type. Further, when the average historical charging amount of the current charging object within various unit time periods is obtained, the computer device can normalize the average historical charging amount of the current charging object to obtain the initial weight value corresponding to each unit time period for the current charging object. For example, when the average historical charging amount of the current charging object is 300, 200, and 500, the initial weight values ​​of the current charging object obtained after normalizing the historical charging amounts are 0.3, 0.2, and 0.5. The sum of the initial weight values ​​is 1.

[0079] Since the computer device performs the above operations for each charging object, it obtains multiple initial weight values ​​for each object. Therefore, the computer device can comprehensively determine the final optimized weights for each unit time period based on these initial weight values. For example, the computer device can obtain a total loss model. For each charging object among multiple charging objects, the computer device can input the multiple initial weight values ​​corresponding to the current charging object into the total loss model, obtaining the total loss corresponding to the current charging object output by the total loss model. The computer device can compare the magnitudes of the total losses corresponding to each charging object, select the charging object with the minimum total loss, and use the multiple initial weight values ​​corresponding to the selected charging object as multiple optimized weight values. Here, the total loss model refers to the model that outputs the total loss, which is the total loss determined above based on the sum of multiple losses.

[0080] Step 206: Referring to multiple optimized weight values, allocate the total charging quota for each charging object within the target statistical period to obtain the unit charging quota for each charging object within each unit time period in the target statistical period.

[0081] Specifically, when the optimization weight value corresponding to each unit time period is obtained, for each of the multiple charging objects, the computer device can allocate the total charging quota of the current charging object according to the optimization weight value corresponding to each unit time period, so as to obtain the unit charging quota of the current charging object in each unit time period in the target statistical period.

[0082] For example, when the charging target is an electric vehicle, the unit time period is one month, the target statistical period is the next year, and the computer device obtains the optimization weight value corresponding to each month within the year, for each of the multiple optimization weight values, the computer device can multiply the current optimization weight value by the total charging quota corresponding to the current electric vehicle to obtain the unit charging quota of the current electric vehicle in the month corresponding to the current optimization weight value. For example, when the total charging quota is 100 and the optimization weight value corresponding to February in the next year is 0.3, the unit charging quota allocated to the current charging target in February of the next year is 0.3 * 100 = 30.

[0083] In one embodiment, reference Figure 3 The computer equipment can divide the total charging quota according to each optimization weight value to obtain the unit charging quota in each unit time period within the target statistical period. Figure 3 The diagram illustrates the division of total charging quotas in one embodiment. Here, t represents a unit time period; for example, t1 represents unit time period t1, t2 represents unit time period t2, etc.

[0084] In one embodiment, reference Figure 4 When the object of charging is an electric vehicle, and the unit charging quota for each month of the electric vehicle in the coming year has been determined, the electric vehicle can go to the charging station to charge in the corresponding month. Figure 4 A schematic diagram of a charging object being charged is shown in one embodiment.

[0085] In the aforementioned method for determining charging quotas, by acquiring the historical charging volume of multiple charging objects within each unit time period across multiple historical statistical periods, the total charging quota for each charging object within the target statistical period and the corresponding optimized weight value for each unit time period can be determined based on the historical charging volume. By determining the total charging quota and optimized weight values, the total charging quota can be allocated based on each optimized weight value, thereby obtaining the unit charging quota for each charging object within each unit time period of the target statistical period. Since the total charging quota is determined based on historical charging volume, the determined total charging quota conforms to the historical charging patterns of the charging objects, thus making the predicted total charging quota more accurate. This, in turn, reduces the waste of energy resources such as electricity based on the accurately determined total charging quota.

[0086] Furthermore, since loss refers to the difference between the total charging amount actually used within the statistical period and the total charging amount predicted based on multiple optimized weight values, when the total loss determined based on each optimized weight value satisfies the condition of minimizing the total loss, it indicates that the total charging amount predicted based on each optimized weight value within the statistical period is closer to the total charging amount actually needed by the charging object within the statistical period. Since the optimized weight values ​​correspond to the corresponding unit time period, when the total charging amount predicted based on each optimized weight value within the statistical period matches the actual usage, the unit charging amount predicted based on a single optimized weight value within a unit time period will also match the actual usage of the charging object within that unit time period. Therefore, the unit charging quota obtained by allocating the total charging quota through optimized weight values ​​will also be closer to the actual charging amount needed by the charging object within that unit time period, thus achieving the goal of accurately determining the unit charging quota. Because the unit charging quota for each unit time period is accurately determined, the arbitrary abuse of electricity and other energy resources by the charging object can be reduced, thereby saving electricity and other energy resources.

[0087] In one embodiment, each of the multiple charging objects is taken as a target charging object. The step of determining the total charging quota of the target charging object in the target statistical period includes: for each historical statistical period, the historical charging amount of the target charging object in each unit time period in the current historical statistical period is superimposed to obtain the historical total charging amount of the target charging object in the current historical statistical period; the historical total charging amount of the target charging object in each historical statistical period is averaged to obtain the historical total charging amount mean of the target charging object; the historical total charging amount variance is determined based on the historical total charging amount mean and the historical total charging amount of the target charging object in each historical statistical period; and the total charging quota to be allocated to the target charging object in the target statistical period under a preset confidence level is determined based on the historical total charging amount mean and the historical total charging amount variance of the target charging object.

[0088] Specifically, each of the multiple charging objects is taken as the target charging object. The following explanation uses charging object i as the target charging object, where i is a positive integer and less than or equal to the total number of charging objects. When it is necessary to determine the total charging quota for charging object i, the computer device can filter the target historical charging amount of charging object i within each unit time period in multiple historical statistical periods from the obtained historical charging amounts. For each historical statistical period, the computer device can sum up the target historical charging amounts of charging object i within each unit time period in the current historical statistical period to obtain the total historical charging amount of charging object i in the current historical statistical period.

[0089] Furthermore, the computer equipment determines the total quantity of multiple historical statistical periods and sums up the historical total charging amount of charging object i within each historical statistical period to obtain the summed historical total charging amount. The computer equipment divides the summed historical total charging amount by the total quantity of multiple historical statistical periods to obtain the average historical total charging amount of charging object i. Furthermore, the computer equipment determines the variance of the historical total charging amount based on the average historical total charging amount of charging object i and the historical total charging amount of charging object i within each historical statistical period, and determines the total charging quota to be allocated to charging object i within the target statistical period under a preset confidence level based on the average historical total charging amount and the variance of the historical total charging amount.

[0090] It is easy to understand that a computer device can simultaneously determine the total charging quota of multiple charging objects. In the process of simultaneously determining the total charging quota of multiple charging objects, the computer device can also determine the historical total charging amount of charging object i in multiple historical statistical periods.

[0091] In one embodiment, the computer device can determine the total charging quota corresponding to each charging object using the following formula:

[0092] Total charging quota:

[0093] in,

[0094]

[0095] {x ik |i=1,...,m;k=1,...,N}

[0096] Among them, {x ik x in |i=1,...,m;k=1,...,N} ik This represents the total historical charging amount of charging object i within the historical statistical period K. The historical average total charge value of charging object i; This represents the historical total charging volume variance of charging object i; the total charging quota can be an interval. This represents the lower limit of the total charging quota for charging object i. is the upper limit of the total charging quota for charging object i; i represents charging object i; m represents the total number of multiple charging objects; k represents the historical statistical period k; N represents the total number of multiple historical statistical periods; (1-α) is the confidence level.

[0097] In one embodiment, the computer device can determine the total historical charging amount of each charging object within the historical statistical period using the following formula:

[0098]

[0099] Among them, {x ijk |i=1,...,m;j=1,...,g;k=1,...,N}

[0100] Where, x ijk This represents the historical charging amount of charging object i within a unit time period j in the historical statistical period k; x ik This represents the total historical charging amount of charging object i within the historical statistical period k. i represents charging object i; m represents the total number of multiple charging objects; j represents the unit time period j; g represents the total number of different unit time periods; k represents the historical statistical period k; and N represents the total number of multiple historical statistical periods.

[0101] In the above embodiments, the total charging quota is determined by the mean and variance, which can make the determined total charging quota meet the variation law of the total charging amount, so that the determined total charging quota can better match the actual total charging amount of the charging object, and thus make the determined total charging quota more accurate.

[0102] In one embodiment, the step of determining the optimization weight value corresponding to each unit time period includes: obtaining a Gaussian mixture model; the Gaussian mixture model represents the combination of probability distributions of the charging object's charging amount in each unit time period, and includes Gaussian weight parameters corresponding to each unit time period; based on each historical charging amount, determining the average historical charging amount of each charging object in each unit time period, and adjusting the values ​​of the Gaussian weight parameters corresponding to each unit time period in the Gaussian mixture model according to the total loss model and the average historical charging amount, to obtain the optimization weight value corresponding to each unit time period.

[0103] A Gaussian mixture model (GMM) refers to the precise quantification of a phenomenon using a Gaussian probability density function (normal distribution curve). It decomposes a phenomenon into several models based on Gaussian probability density functions (normal distribution curves). For example, in this embodiment, the GMM can decompose the charging amount of a charging object within a unit time period into several models based on Gaussian probability density functions (normal distribution curves). The Gaussian probability density function characterizes the probability distribution of the charging amount used by the charging object within a unit time period. In other words, the GMM can characterize the combination of probability distributions of the charging amount used by the charging object within each unit time period. The GMM may include multiple Gaussian weight parameters, where the value of each Gaussian weight parameter represents the weight value corresponding to the corresponding Gaussian probability density function. The number of Gaussian probability density functions in the GMM can be consistent with the total number of classes within a unit time period, so that each Gaussian probability density function can correspond to a unit time period. Since each Gaussian probability density function can correspond to a unit time period, the Gaussian weight parameter corresponding to each Gaussian probability distribution also corresponds to a unit time period. For example, Gaussian weight parameter w1 corresponds to the Gaussian weight parameter in January; Gaussian weight parameter w2 corresponds to the Gaussian weight parameter in February. Accordingly, 0.3 in w1 = 0.3 is the value of the Gaussian weight parameter.

[0104] Since the Gaussian probability density function characterizes the probability distribution of the charging object's charge usage within a unit of time period, the Gaussian weight parameter corresponding to the Gaussian probability density function characterizes the weight of the probability distribution of the charging object's charge usage within a unit of time period. Therefore, computer equipment can determine the optimization weight value based on the value of the Gaussian weight parameter.

[0105] Specifically, when it is necessary to determine the optimization weight values, the computer device can acquire a pre-set Gaussian mixture model and determine the average historical charging amount of each charging object in each unit time period based on the acquired historical charging amounts. Further, the computer device acquires a total loss model and, based on the total loss model and the average historical charging amounts, adjusts the values ​​of the Gaussian weight parameters in the Gaussian mixture model corresponding to each unit time period until the total loss determined based on the values ​​of each Gaussian weight parameter is minimized. The final values ​​of each Gaussian weight parameter are then used as the optimization weight values.

[0106] In one embodiment, the computer device can determine the Gaussian mixture model and the total loss model using the following formula:

[0107]

[0108] Wherein, P(x i |μ,σ) represents a Gaussian mixture model; L represents the total loss model; μ j σ represents the average charging value of charging object i over a unit time period j; j φ(x) represents the variance of the amount of charge applied to object i within a unit time period j; ij |μ j ,σ j ) represents the Gaussian probability density. w j For the weight parameters, when w j lie in In the middle of the time, w j These are called Gaussian weight parameters; when w j lie in In the middle of the time, w j These are called loss weight parameters; therefore, w in j Let j be the Gaussian weight parameter corresponding to the unit time period j; w in j Let y be the loss weight parameter j corresponding to unit time period j. i When calculating the total charge of charging object i, and in the actual calculation of the total loss L, for y i The historical average total charging amount of charging object i can be substituted into it; The predicted total charging amount for charging object i is used in the actual calculation of the total loss L. x in ij The historical average charging amount of charging object i within each unit time period can be substituted; x ij The charging amount of charging object i within a unit time period j; j represents the unit time period j; g represents the total number of different unit time periods; m represents the total number of multiple charging objects; λ represents the bias parameter; ||w j|| represents the norm of the loss weight parameter j (j=1,...,g).

[0109] It is easy to understand that since both the Gaussian mixture model and the total loss model include weight parameters, the weight parameter is called the Gaussian weight parameter when it is in the Gaussian mixture model, and the weight parameter is called the loss weight parameter when it is in the total loss model. Therefore, the value of the Gaussian weight parameter can be input into the total loss model, and the total loss model will output the corresponding total loss.

[0110] In the above embodiments, by constructing a Gaussian mixture model, the optimized weight values ​​can be determined based on the Gaussian weight parameters of the Gaussian mixture model; the values ​​of the Gaussian weight parameters are adjusted by the total loss model, so that the optimized weight values ​​determined based on the values ​​of the Gaussian weight parameters are more accurate.

[0111] In one embodiment, each of the multiple charging objects is taken as a target charging object, and the step of determining the average historical charging amount of the target charging object in each unit time period includes: extracting the target historical charging amount of the target charging object in each unit time period in the multiple historical statistical periods from the acquired historical charging amounts; and determining the average historical charging amount of the target charging object in each unit time period based on the extracted target historical charging amounts.

[0112] Specifically, each of the multiple charging objects is designated as a target charging object. The following explanation uses charging object i as the target charging object, where i is a positive integer and is less than or equal to the total number of charging objects. When it is necessary to determine the average historical charging amount of charging object i in each unit time period, the computer device can extract the target historical charging amount used by charging object i in each unit time period of the multiple historical statistical periods from the acquired historical charging amounts. For example, the computer device can extract the historical charging amount used by charging object i in each month over N years and use the obtained historical charging amount as the target historical charging amount. Further, the computer device determines the average historical charging amount of charging object i in each unit time period based on the extracted target historical charging amounts.

[0113] In one embodiment, the unit statistical period includes multiple unit time periods; each unit time period among the multiple unit time periods is taken as the target unit time period, and the step of determining the historical average charging amount of the target charging object in the target unit time period includes: superimposing the target historical charging amount of the target charging object in the target unit time period in each historical statistical period to obtain the unit historical total charging amount in the target unit time period; dividing the unit historical total charging amount by the total number of multiple historical statistical periods to obtain the historical average charging amount of the target charging object in the target unit time period.

[0114] Specifically, the target charging object can be charging object i. When the target historical charging amount of charging object i in each unit time period of multiple historical statistical periods is obtained, for unit time period j, the computer device can sum up the target historical charging amounts of charging object i in each unit time period j of the various historical statistical periods to obtain the total historical charging amount of charging object i in unit time period j. Here, the unit statistical period includes multiple unit time periods; each unit time period among the multiple unit time periods is taken as the target unit time period, and the target unit time period can be unit time period j.

[0115] For example, a computer device can sum up the target historical charging amount used by charging object i in the i-th month over N years to obtain the unit historical total charging amount of charging object i in the i-th month. Further, the computer device determines the total number of multiple historical statistical periods and divides the unit historical total charging amount of charging object i in a unit period j by the total number of multiple historical statistical periods to obtain the average historical charging amount of charging object i in a unit period j.

[0116] In one embodiment, the computer device can determine the historical average charging amount using the following formula:

[0117]

[0118] {x ijk |i=1,...,m;j=1,...,g;k=1,...,N}

[0119] in, The value represents the average historical charging amount of charging object i within a unit time period j; i represents charging object i; m represents the total number of multiple charging objects; j represents the unit time period j; g represents the total number of categories across multiple unit time periods; k represents the historical statistical time period k; N represents the total number across multiple historical statistical time periods; x ijk This represents the historical charging amount of charging object i within a unit time period j in the historical statistical time period k.

[0120] In the above embodiments, by determining the historical average charging amount, the subsequent optimization weight value can be determined based on the historical average charging amount, and the unit charging quota can be determined based on the determined optimization weight value.

[0121] In one embodiment, based on the total loss model and the historical average charging amount of each charging object within each unit time period, the values ​​of the Gaussian weight parameters in the Gaussian mixture model corresponding to each unit time period are adjusted to obtain multiple optimized weight values. This includes: obtaining an initialized model parameter vector group; the model parameter vector group includes multiple model parameter vectors; the model parameter vectors are vectors determined based on the values ​​of various types of model parameters in the Gaussian mixture model; determining the target model parameter vector group for the first round based on the initialized model parameter vector group and the historical average charging amount of each charging object within each unit time period; and in the current round starting from the second round after the first round, based on... The target model parameter vector set from the previous round determines the initial model parameter vector set for the current round. Based on the total loss model, the initial model parameter vector set for the current round is adjusted to obtain the target model parameter vector set for the current round. The next round is taken as the new current round, and the process of determining the initial model parameter vector set for the current round based on the target model parameter vector set from the previous round is repeated in the current rounds starting from the second round after the first round. This process continues until a preset stopping condition is met, resulting in the target model parameter vector set for the final round. The Gaussian weight parameter values ​​corresponding to each unit time period in the target model parameter vector set for the final round are used as the optimization weight values.

[0122] A Gaussian mixture model can include multiple parameters. For example, referring to the formula for a Gaussian mixture model mentioned above, it can include Gaussian weight parameters w, mean parameters μ, and standard deviation parameters σ. The model parameter vector can be a vector containing the values ​​of the Gaussian weight parameters w, the mean parameters μ, and the standard deviation parameters σ. Since each time period can correspond to one Gaussian weight parameter, one mean parameter, and one standard deviation parameter, each time period can correspond to a separate model parameter vector. By combining the model parameter vectors corresponding to each time period, a model parameter vector group can be obtained. For example, the model parameter vector group can be {w} j ,μ j ,σ j}(j=1,...,g). That is, the model parameter vector group includes g model parameter vectors.

[0123] Specifically, the computer device can adjust the values ​​of each Gaussian weight parameter in multiple rounds using the total loss model, so that the adjusted values ​​of each Gaussian weight parameter minimize the total loss. When multiple rounds of adjustment of the values ​​of each Gaussian weight parameter are required, the computer device can obtain an initialized model parameter vector, where the initialized model parameter vector set can be a vector set with random values. The computer device can determine the target model parameter vector set for the first round based on the initialized model parameter vector set and the historical average charging amount of each charging object in each unit time period. From the second round onwards after the first round, the computer device can determine the initial model parameter vector set for the current round based on the target model parameter vector set of the previous round. Then, based on the total loss model, the initial model parameter vector set for the current round is adjusted to obtain the target model parameter vector set for the current round. For example, a computer device can adjust the values ​​of the Gaussian weight parameters corresponding to each unit time period in the initial model parameter vector group of the current round through the total loss model, obtain the adjusted values ​​of the Gaussian weight parameters corresponding to each unit time period, and use the model parameter vector group of the current round, which includes the adjusted values ​​of the Gaussian weight parameters corresponding to each unit time period, as the target model parameter vector group of the current round.

[0124] Furthermore, the process proceeds to the next round, which is then taken as the new current round. Returning to the current round starting from the second round after the first, the steps of determining the initial model parameter vector group for the current round based on the target model parameter vector group from the previous round are repeated until a preset stopping condition is met. This yields the target model parameter vector group for the final round. The Gaussian weight parameters corresponding to each unit time period in the final round's target model vector group are then used as the optimized weight values. Thus, the optimized weight values ​​for each unit time period are obtained. The preset stopping condition can be freely set according to requirements. For example, it can be determined that the preset stopping condition is met when a preset number of rounds is reached. Alternatively, it can be determined that the preset stopping condition is met when the difference between the target model parameter vector groups from multiple rounds is less than or equal to a preset difference threshold, at which point iteration can stop. Another example is when it is determined that the total loss obtained based on the target model parameter vector group of the current round reaches its minimum value, at which point iteration can stop.

[0125] In one embodiment, reference Figure 5 The computer device can adjust the model parameter vector group multiple times until a stopping condition is met. When iteration stops, the computer device uses the values ​​of multiple Gaussian weight parameters in the target model parameter vector group of the final iteration as the optimized weight values. Figure 5 A schematic diagram of the iterative adjustment of parameter values ​​in one embodiment is shown.

[0126] In the above embodiments, by adjusting the model parameter vector in multiple rounds, the values ​​of multiple Gaussian weight parameters in the adjusted target model parameter vector group can satisfy the condition of minimum total loss.

[0127] In one embodiment, the target model parameter vector set for the first round is determined based on the initialized model parameter vector set and the historical average charging amount of each charging object within each unit time period. This includes: obtaining a probability density parameter model corresponding to the probability density parameters; obtaining multiple initial probability density parameter values ​​through the probability density parameter model and based on the initialized model parameter vector set and the historical average charging amount of each charging object within each unit time period; obtaining a Gaussian weight parameter model corresponding to the Gaussian weight parameters, a mean parameter model corresponding to the mean parameter, and a standard deviation parameter model corresponding to the standard deviation parameter; and obtaining the Gaussian weight parameter model, the mean parameter model corresponding to the mean parameter, and the ... mean parameter; and obtaining the Gaussian weight parameter model, the mean parameter model corresponding to the mean parameter, and the standard deviation parameter model corresponding to the mean parameter. The mean parameter model and standard deviation parameter model are used. Based on the values ​​of multiple initial probability density parameters and the average historical charging amount, the values ​​of Gaussian weight parameters, mean parameters, and standard deviation parameters corresponding to each unit time period in the first round are obtained. The values ​​of Gaussian weight parameters corresponding to each unit time period in the first round are adjusted based on the total loss model to obtain the adjusted values ​​of multiple Gaussian weight parameters in the first round. The target model parameter vector group for the first round is obtained by combining the values ​​of multiple standard deviation parameters, multiple mean parameters, and adjusted Gaussian weight parameters in the first round.

[0128] Specifically, when it is necessary to determine the target model parameter vector set for the first round, the computer device can acquire the probability density parameter model and input the initialized model parameter vector set and the historical average charging amount of each charging object in each unit time period into the probability density model. The probability density model then outputs the values ​​of multiple initial probability density parameters. For example, the probability density model can output... in, Here, represents the initial probability density parameter value, where i is the charging object i, and j is the unit time period j. That is, the output value is the initial probability density parameter value. Let be the value of the initial probability density parameter corresponding to the i-th charging object in unit time period j.

[0129] Furthermore, the computer equipment can acquire Gaussian weight parameter model, mean parameter model, and standard deviation parameter model. Through these models, and based on the values ​​of multiple initial probability density parameters and the historical average charging amount of each charging object within each unit time period, the computer obtains the values ​​of multiple Gaussian weight parameters output by the Gaussian weight parameter model for the first round, i.e., the values ​​of the Gaussian weight parameters corresponding to each unit time period; the computer also obtains the values ​​of multiple mean parameters output by the mean parameter model for the first round, i.e., the values ​​of the mean parameters corresponding to each unit time period; and the computer further obtains the values ​​of multiple standard deviation parameters output by the standard deviation parameter model for the first round, i.e., the values ​​of the standard deviation parameters corresponding to each unit time period.

[0130] Furthermore, the computer device combines the values ​​of the Gaussian weight parameters, mean parameters, and standard deviation parameters corresponding to the same time unit to obtain an initial model parameter vector. By combining the initial model parameter vectors corresponding to each time unit, the initial model parameter vector set for the first round is obtained. When the initial model parameter vector set for the first round is obtained, the computer device adjusts the values ​​of multiple Gaussian weight parameters for the first round using the total loss model, obtaining adjusted values ​​for the multiple Gaussian weight parameters for the first round. These adjusted values ​​then replace the values ​​of multiple Gaussian weight parameters in the initial model parameter vector set for the first round, thus obtaining the target model parameter vector set for the first round. in, The values ​​of the multiple Gaussian weight parameters for the first round after adjustment; j represents the unit time period j.

[0131] In one embodiment, the values ​​of multiple Gaussian weight parameters for the first round are adjusted based on the total loss model to obtain adjusted values ​​of the multiple Gaussian weight parameters for the first round. This includes: the computer device determining the first total loss for the first round using the total loss model and based on the values ​​of the multiple Gaussian weight parameters for the first round and the average historical total charging amount corresponding to each charging object; determining the value of the loss weight parameter output for the first round corresponding to the loss weight parameter in the total loss model; determining the second total loss for the first round using the total loss model and based on the value of the loss weight parameter for the first round; when the first total loss for the first round is less than the second total loss for the first round, the values ​​of the multiple Gaussian weight parameters for the first round are used as the adjusted values ​​of the multiple Gaussian weight parameters for the first round; when the first total loss for the first round is greater than the second total loss for the first round, the values ​​of the multiple loss weight parameters for the first round are used as the adjusted values ​​of the multiple Gaussian weight parameters for the first round.

[0132] In the above embodiments, by obtaining the Gaussian weight parameter model, the mean parameter model, and the standard deviation parameter model, the values ​​of multiple Gaussian weight parameters, multiple mean parameters, and multiple standard deviation parameters for the first round can be output based on these models. By obtaining the values ​​of these parameters for the first round, an initial model parameter vector group for the first round can be obtained. To minimize the total loss obtained based on the values ​​of the multiple Gaussian weight parameters in the model parameter vector group, the values ​​of the multiple Gaussian weight parameters for the first round can be adjusted using a total loss model, so that the total loss obtained based on the adjusted values ​​is less than the total loss obtained based on the unadjusted values. This achieves the goal of reducing the total loss.

[0133] In one embodiment, the Gaussian mixture model includes Gaussian weight parameters, mean parameters, and standard deviation parameters. The Gaussian weight parameters are determined based on the probability density parameters. Determining the initial model parameter vector set for the current round based on the target model parameter vector set from the previous round includes: determining the values ​​of multiple probability density parameters for the current round using a probability density parameter model corresponding to the probability density parameters, and based on the target model parameter vector set from the previous round and the historical average charging amount of each charging object within each unit time period; obtaining the initial model parameter vector set for the current round using a Gaussian weight parameter model corresponding to the Gaussian weight parameters, a mean parameter model corresponding to the mean parameters, and a standard deviation parameter model corresponding to the standard deviation parameters, and based on the values ​​of the multiple probability density parameters for the current round and the historical average charging amount of each charging object within each unit time period.

[0134] Specifically, the computer device can input the target model parameter vector group from the previous round and the historical average charging amount of each charging object in each unit time period into the probability density parameter model, and output the values ​​of multiple probability density parameters for the current round through the probability density parameter model. Further, the computer device uses a Gaussian weight parameter model, a mean parameter model, and a standard deviation parameter model, and based on the values ​​of multiple probability density parameters for the current round and the historical average charging amount, to obtain the values ​​of multiple Gaussian weight parameters output by the Gaussian weight parameter model for the current round, that is, the values ​​of the Gaussian weight parameters corresponding to each unit time period for the current round; the values ​​of multiple mean parameters output by the mean parameter model for the current round, that is, the values ​​of the mean parameters corresponding to each unit time period for the current round; and the values ​​of multiple standard deviation parameters output by the standard deviation parameter model for the current round, that is, the values ​​of the standard deviation parameters corresponding to each unit time period for the current round. The computer device combines the values ​​of multiple Gaussian weight parameters, multiple mean parameters, and multiple standard deviation parameters from the current round to obtain the initial model parameter vector set for the current round. Where k represents the kth round, and j represents the unit time period j.

[0135] In one embodiment, the computer device can utilize a probability density parameter model corresponding to the probability density parameter. i = 1,...,m; j = 1,...,g, determine the values ​​of multiple probability density parameters for the current round. Where x ij This represents the amount of charge received by object i within a unit time period j, for x. ij When calculating the probability density parameter, the average historical charging amount of charging object i within a unit time period j can be substituted. j ,μ j ,σ j}(j=1,...,g) is the target model parameter vector group from the previous round.

[0136] In one embodiment, an initial model parameter vector set for the current round is obtained by using a Gaussian weight parameter model corresponding to the Gaussian weight parameters, a mean parameter model corresponding to the mean parameter, and a standard deviation parameter model corresponding to the standard deviation parameter, and based on the values ​​of multiple probability density parameters for the current round and the historical average charging amount of each charging object in each unit time period. This includes: determining the value of the Gaussian weight parameter corresponding to each unit time period of the current round using the Gaussian weight parameter model corresponding to the Gaussian weight parameters and based on the values ​​of multiple probability density parameters for the current round; and determining the value of the Gaussian weight parameter corresponding to each unit time period of the current round using the mean parameter model corresponding to the mean parameter and based on the values ​​of multiple probability density parameters for the current round. The values ​​of probability density parameters and the historical average charging amount of each charging object in each unit time period are used to determine the value of the mean parameter corresponding to each unit time period in the current round. Using the standard deviation parameter model corresponding to the standard deviation parameter, and based on the values ​​of multiple probability density parameters in the current round, the historical average charging amount of each charging object in each unit time period, and the value of the mean parameter corresponding to each unit time period in the current round, the value of the standard deviation parameter corresponding to each unit time period in the current round is determined. Finally, by combining the values ​​of multiple Gaussian weight parameters, multiple mean parameters, and multiple standard deviation parameters in the current round, the initial model parameter vector set for the current round is obtained.

[0137] Specifically, the Gaussian weighting computer device first inputs the values ​​of multiple probability density parameters for the current round into the Gaussian weighting parameter model, which then outputs the values ​​of multiple Gaussian weighting parameters for the current round. Further, the computer device inputs the values ​​of multiple probability density parameters for the current round and the historical average charging amount of each charging object within each unit time period into the mean parameter model, which then outputs the values ​​of multiple mean parameters for the current round. The computer device then inputs the values ​​of multiple probability density parameters for the current round, the historical average charging amount of each charging object within each unit time period, and the values ​​of multiple mean parameters for the current round into the standard deviation parameter model, which then outputs the values ​​of multiple standard deviation parameters for the current round. Finally, the computer device combines the values ​​of multiple Gaussian weighting parameters, multiple mean parameters, and multiple standard deviation parameters for the current round to obtain the initial model parameter vector set for the current round.

[0138] In one embodiment, the Gaussian weight parameter model is used to output the Gaussian weight parameters that characterize the weights of the Gaussian distribution density function in the Gaussian mixture model; the mean parameter model is used to determine the value of the mean parameter in the Gaussian mixture model; the mean parameter characterizes the mean amount of charging of the charging object within a unit time period; the standard deviation parameter model is used to determine the value of the standard deviation parameter in the Gaussian mixture model; the standard deviation parameter characterizes the variance of charging of the charging object within a unit time period.

[0139] In one embodiment, the computer device can utilize the Gaussian weight parameter model corresponding to the Gaussian weight parameters. j = 1, ..., g, determines the values ​​of multiple Gaussian weight parameters for the current round. γ ij The value of the probability density parameter represents the current charging object; in actual calculations, the value of the probability density parameter for the current cycle can be substituted into it. i represents the charging object i; j represents the unit time period j.

[0140] In one embodiment, the computer device can use a mean parameter model corresponding to the mean parameter. j = 1, ..., g, determine the values ​​of multiple mean parameters for the current round. Among them, γ ij This represents the value of the probability density parameter, which can be substituted into the value of the probability density parameter for the current round during actual calculations; x ij This represents the amount of charge received by charging object i within a unit time period j; when actually calculating the value of the average parameter, the average amount of charge received by charging object i within a unit time period j can be substituted.

[0141] In one embodiment, the computer device can utilize a standard deviation parameter model corresponding to the standard deviation parameter. j = 1, ..., g, determine the values ​​of multiple standard deviation parameters for the current round. γ ij This represents the value of the probability density parameter, which can be substituted into the value of the probability density parameter for the current round during actual calculation; x ij This represents the amount of charge received by charging object i within a unit time period j. When calculating the standard deviation parameter, the average amount of charge received by charging object i within a unit time period j can be substituted into it; μ j This represents the value of the mean parameter, which can be substituted into the mean parameter value of the current round during actual calculation.

[0142] In the above embodiments, by obtaining the Gaussian weight parameter model, mean parameter model, and standard deviation parameter model, the initial model parameter vector group for the current round can be output based on the obtained Gaussian weight parameter model, mean parameter model, and standard deviation parameter model, so that the target model parameter vector group for the current round can be obtained based on the initial model parameter vector group for the current round.

[0143] In one embodiment, adjusting the initial model parameter vector set for the current round based on the total loss model to obtain the target model parameter vector set for the current round includes: adjusting the values ​​of the Gaussian weight parameters corresponding to each unit time period in the initial model parameter vector set for the current round using the total loss model to obtain the adjusted values ​​of the Gaussian weight parameters corresponding to each unit time period in the current round; updating the initial model parameter vector set for the current round based on the adjusted values ​​of the Gaussian weight parameters corresponding to each unit time period in the current round to obtain the target model parameter vector set for the current round.

[0144] Specifically, when generating the initial model parameter vector set for the current round, the computer device adjusts the value of each Gaussian weight parameter in each initial model parameter vector using the total loss model to obtain adjusted values ​​for multiple Gaussian weight parameters in the current round. Furthermore, for each of the various time units, the computer device replaces the original Gaussian weight parameter value with the adjusted value for that time unit. This updates the initial model parameter vector set for the current round based on the adjusted Gaussian weight parameter values ​​for each time unit in the current round, thus obtaining the target model parameter vector set for the current round.

[0145] In one embodiment, the Gaussian weight parameters corresponding to each unit time period in the initial model parameter vector group of the current round are adjusted using the total loss model to obtain the adjusted Gaussian weight parameter values ​​corresponding to each unit time period of the current round. This includes: determining the first total loss of the current round using the total loss model and based on the Gaussian weight parameter values ​​corresponding to each unit time period of the current round and the historical average total charging amount of each charging object; determining the loss weight parameter values ​​corresponding to each unit time period of the current round output by the loss weight parameter model corresponding to the loss weight parameter in the total loss model; determining the second total loss of the current round using the total loss model and based on the loss weight parameter values ​​corresponding to each unit time period of the previous round; and determining the adjusted Gaussian weight parameter values ​​corresponding to each unit time period of the current round based on the first total loss and the second total loss of the current round.

[0146] Specifically, the computer device can input the values ​​of multiple Gaussian weight parameters for the current round and the historical average total charging value for each charging object into a pre-set total loss model, and output the first total loss for the current round through the total loss model. Further, the total loss model may include loss weight parameters, which represent the proportion of the charging amount of the charging object within a unit time period to the total charging amount within the statistical time period. The computer device can obtain the values ​​of the loss weight parameters for the current round output by the loss weight parameter model, and input the values ​​of multiple loss weight parameters for the current round into the total loss model, and output the second total loss for the current round through the total loss model. Once the first and second total losses for the current round are obtained, the computer device can determine the adjusted values ​​of the Gaussian weight parameters based on the first and second total losses for the current round.

[0147] In this embodiment, since the smaller the total loss output by the total loss model, the closer the predicted total charging amount generated based on the values ​​of multiple Gaussian weight parameters or multiple loss weight parameters is to the statistical charging amount of the charging object in a unit statistical period, the more accurate the values ​​of the Gaussian weight parameters or loss weight parameters are. Therefore, the computer device can determine the adjusted Gaussian weight parameter values ​​based on the first and second total losses of the current round in the opposite direction of reducing the total loss output by the total loss model, so that the total loss determined based on the adjusted Gaussian weight parameter values ​​can be less than the total loss determined based on the unadjusted Gaussian weight parameter values.

[0148] In one embodiment, determining the adjusted Gaussian weight parameter value for each unit time period of the current round based on the first total loss and the second total loss of the current round includes: when the first total loss of the current round is less than the second total loss of the current round, using the Gaussian weight parameter value for each unit time period of the current round as the adjusted Gaussian weight parameter value for each unit time period of the current round; when the first total loss of the current round is greater than the second total loss of the current round, using the loss weight parameter value for each unit time period of the current round as the adjusted Gaussian weight parameter value for each unit time period of the current round.

[0149] Specifically, the computer device compares the first total loss with the second total loss. When the first total loss is less than the second total loss, the computer device directly uses the values ​​of the multiple Gaussian weight parameters from the current round as the adjusted values ​​of the multiple Gaussian weight parameters. When the first total loss is greater than the second total loss, the computer device uses the values ​​of the multiple loss weight parameters from the current round as the adjusted values ​​of the multiple Gaussian weight parameters.

[0150] In one embodiment, reference Figure 6 The computer device can input the values ​​of multiple Gaussian weight parameters and multiple loss weight parameters of the current round into the total loss model, output the first total loss and the second total loss through the total loss model, and use the values ​​of multiple Gaussian weight parameters or multiple loss weight parameters as the adjusted values ​​of multiple Gaussian weight parameters of the current round according to the relationship between the first total loss and the second total loss. Figure 6 A schematic diagram illustrating the adjustment of the Gaussian weight parameter value in one embodiment is shown.

[0151] In the above embodiments, by using the value of the weight parameter corresponding to the smaller total loss as the value of the adjusted Gaussian weight parameter, the goal of gradually reducing the total loss in each round can be achieved.

[0152] In one embodiment, reference Figure 7, Figure 7 A schematic diagram of the overall process for determining multiple optimization weight values ​​in one embodiment is shown:

[0153] The algorithm flow of this scheme is mainly divided into the following stages: data input stage, annual charging capacity range estimation stage, federated hybrid distribution and total loss model construction stage, parameter model construction stage, federated hybrid distribution and total loss joint decision mechanism parameter construction stage, and parameter model interactive iterative calculation stage.

[0154] S702, Data Input Stage. Input the historical charging data (also known as historical charging amount) of the electric vehicle (also known as the charging object) for the past N years (also known as the historical statistical period) for each of the 12 months (also known as each unit period). ijk |i=1,...,m;j=1,...,12;k=1,...,N}. The charging data for each electric vehicle over 12 months of the year is then accumulated annually to obtain the annual charging data for each electric vehicle (also known as the historical total charging data) {x ik |i=1,...,m;k=1,...,N}, where, The charging data for each electric vehicle over the 12 months of the year is then averaged monthly to obtain the average monthly charging data for each electric vehicle (also known as the historical average charging data). in, Where i represents electric vehicle; k represents year; and j represents month.

[0155] S704, Annual Charging Volume Range Estimation Stage. Input the annual charging volume data (also known as historical total charging volume) for each electric vehicle in S702 {x ik |i=1,...,m;k=1,...,N}. Calculate the average annual electricity consumption for each user (also known as the average historical total charging amount). Sum of variance (also known as the variance of total historical electricity volume) At a confidence level of 1-a, the confidence interval for the annual charging amount of each electric vehicle (also known as the total charging quota within the target statistical period) is calculated.

[0156] S706, Gaussian Mixture Model and Total Loss Model Construction Phase. Constructing a federated Gaussian Mixture Model: And constructing the total loss model:

[0157] S708, Parametric Model Construction Stage. For the probability density parameters in the Gaussian mixture model, the probability density parameter model is as follows: i=1,...,m; j=1,...,12. x ijFor the charging amount of charging object i within a unit time period j, when actually calculating the value of the probability density parameter, the average historical charging amount of charging object i within a unit time period j can be substituted. j ,μ j ,σ j}(j=1,...,g) is the model parameter vector group. When actually calculating the value of the probability density parameter, the target model parameter vector group obtained in the previous round can be input.

[0158] For the mean parameter in the Gaussian mixture model, the mean parameter model is as follows: j = 1,...,12, x ij The charging amount of charging object i within a unit time period j can be substituted into the historical average charging amount of charging object i within a unit time period j when calculating the total loss parameter.

[0159] For the standard deviation parameter in the Gaussian mixture model, the standard deviation parameter model is as follows: j = 1, ..., 12. x ij The charging amount of charging object i within a unit time period j can be substituted into the historical average charging amount of charging object i within a unit time period j when calculating the total loss parameter.

[0160] For the Gaussian weight parameters in the Gaussian mixture model, the Gaussian weight parameter model is as follows: j = 1, ..., 12.

[0161] For the total loss parameter in the total loss model, the total loss parameter calculation model is as follows: y i The total charging amount for charging object i can be calculated by substituting the historical average total charging amount of charging object i into the total loss parameter. ij The charging amount of charging object i within a unit time period j can be substituted into the historical average charging amount of charging object i within a unit time period j when calculating the total loss parameter.

[0162] S710, Parameter Construction Stage for the Joint Decision-Making Mechanism of Gaussian Mixture Model and Total Loss Model. Input the parameter estimation models of the Gaussian mixture model and the total loss model from S706 to construct the parameter model of the joint decision-making mechanism:

[0163]

[0164] S712, Parametric Model Interactive Iterative Calculation Stage. First, the model parameter vector set is initialized to obtain the initial model parameter vector set. Input the average monthly charging data (also known as the historical average charging data) for each electric vehicle in S702. and into the above probability density parameter model, and output the values of a plurality of initial probability density parameters through the probability density parameter model the calculated are input into the mean parameter model μ j , the standard deviation parameter model and the Gaussian weight parameter model w j to obtain the initial model parameter vector group of the first round: substituting the values of a plurality of Gaussian parameter weights in the initial model parameter vector group of the first round into the total loss model, and calculating the first total loss under the first round wherein, when calculating the total loss, for y in the total loss model i , the historical total charging mean of the charging object i is substituted. For x in the total loss model ij , the historical charging mean of the charging object i in the unit time period j is substituted.

[0165] The computer device inputs the parameter λ into the loss weight parameter model to obtain the weight parameter value output by the weight parameter model when calculating the weight parameter value, for y in the weight parameter model i , the historical total charging mean of the charging object i is substituted. For x in the weight parameter model ij , the historical charging mean of the charging object i in the unit time period j is substituted. Further, the computer device substitutes the weight parameter value w′ j into the total loss model, and outputs the second total loss under the first round through the total loss model

[0166] if L′<L1, the optimized weight parameter value is: otherwise, the optimized weight parameter value is: recording the optimized weight parameter value of the first round as and updating the initial model parameter vector group of the first round to: thereby obtaining the target model parameter vector group of the first round. By parity of reasoning, the target model parameter vector group of the final round is obtained, and the values of a plurality of Gaussian weight parameters in the target model parameter vector group of the final round are all used as final optimized weight values.

[0167] In one embodiment, each of the multiple charging objects is taken as a target charging object. The total charging quota of the target charging object in the target statistical period includes an upper limit and a lower limit of the total charging quota. The step of determining the unit charging quota of the target charging object in each unit time period in the target statistical period includes: for each of the multiple optimization weight values, multiplying the current optimization weight value by the upper limit of the total charging quota of the target charging object to obtain the upper limit of the unit charging quota of the unit time period corresponding to the current optimization weight value; multiplying the current optimization weight value by the lower limit of the total charging quota of the target charging object to obtain the lower limit of the unit charging quota of the unit time period corresponding to the current optimization weight value; and combining the upper limit and the lower limit of the unit charging quota to obtain the unit charging quota of the target charging object in the unit time period corresponding to the current optimization weight value.

[0168] Specifically, when the optimized weight value corresponding to each unit time period is obtained, for each of the multiple optimized weight values, the current optimized weight value is multiplied by the total charging quota limit of charging object i to obtain the unit charging quota limit for the unit time period corresponding to the current optimized weight value. For example, when the optimized weight value for February is obtained, the computer device can multiply the optimized weight value for February by the total charging quota limit of charging object i to obtain the unit charging quota limit for charging object i in February. Here, charging object i is the target charging object mentioned above.

[0169] Furthermore, the computer device can multiply the current optimized weight value by the total lower limit of the charging target i to obtain the lower limit of the unit charging quota for a unit time period corresponding to the current optimized weight value. For example, the computer device can multiply the optimized weight value for February by the lower limit of the total charging quota for charging target i to obtain the lower limit of the unit charging quota for charging target i in February. The computer device combines the upper limit and lower limit of the unit charging quota to obtain the unit charging quota for charging target i within a unit time period corresponding to the current optimized weight value. For example, the computer device can combine the upper limit and lower limit of the unit charging quota for February to obtain the unit charging quota for charging target i in February of the next year in interval form.

[0170] In one embodiment, the computer device can determine the unit charging quota using the following formula:

[0171]

[0172] in, To optimize the weight values; C i This is the lower limit of the total charging quota; U iis the upper limit of the total charging quota; i is the charging object i; m is the total number of multiple charging objects; j is the unit time period j; g is the total number of multiple unit time periods.

[0173] In the above embodiments, since the determined unit charging quota is a range, the charging amount can be flexibly selected based on the range when allocating charging amount to the charging object based on the unit charging quota, making the allocation of power more flexible.

[0174] In one embodiment, the unit statistical period includes multiple unit time periods, with each of the multiple charging objects being designated as a target charging object, and each of the multiple unit time periods being designated as a target unit time period. The method further includes: determining the actual charging amount of the target charging object within the target unit time period of the target statistical period; when the actual charging amount is less than the unit charging quota of the target charging object within the target unit time period of the target statistical period, storing the difference between the unit charging quota of the target charging object within the target unit time period of the target statistical period and the actual charging amount in the cloud; and when the actual charging amount is greater than the unit charging quota of the target charging object within the target unit time period of the target statistical period, requesting additional charging quota from the cloud.

[0175] Specifically, the computer device can obtain the actual charging amount of charging object i within a unit time period j of the target statistical period, and determine the difference between the unit charging quota and the actual charging amount of charging object i within the unit time period j of the target statistical period. When the actual charging amount is less than the unit charging quota, the excess charging quota can be uploaded to the cloud for storage. When the actual charging amount is less than the unit charging quota, additional charging quota can be requested from the cloud. The additional charging quota can be the charging quota stored in the cloud for other charging objects besides charging object i, or it can be a pre-set charging quota. Here, charging object i is the aforementioned target charging object. Unit time period j is the aforementioned target unit time period.

[0176] In one embodiment, reference Figure 8 The charging quota determination method of this application can be applied to a charging quota determination system, which may include endpoints, edges, pipes, and the cloud. Here, "endpoint" can be a charging object; "edge" can be a switch, edge server, etc., wherein the edge server is used to determine the unit charging quota for each charging object within each unit time period of the target statistical period, and to collect the actual charging amount of the charging object within the corresponding unit time period of the target period; "pipe" refers to reporting the data generated and acquired by the edge server to the cloud; "cloud" can be a cloud server used to store the data reported by the edge server, wherein the data reported by the edge server may include additional charging quotas allocated to the charging object. Figure 8 A schematic diagram of a charging quota determination system in one embodiment is shown.

[0177] In one embodiment, reference Figure 9 When the edge server determines that the actual charging amount of charging object i within a unit time period j of the target statistical period is less than the unit charging quota of charging object i within a unit time period j of the target statistical period, the edge server can determine the first difference between the actual charging amount and the unit charging quota, and generate a charging quota storage request based on the first difference, and send the charging quota storage request to the cloud so that the cloud stores the charging quota of the first difference into the account corresponding to charging object i.

[0178] When the edge server determines that the actual charging amount of charging object i within a unit time period j of the target statistical period is greater than the unit charging quota of charging object i within the unit time period j of the target statistical period, the edge server can determine a second difference between the actual charging amount and the unit charging quota, and generate a charging quota application request based on the second difference, and send the charging quota application request to the cloud. When the cloud receives the charging quota application request, the cloud can first determine whether the account corresponding to charging object i has stored charging quota. If it has stored charging quota, it indicates that charging object i has remaining charging quota in some unit time periods before the unit time period j of the target statistical period. At this time, the cloud will distribute the charging quota of the second difference in the account corresponding to charging object i to charging object i. If the charging quota stored in the account corresponding to charging object i is less than the charging quota of the second difference, the cloud will first distribute all the charging quota in the account corresponding to charging object i to charging object i, and obtain charging quota from the accounts of other charging objects besides charging object i, and distribute the obtained charging quota to charging object i. Alternatively, the cloud can obtain charging quotas from the public account and distribute them to charging object i. Conversely, if the account corresponding to charging object i does not have any stored charging quotas, the cloud can directly obtain charging quotas from the accounts of other charging objects or the public account and distribute them to charging object i.

[0179] When the cloud obtains charging quotas from the accounts or public accounts of other charging objects and distributes the obtained charging quotas to charging object i, the cloud can transfer a corresponding amount of target type resources from the account corresponding to charging object i to the accounts or public accounts of other charging objects. The target type of resources and electricity resources are not the same type of resource; there is a resource conversion rate between the target type of resources and electricity resources. Therefore, the cloud can determine the amount of target type resources that needs to be transferred from the account of charging object i based on the resource conversion rate. Figure 9 A schematic diagram of charging quota storage and application is shown in one embodiment.

[0180] In the above embodiments, by reasonably allocating the unit charging quota for each charging object, and through the charging quota storage mechanism and the charging quota paid application mechanism, the charging object can be urged to use electricity reasonably, thereby reducing the waste of energy resources such as electricity.

[0181] In one specific embodiment, reference is made to Figure 10 , Figure 10 The diagram illustrates a flowchart of a method for determining charging quotas in one embodiment, where charging object i is any one of a plurality of charging objects; and unit time period j is any one of a plurality of unit time periods.

[0182] S1002, The computer device acquires the historical charging amount of multiple charging objects in each unit time period in multiple historical statistical periods.

[0183] S1004. For each historical statistical period in multiple historical statistical periods, the computer device sums up the historical charging amount of charging object i in each unit time period in the current historical statistical period to obtain the total historical charging amount of charging object i in the current historical statistical period.

[0184] S1006, the computer device averages the historical total charging amount of charging object i in each historical statistical period to obtain the historical total charging amount mean of charging object i; based on the historical total charging amount mean and the historical total charging amount of charging object i in each historical statistical period, the historical total power variance is determined.

[0185] S1008, the computer equipment determines the total charging quota that should be allocated to the charging object i within the target statistical period under the preset confidence level, based on the historical average total charging amount and historical total power variance of the charging object i.

[0186] S1010, The computer device acquires a Gaussian mixture model; the Gaussian mixture model represents the combination of probability distributions of the charging object's charging amount in each unit time period, and includes the Gaussian weight parameters corresponding to each unit time period.

[0187] S1012, the computer device extracts the target historical charging amount of the charging object i in each unit time period in multiple historical statistical periods from the acquired historical charging amounts; and determines the average historical charging amount of the charging object i in each unit time period based on the extracted target historical charging amounts.

[0188] S1014, The computer device obtains the initialized model parameter vector group; the model parameter vector group includes multiple model parameter vectors; the model parameter vectors are vectors determined based on the values ​​of various types of model parameters in the Gaussian mixture model; the target model parameter vector group for the first round is determined according to the initialized model parameter vector group and the historical average charging amount of each charging object in each unit time period.

[0189] S1016, In the current round starting from the second round after the first round, the computer device determines the initial model parameter vector group for the current round based on the target model parameter vector group of the previous round, and adjusts the initial model parameter vector group for the current round based on the total loss model to obtain the target model parameter vector group for the current round.

[0190] S1018, the computer device takes the next round as the new current round and returns to the current round starting from the second round after the first round. It continues to execute the step of determining the initial model parameter vector group of the current round based on the target model parameter vector group of the previous round until the preset stopping condition is met, and obtains the target model parameter vector group of the final round. The value of the Gaussian weight parameter corresponding to each unit time period in the target model parameter vector group of the final round is used as the optimization weight value.

[0191] S1020, for each of the multiple optimization weight values, the computer device multiplies the current optimization weight value by the total charging quota limit of charging object i to obtain the unit charging quota limit for the unit time period corresponding to the current optimization weight value.

[0192] S1022, the computer device multiplies the current optimization weight value by the lower limit of the total charging quota of charging object i to obtain the lower limit of the unit charging quota for the unit time period corresponding to the current optimization weight value; combining the upper limit of the unit charging quota and the lower limit of the unit charging quota, the unit charging quota of charging object i in the unit time period corresponding to the current optimization weight value is obtained.

[0193] S1024, The computer equipment determines the actual charging amount of the charging object i within a unit time period j of the target statistical period.

[0194] S1026, when the actual charging amount is less than the unit charging quota of charging object i in the unit time period j of the target statistical period, the computer device stores the difference between the unit charging quota of charging object i in the unit time period j of the target statistical period and the actual charging amount in the cloud; when the actual charging amount is greater than the unit charging quota of charging object i in the unit time period j of the target statistical period, the computer device requests an additional charging quota from the cloud; the additional charging is configured as the charging quota stored in the cloud for other charging objects besides charging object i.

[0195] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily...

[0196] Execution is performed sequentially as indicated by the arrows. Unless otherwise expressly stated herein, there is no strict order restriction on the execution of these steps, and they may be executed in other orders. Moreover, at least some steps in the flowcharts involved in the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but may be executed at different times. The execution order of these steps or stages is not necessarily sequential, but may be performed alternately or in turn with other steps or at least some of the steps or stages in other steps.

[0197] This application also provides an application scenario in which the above-described method for determining charging quotas is applied. Specifically, the method for determining charging quotas is applied in this application scenario as follows:

[0198] Currently, logistics parks can utilize multiple driverless electric vehicles for cargo transportation. When unified management of these vehicles is required, computer equipment can retrieve the historical charging data for each vehicle, generate a total charging quota for each vehicle, and create multiple optimized weight values. Furthermore, based on these optimized weight values ​​and the total charging quota, the computer equipment can determine the monthly charging quota for each vehicle over the next year. Once the monthly charging quota for each vehicle is obtained, it can charge at designated charging stations in the corresponding month. This achieves unified management of the vehicle network, reducing indiscriminate and uncontrolled charging and conserving resources such as electricity. It's easy to understand that driverless electric vehicles can also be manned electric vehicles.

[0199] The above application scenarios are merely illustrative. It is understood that the application of determining the charging quota provided in the various embodiments of this application is not limited to the above scenarios.

[0200] Based on the same inventive concept, this application also provides a charging quota determination apparatus for implementing the charging quota determination method described above. The solution provided by this apparatus is similar to the implementation described in the above method; therefore, the specific limitations in one or more embodiments of the charging quota determination apparatus provided below can be found in the limitations of the charging quota determination method described above, and will not be repeated here.

[0201] In one embodiment, such as Figure 11 A charging quota determination device 1100 is provided, comprising: a historical charging amount determination module 1102, an information determination module 1104, and a quota allocation module 1106, wherein:

[0202] The historical charging amount determination module 1102 is used to obtain the historical charging amount of multiple charging objects in each unit time period in multiple historical statistical time periods.

[0203] The information determination module 1104 is used to determine the total charging quota for each charging object within the target statistical period and the corresponding optimized weight value for each unit period based on the historical charging volume; the total loss determined based on each optimized weight value satisfies the condition of minimum total loss; the total loss is determined based on the sum of multiple losses; each loss represents the difference between the statistical total charging volume and the predicted total charging volume of the corresponding charging object; the statistical total charging volume represents the charging volume of the corresponding charging object within the unit statistical period obtained from statistics; the predicted total charging volume represents the charging volume of the corresponding charging object within the unit statistical period obtained based on each optimized weight value.

[0204] The quota allocation module 1106 is used to allocate the total charging quota of each charging object in the target statistical period by referring to the optimization weight value corresponding to each unit time period, so as to obtain the unit charging quota of each charging object in each unit time period in the target statistical period.

[0205] In one embodiment, the information determination module 1104 further includes a total charging quota determination module 1141, which is used to take each of the multiple charging objects as a target charging object, and for each of the multiple historical statistical periods, to sum up the historical charging amount of the target charging object in each unit time period in the current historical statistical period to obtain the historical total charging amount of the target charging object in the current historical statistical period; to average the historical total charging amount of the target charging object in each historical statistical period to obtain the average historical total charging amount of the target charging object; to determine the historical total power variance based on the historical total charging average and the historical total charging amount of the target charging object in each historical statistical period; and to determine the total charging quota to be allocated to the target charging object in the target statistical period under a preset confidence level based on the historical total charging average and the historical total power variance of the target charging object.

[0206] In one embodiment, the information determination module 1104 further includes an optimization weight value determination module 1142, used to obtain a Gaussian mixture model; the Gaussian mixture model represents the combination of probability distributions of the charging object's charging amount in each unit time period, and includes Gaussian weight parameters corresponding to each unit time period; based on each historical charging amount, the average historical charging amount of each charging object in each unit time period is determined, and based on the total loss model and the average historical charging amount, the values ​​of the Gaussian weight parameters in the Gaussian mixture model corresponding to each unit time period are adjusted to obtain the optimization weight values ​​corresponding to each unit time period.

[0207] In one embodiment, the optimization weight value determination module 1142 is further configured to take each of the multiple charging objects as the target charging object, extract the target historical charging amount of the target charging object in each unit time period in multiple historical statistical periods from the acquired historical charging amounts, and determine the average historical charging amount of the target charging object in each unit time period based on the extracted target historical charging amounts.

[0208] In one embodiment, the unit statistical time period includes multiple unit time periods. The optimization weight value determination module 1142 is further used to take each unit time period among the multiple unit time periods as the target unit time period, and to superimpose the target historical charging amount of the target charging object in the target unit time period in each historical statistical time period to obtain the unit historical total charging amount in the target unit time period; and to divide the unit historical total charging amount by the total number of multiple historical statistical time periods to obtain the average historical charging amount of the target charging object in the target unit time period.

[0209] In one embodiment, the optimization weight value determination module 1142 is further configured to obtain an initialized model parameter vector group; the model parameter vector group includes multiple model parameter vectors; the model parameter vectors are vectors determined based on the values ​​of multiple types of model parameters in a Gaussian mixture model; the target model parameter vector group for the first round is determined based on the initialized model parameter vector group and the historical average charging amount of each charging object in each unit time period; in the current round starting from the second round after the first round, the initial model parameter vector group for the current round is determined based on the target model parameter vector group of the previous round, and the initial model parameter vector group for the current round is adjusted based on the total loss model to obtain the target model parameter vector group for the current round; the next round is taken as the new current round, and the process of determining the initial model parameter vector group for the current round based on the target model parameter vector group of the previous round is returned to the current round starting from the second round after the first round and the process continues until a preset stopping condition is met, thus obtaining the target model parameter vector group for the final round, and the values ​​of the Gaussian weight parameters corresponding to each unit time period in the target model parameter vector group for the final round are all used as optimization weight values.

[0210] In one embodiment, the Gaussian mixture model includes Gaussian weight parameters, mean parameters, and standard deviation parameters. The Gaussian weight parameters are determined based on the probability density parameters. The optimization weight value determination module 1142 is further used to obtain a probability density parameter model corresponding to the probability density parameters. Through the probability density parameter model, and based on the initialized model parameter vector group and the historical average charging amount of each charging object in each unit time period, multiple initial probability density parameter values ​​are obtained. The Gaussian weight parameter model corresponding to the Gaussian weight parameters, the mean parameter model corresponding to the mean parameters, and the standard deviation parameter model corresponding to the standard deviation parameters are obtained. The Gaussian weight parameters are determined based on the Gaussian weight parameters, mean parameter model corresponding to the mean parameters, and standard deviation parameter model corresponding to the standard deviation parameters. The model consists of a number model, a mean parameter model, and a standard deviation parameter model. Based on the values ​​of multiple initial probability density parameters and the average historical charging amount, the values ​​of the Gaussian weight parameters, the mean parameter, and the standard deviation parameter for each unit time period in the first round are obtained. The values ​​of the Gaussian weight parameters for each unit time period in the first round are adjusted based on the total loss model, resulting in the adjusted values ​​of the multiple Gaussian weight parameters for the first round. The target model parameter vector set for the first round is obtained by combining the values ​​of the multiple standard deviation parameters, the multiple mean parameters, and the adjusted values ​​of the multiple Gaussian weight parameters from the first round.

[0211] In one embodiment, the Gaussian mixture model includes Gaussian weight parameters, mean parameters, and standard deviation parameters. The Gaussian weight parameters are determined based on the probability density parameters. The optimization weight value determination module 1142 is further used to determine the values ​​of multiple probability density parameters for the current round using the probability density parameter model corresponding to the probability density parameters, and based on the target model parameter vector group of the previous round and the historical average charging amount of each charging object in each unit time period. The initial model parameter vector group for the current round is obtained using the Gaussian weight parameter model corresponding to the Gaussian weight parameters, the mean parameter model corresponding to the mean parameters, and the standard deviation parameter model corresponding to the standard deviation parameters, and based on the values ​​of multiple probability density parameters for the current round and the historical average charging amount of each charging object in each unit time period.

[0212] In one embodiment, the optimization weight value determination module 1142 is further configured to: determine the value of the Gaussian weight parameter corresponding to each unit time period of the current round using a Gaussian weight parameter model corresponding to the Gaussian weight parameter and based on the values ​​of multiple probability density parameters of the current round; determine the value of the mean parameter corresponding to each unit time period of the current round using a mean parameter model corresponding to the mean parameter and based on the values ​​of multiple probability density parameters of the current round and the historical average charging amount of each charging object in each unit time period; determine the value of the standard deviation parameter corresponding to each unit time period of the current round using a standard deviation parameter model corresponding to the standard deviation parameter and based on the values ​​of multiple probability density parameters of the current round, the historical average charging amount of each charging object in each unit time period, and the value of the mean parameter corresponding to each unit time period of the current round; and obtain the initial model parameter vector group of the current round by combining the values ​​of multiple Gaussian weight parameters, multiple mean parameters, and multiple standard deviation parameters of the current round.

[0213] In one embodiment, the optimization weight value determination module 1142 is further configured to adjust the values ​​of the Gaussian weight parameters corresponding to each unit time period in the initial model parameter vector group of the current round using the total loss model, so as to obtain the adjusted values ​​of the Gaussian weight parameters corresponding to each unit time period in the current round; and update the initial model parameter vector group of the current round according to the adjusted values ​​of the Gaussian weight parameters corresponding to each unit time period in the current round, so as to obtain the target model parameter vector group of the current round.

[0214] In one embodiment, the optimization weight value determination module 1142 is further configured to: determine the first total loss of the current round using the total loss model and based on the values ​​of the Gaussian weight parameters corresponding to each unit time period of the current round and the historical average total charging amount of each charging object; determine the values ​​of the loss weight parameters output by the loss weight parameter model corresponding to each unit time period of the current round, and determine the second total loss of the current round using the total loss model and based on the values ​​of the loss weight parameters corresponding to each unit time period of the current round; and determine the adjusted values ​​of the Gaussian weight parameters corresponding to each unit time period of the current round based on the first total loss and the second total loss of the current round.

[0215] In one embodiment, the optimization weight value determination module 1142 is further configured to, when the first total loss of the current round is less than the second total loss of the current round, use the value of the Gaussian weight parameter corresponding to each unit time period of the current round as the adjusted value of the Gaussian weight parameter corresponding to each unit time period of the current round; and when the first total loss of the current round is greater than the second total loss of the current round, use the value of the loss weight parameter corresponding to each unit time period of the current round as the adjusted value of the Gaussian weight parameter corresponding to each unit time period of the current round.

[0216] In one embodiment, each of the multiple charging objects is taken as a target charging object. The total charging quota of the target charging object in the target statistical period includes a total charging quota upper limit and a total charging quota lower limit. The quota allocation module 1106 is further used to multiply the current optimization weight value with the total charging quota upper limit of the target charging object for each of the multiple optimization weight values ​​to obtain the unit charging quota upper limit of the unit time period corresponding to the current optimization weight value; multiply the current optimization weight value with the total charging quota lower limit of the target charging object to obtain the unit charging quota lower limit of the unit time period corresponding to the current optimization weight value; and combine the unit charging quota upper limit and the unit charging quota lower limit to obtain the unit charging quota of the target charging object in the unit time period corresponding to the current optimization weight value.

[0217] In one embodiment, the unit statistical period includes multiple unit time periods, with each of the multiple charging objects being designated as a target charging object, and each of the multiple unit time periods being designated as a target unit time period. The charging quota determination device 1100 further includes a cloud processing module, used to determine the actual charging amount of the target charging object within the target unit time period of the target statistical period; when the actual charging amount is less than the unit charging quota of the target charging object within the target unit time period of the target statistical period, the difference between the unit charging quota of the target charging object within the target unit time period of the target statistical period and the actual charging amount is stored in the cloud; when the actual charging amount is greater than the unit charging quota of the target charging object within the target unit time period of the target statistical period, an additional charging quota is requested from the cloud.

[0218] Each module in the aforementioned charging quota determination device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0219] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 12 As shown, this computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores data for determining charging quotas. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When executed by the processor, the computer program implements a method for determining charging quotas.

[0220] Those skilled in the art will understand that Figure 12 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0221] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0222] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0223] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the steps in the above method embodiments.

[0224] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data shall comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0225] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0226] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0227] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for determining charging quotas, characterized in that, The method includes: Get the historical charging amount of multiple charging objects within each unit time period in multiple historical statistical periods; Based on the historical charging volume, the total charging quota for each charging object within the target statistical period and the corresponding optimized weight value for each unit period are determined. The total loss determined based on the optimized weight values ​​satisfies the condition of minimum total loss. The total loss is determined based on the sum of multiple losses. Each loss represents the difference between the statistical total charging volume and the predicted total charging volume of the corresponding charging object. The statistical total charging volume represents the charging volume of the corresponding charging object within the unit statistical period, obtained from statistics. The predicted total charging volume represents the charging volume of the corresponding charging object within the unit statistical period, determined based on the optimized weight values. Referring to the optimization weight value corresponding to each of the unit time periods, the total charging quota of each of the charging objects in the target statistical period is allocated to obtain the unit charging quota of each of the unit time periods in the target statistical period; The steps for determining the optimization weight value corresponding to each of the aforementioned time periods include: Obtain a Gaussian mixture model; the Gaussian mixture model represents the combination of probability distributions of the charging object's charging amount in each unit time period, and includes Gaussian weight parameters corresponding to each unit time period; Based on the historical charging amounts, the average historical charging amount of each charging object in each unit time period is determined. Then, based on the total loss model and the average historical charging amounts, the values ​​of the Gaussian weight parameters in the Gaussian mixture model corresponding to each unit time period are adjusted to obtain the optimized weight values ​​corresponding to each unit time period.

2. The method according to claim 1, characterized in that, Taking each of the plurality of charging objects as a target charging object, the steps for determining the total charging quota of the target charging object within the target statistical period include: For each of the multiple historical statistical periods, the historical charging amount of the target charging object in each unit time period in the current historical statistical period is summed to obtain the total historical charging amount of the target charging object in the current historical statistical period. The average of the historical total charging amount of the target charging object in each of the historical statistical periods is calculated to obtain the average historical total charging amount of the target charging object. The historical total charge variance is determined based on the historical average total charge amount and the historical total charge amount of the target charging object in each historical statistical period. Based on the historical average total charging amount and the historical variance of the target charging object, determine the total charging quota that should be allocated to the target charging object within the target statistical period under a preset confidence level.

3. The method according to claim 1, characterized in that, The charging object refers to an object powered by electrical resources.

4. The method according to claim 1, characterized in that, Taking each of the plurality of charging objects as a target charging object, the step of determining the historical average charging amount of the target charging object in each of the unit time periods includes: From the acquired historical charging amounts, extract the target historical charging amount of the target charging object within each unit time period in the plurality of historical statistical time periods; Based on the extracted historical charging amounts of each target, the average historical charging amount of the target charging object in each unit time period is determined.

5. The method according to claim 4, characterized in that, The statistical time period for each unit includes multiple time periods; Taking each of the multiple time periods as a target time period, the step of determining the historical average charging amount of the target charging object in the target time period includes: The target historical charging amount of the target charging object in each of the historical statistical periods is summed to obtain the total historical charging amount per unit time period in the target unit time period. Divide the total historical charging amount per unit by the total number of the multiple historical statistical periods to obtain the average historical charging amount of the target charging object within the target unit time period.

6. The method according to claim 1, characterized in that, The step involves adjusting the values ​​of the Gaussian weight parameters in the Gaussian mixture model corresponding to each of the given time periods based on the total loss model and the average historical charging amount, to obtain optimized weight values ​​for each of the given time periods, including: Obtain an initialized set of model parameter vectors; the set of model parameter vectors includes multiple model parameter vectors; the model parameter vectors are vectors determined based on the values ​​of various types of model parameters in the Gaussian mixture model; Based on the initialized model parameter vector set and the historical average charging amount of each charging object in each unit time period, the target model parameter vector set for the first round is determined; From the second round after the first round, the initial model parameter vector group of the current round is determined based on the target model parameter vector group of the previous round. The initial model parameter vector group of the current round is adjusted based on the total loss model to obtain the target model parameter vector group of the current round. The next round is taken as the new current round, and the process of determining the initial model parameter vector group of the current round based on the target model parameter vector group of the previous round is returned to the current round starting from the second round after the first round. This process continues until a preset stopping condition is met, and the target model parameter vector group of the final round is obtained. The value of the Gaussian weight parameter corresponding to each unit time period in the target model parameter vector group of the final round is used as the optimization weight value.

7. The method according to claim 6, characterized in that, The Gaussian mixture model includes Gaussian weight parameters, mean parameters, and standard deviation parameters. The Gaussian weight parameters are determined based on probability density parameters. Determining the target model parameter vector set for the first round based on the initialized model parameter vector set and the historical average charging amount of each charging object within each unit time period includes: Obtain the probability density parameter model corresponding to the probability density parameter. Through the probability density parameter model, and based on the initialized model parameter vector group and the historical average charging amount of each charging object in each unit time period, obtain the values ​​of multiple initial probability density parameters. Obtain the Gaussian weight parameter model corresponding to the Gaussian weight parameter, the mean parameter model corresponding to the mean parameter, and the standard deviation parameter model corresponding to the standard deviation parameter; The Gaussian weight parameter model, the mean parameter model, and the standard deviation parameter model are used respectively, and based on the values ​​of the multiple initial probability density parameters and the average historical charging amount, the values ​​of the Gaussian weight parameter, the mean parameter, and the standard deviation parameter for each unit time period in the first round are obtained. Based on the total loss model, the values ​​of the Gaussian weight parameters corresponding to each unit time period in the first round are adjusted to obtain the adjusted values ​​of the multiple Gaussian weight parameters in the first round. By combining the values ​​of multiple standard deviation parameters, multiple mean parameters, and adjusted Gaussian weight parameters from the first round, the target model parameter vector group for the first round is obtained.

8. The method according to claim 6, characterized in that, The Gaussian mixture model includes Gaussian weight parameters, mean parameters, and standard deviation parameters, wherein the Gaussian weight parameters are determined based on the probability density parameters; the step of determining the initial model parameter vector set for the current round based on the target model parameter vector set from the previous round includes: The values ​​of multiple probability density parameters for the current round are determined by using the probability density parameter model corresponding to the probability density parameters, and based on the target model parameter vector group of the previous round and the historical average charging amount of each charging object in each unit time period. The initial model parameter vector group for the current round is obtained by using the Gaussian weight parameter model corresponding to the Gaussian weight parameter, the mean parameter model corresponding to the mean parameter, and the standard deviation parameter model corresponding to the standard deviation parameter, and based on the values ​​of multiple probability density parameters for the current round and the historical average charging amount of each charging object in each unit time period.

9. The method according to claim 8, characterized in that, The process involves obtaining an initial model parameter vector set for the current round using a Gaussian weight parameter model corresponding to the Gaussian weight parameters, a mean parameter model corresponding to the mean parameter, and a standard deviation parameter model corresponding to the standard deviation parameter, and based on the values ​​of multiple probability density parameters for the current round and the historical average charging amount of each charging object within each unit time period. This includes: By using the Gaussian weight parameter model corresponding to the Gaussian weight parameters, and based on the values ​​of multiple probability density parameters of the current round, the values ​​of the Gaussian weight parameters corresponding to each unit time period of the current round are determined. The mean parameter model corresponding to the mean parameter is used, and the values ​​of multiple probability density parameters of the current round and the historical average charging amount of each charging object in each unit time period are used to determine the value of the mean parameter corresponding to each unit time period of the current round. The standard deviation parameter is determined by using the standard deviation parameter model corresponding to the standard deviation parameter, and based on the values ​​of multiple probability density parameters of the current round, the historical average charging amount of each charging object in each unit time period, and the value of the mean parameter corresponding to each unit time period of the current round. By combining the values ​​of multiple Gaussian weight parameters, multiple mean parameters, and multiple standard deviation parameters for the current round, the initial model parameter vector set for the current round is obtained.

10. The method according to claim 6, characterized in that, The step of adjusting the initial model parameter vector set for the current round based on the total loss model to obtain the target model parameter vector set for the current round includes: Using the total loss model, the values ​​of the Gaussian weight parameters corresponding to each unit time period in the initial model parameter vector group of the current round are adjusted to obtain the adjusted values ​​of the Gaussian weight parameters corresponding to each unit time period in the current round. Based on the adjusted Gaussian weight parameter values ​​for each unit time period in the current round, the initial model parameter vector group for the current round is updated to obtain the target model parameter vector group for the current round.

11. The method according to claim 10, characterized in that, The step of adjusting the values ​​of the Gaussian weight parameters corresponding to each unit time period in the initial model parameter vector group of the current round using the total loss model to obtain the adjusted values ​​of the Gaussian weight parameters corresponding to each unit time period in the current round includes: The first total loss for the current round is determined by the total loss model and based on the values ​​of the Gaussian weight parameters corresponding to each unit time period in the current round and the average historical total charging amount of each charging object. Determine the values ​​of the loss weight parameters corresponding to each unit time period in the current round, as output by the loss weight parameter model in the total loss model. The second total loss for the current round is determined using the total loss model and based on the values ​​of the loss weight parameters corresponding to each unit time period in the current round. Based on the first total loss and the second total loss of the current round, determine the value of the Gaussian weight parameter corresponding to each unit time period of the current round after adjustment.

12. A device for determining charging quotas, characterized in that, The device includes: The historical charging amount determination module is used to obtain the historical charging amount of multiple charging objects in each unit time period in multiple historical statistical time periods; An information determination module is used to determine the total charging quota for each charging object within a target statistical period based on the historical charging amounts; obtain a Gaussian mixture model; the Gaussian mixture model represents the combination of probability distributions of charging amounts used by the charging object in each unit period, and includes Gaussian weight parameters corresponding to each unit period; determine the average historical charging amount of each charging object in each unit period based on the historical charging amounts, and adjust the values ​​of the Gaussian weight parameters in the Gaussian mixture model corresponding to each unit period based on the total loss model and the average historical charging amounts, to obtain optimized weight values ​​corresponding to each unit period; the total loss determined based on the optimized weight values ​​satisfies the condition of minimum total loss; the total loss is determined based on the sum of multiple losses; each loss represents the difference between the statistical total charging amount and the predicted total charging amount of the corresponding charging object; the statistical total charging amount represents the charging amount of the corresponding charging object within a unit statistical period obtained from statistics; the predicted total charging amount represents the charging amount of the corresponding charging object within a unit statistical period determined based on the optimized weight values. The quota allocation module is used to allocate the total charging quota of each charging object in the target statistical period by referring to the optimization weight value corresponding to each of the unit time periods, so as to obtain the unit charging quota of each charging object in each of the unit time periods in the target statistical period.

13. The apparatus according to claim 12, characterized in that, The information determination module further includes a total charging quota determination module, which is used to take each of the plurality of charging objects as a target charging object; for each of the plurality of historical statistical periods, the historical charging amount of the target charging object in each unit time period in the current historical statistical period is superimposed to obtain the historical total charging amount of the target charging object in the current historical statistical period; and the historical total charging amount of the target charging object in each of the historical statistical periods is averaged to obtain the average historical total charging amount of the target charging object. Based on the historical average total charging amount and the historical total charging amount of the target charging object in each historical statistical period, the historical total power variance is determined; based on the historical average total charging amount and the historical total power variance of the target charging object, the total charging quota to be allocated to the target charging object in the target statistical period under the preset confidence level is determined.

14. The apparatus according to claim 12, characterized in that, The information determination module further includes an optimization weight value determination module, which is used to take each of the plurality of charging objects as a target charging object; extract the target historical charging amount of the target charging object in each unit time period in the plurality of historical statistical periods from the acquired historical charging amounts; and determine the average historical charging amount of the target charging object in each unit time period based on the extracted target historical charging amounts.

15. The apparatus according to claim 14, characterized in that, The unit statistical time period includes multiple unit time periods; the optimization weight value determination module is also used to take each of the multiple unit time periods as the target unit time period; The target historical charging amount of the target charging object in each of the historical statistical periods is summed to obtain the total historical charging amount per unit time period in the target unit time period. Divide the total historical charging amount per unit by the total number of the multiple historical statistical periods to obtain the average historical charging amount of the target charging object within the target unit time period.

16. The apparatus according to claim 12, characterized in that, The information determination module further includes an optimization weight value determination module, used to obtain an initialized model parameter vector group; the model parameter vector group includes multiple model parameter vectors; the model parameter vectors are vectors determined based on the values ​​of multiple types of model parameters in the Gaussian mixture model; the target model parameter vector group for the first round is determined according to the initialized model parameter vector group and the historical average charging amount of each charging object in each unit time period; Starting from the second round after the first round, the initial model parameter vector group for the current round is determined based on the target model parameter vector group of the previous round. The initial model parameter vector group for the current round is adjusted based on the total loss model to obtain the target model parameter vector group for the current round. The next round is taken as the new current round, and the process of determining the initial model parameter vector group for the current round based on the target model parameter vector group of the previous round is returned and continued until a preset stopping condition is met, thus obtaining the target model parameter vector group for the final round. The values ​​of the Gaussian weight parameters corresponding to each unit time period in the target model parameter vector group of the final round are used as the optimized weight values.

17. The apparatus according to claim 16, characterized in that, The Gaussian mixture model includes Gaussian weight parameters, mean parameters, and standard deviation parameters. The Gaussian weight parameters are determined based on probability density parameters. The optimization weight value determination module is further used to obtain a probability density parameter model corresponding to the probability density parameters. Through the probability density parameter model, and based on the initialized model parameter vector group and the historical average charging amount of each charging object in each unit time period, multiple initial probability density parameter values ​​are obtained. The module also obtains the Gaussian weight parameter model corresponding to the Gaussian weight parameters, the mean parameter model corresponding to the mean parameters, and the standard deviation parameter model corresponding to the standard deviation parameters. The module then uses the Gaussian weight parameter model, mean parameter model corresponding to the mean parameters, and standard deviation parameter model corresponding to the standard deviation parameters. The mean parameter model and the standard deviation parameter model are used to obtain the values ​​of the Gaussian weight parameter, the mean parameter, and the standard deviation parameter for each unit time period in the first round, based on the values ​​of the multiple initial probability density parameters and the average historical charging amount. The values ​​of the Gaussian weight parameters for each unit time period in the first round are adjusted based on the total loss model to obtain the adjusted values ​​of the multiple Gaussian weight parameters for the first round. The target model parameter vector group for the first round is obtained by combining the values ​​of the multiple standard deviation parameters, the multiple mean parameters, and the adjusted values ​​of the multiple Gaussian weight parameters for the first round.

18. The apparatus according to claim 16, characterized in that, The Gaussian mixture model includes Gaussian weight parameters, mean parameters, and standard deviation parameters. The Gaussian weight parameters are determined based on probability density parameters. The optimization weight value determination module is further used to determine the values ​​of multiple probability density parameters for the current round using a probability density parameter model corresponding to the probability density parameters, and based on the target model parameter vector group of the previous round and the historical average charging amount of each charging object in each unit time period. The module also obtains the initial model parameter vector group for the current round using the Gaussian weight parameter model corresponding to the Gaussian weight parameters, the mean parameter model corresponding to the mean parameters, and the standard deviation parameter model corresponding to the standard deviation parameters, and based on the values ​​of multiple probability density parameters for the current round and the historical average charging amount of each charging object in each unit time period.

19. The apparatus according to claim 18, characterized in that, The optimization weight value determination module is further configured to: determine the value of the Gaussian weight parameter corresponding to each unit time period of the current round using a Gaussian weight parameter model corresponding to the Gaussian weight parameter and based on the values ​​of multiple probability density parameters of the current round; determine the value of the mean parameter corresponding to each unit time period of the current round using a mean parameter model corresponding to the mean parameter and based on the values ​​of multiple probability density parameters of the current round and the historical average charging amount of each charging object in each unit time period; determine the value of the standard deviation parameter corresponding to each unit time period of the current round using a standard deviation parameter model corresponding to the standard deviation parameter and based on the values ​​of multiple probability density parameters of the current round, the historical average charging amount of each charging object in each unit time period, and the value of the mean parameter corresponding to each unit time period of the current round; and obtain the initial model parameter vector group of the current round by combining the values ​​of multiple Gaussian weight parameters, multiple mean parameters, and multiple standard deviation parameters of the current round.

20. The apparatus according to claim 16, characterized in that, The optimization weight value determination module is further configured to adjust the values ​​of the Gaussian weight parameters corresponding to each unit time period in the initial model parameter vector group of the current round using the total loss model, to obtain the adjusted values ​​of the Gaussian weight parameters corresponding to each unit time period in the current round; and update the initial model parameter vector group of the current round according to the adjusted values ​​of the Gaussian weight parameters corresponding to each unit time period in the current round, to obtain the target model parameter vector group of the current round.

21. The apparatus according to claim 20, characterized in that, The optimization weight value determination module is further configured to: determine the first total loss of the current round using the total loss model and based on the values ​​of the Gaussian weight parameters corresponding to each unit time period of the current round and the historical average total charging amount corresponding to each charging object; determine the values ​​of the loss weight parameters output by the loss weight parameter model corresponding to the loss weight parameters in the total loss model and corresponding to each unit time period of the current round; determine the second total loss of the current round using the total loss model and based on the values ​​of the loss weight parameters corresponding to each unit time period of the current round; and determine the adjusted values ​​of the Gaussian weight parameters corresponding to each unit time period of the current round based on the first total loss and the second total loss of the current round.

22. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 11.

23. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 11.

24. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 11.

Images