Multi-subject data value marketization distribution method and system based on vehicle pile network scene

Abstract

The invention relates to the technical field of multi-subject data value marketization distribution based on a vehicle pile network scene, in particular to a multi-subject data value marketization distribution method and system based on the vehicle pile network scene. Determining the distributable data effect value, carrying out multi-subject model contribution degree calculation, and evaluating the feature contribution degree of each subject in the data model; comprehensively adjusting and balancing multi-main-body feature contribution degrees; estimating a multi-subject data value according to the feature contribution degree; and determining a distribution proportion according to the contribution degree of each main body and other adjustment factors, and distributing the service effect value which can be distributed to each main body according to the distribution proportion. According to the method, the multi-dimensional value of the data can be fully mined and utilized, and meanwhile, the fair, transparent and efficient distribution of the data element effect value by multiple participants is realized in combination with a multi-host value co-creation framework.

Description

Technical Field

[0001] The present invention relates to the technical field of market-based allocation of multi-agent data value in a vehicle-pile network scenario, and in particular to a method and system for market-based allocation of multi-agent data value based on a vehicle-pile network scenario. Background Art

[0002] In the existing technology, the methods for market-oriented distribution of data element value mainly include distribution according to work, distribution according to production factors, profit distribution, and distribution of benefits through experience negotiation. However, these methods fail to fully take into account the special nature of data elements. Data elements are different from traditional physical elements. They have the characteristics of non-physicality, reproducibility, non-competitiveness, weak exclusivity, diversity, processability, timeliness, and value volatility. These characteristics will cause difficulties in the use and distribution of data, making it difficult to fully explore and utilize the multi-dimensional value of data.

[0003] In addition, the existing data value market allocation method also lacks application practice in the multi-subject value co-creation model, and cannot meet the data demander's requirements for fair, transparent and efficient distribution of data value. Therefore, a new data value market allocation method is needed that can fully consider the characteristics of data elements and the multi-subject interactive value co-creation mechanism, so that data value can be fairly, transparently and efficiently distributed among multiple subjects. Summary of the invention

[0004] In view of the above problems existing in the prior art, the present invention is proposed.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for market-based allocation of multi-agent data value based on a vehicle-pile network scenario, including: clarifying the allocable primary data effectiveness value, calculating the contribution of a multi-agent model, and evaluating the feature contribution of each subject in the data model;

[0006] Comprehensive adjustment and balance of multi-subject feature contribution;

[0007] Estimate the value of multi-subject data based on feature contribution;

[0008] The distribution ratio is determined based on the contribution of each entity and other adjustment factors, and the distributable service performance value is distributed to each entity based on the distribution ratio.

[0009] As a preferred solution of the multi-subject data value market-oriented distribution method based on the vehicle-pile network scenario described in the present invention, the clearly distributable data effectiveness value includes various fees obtained in the process of external services of statistical data service products collected by the operator of the data circulation, transaction and sharing system, including direct charges, indirect charges and other charging data, and the distributable service effectiveness value is calculated.

[0010] As a preferred solution of the multi-agent data value market allocation method based on the vehicle-pile network scenario described in the present invention, wherein:: the multi-agent model contribution degree calculation includes:

[0011] When the value of the performance available for distribution is clearly defined, the data contribution of each entity is measured based on the primary model.

[0012] As a preferred solution of the multi-agent data value market allocation method based on the vehicle-pile network scenario described in the present invention, wherein: the multi-agent model contribution degree measurement includes:

[0013] When the value of the performance available for distribution is clearly defined, the data contribution of each entity is measured based on the primary model.

[0014] As a preferred solution of the multi-agent data value market allocation method based on the vehicle-pile network scenario described in the present invention, wherein: the feature contribution of each subject in the data model includes:

[0015] The average of editorial contributions was calculated repeatedly for all multi-agent coalitions using the first-level model.

[0016] As a preferred solution of the multi-agent data value market allocation method based on the vehicle-pile network scenario described in the present invention, the comprehensive adjustment and balance of the multi-agent feature contribution includes:

[0017] By calculating the overall characteristic contribution of the data subject, the overall characteristic contribution of each data subject is adjusted.

[0018] As a preferred solution of the market-oriented distribution method of multi-subject data value based on the vehicle-pile network scenario described in the present invention, wherein: the estimation of the multi-subject data value based on the feature contribution includes estimating through all possible feature alliances using and not using the features. In the vehicle-pile network scenario, due to the large number of features of each data subject, the number of multi-subject alliances will increase exponentially with the increase of features. This is an approximate valuation method for estimating the multi-subject data value based on the overall feature contribution based on the vehicle-pile network scenario.

[0019] As a preferred solution of the multi-agent data value market distribution method based on the vehicle-pile network scenario described in the present invention, wherein: the distribution ratio is determined according to the contribution of each subject and other adjustment factors, including:

[0020] Based on the contribution of each entity and other adjustment factors, calculate the performance value distribution that each entity should receive.

[0021] As a preferred solution of the multi-agent data value market distribution system based on the vehicle-pile network scenario described in the present invention, it includes: a contribution degree measurement module, an adjustment and balance module, a multi-agent value calculation module, and a distribution module;

[0022] The contribution calculation module clarifies the distributable data effectiveness value, performs multi-agent model contribution calculation, and evaluates the characteristic contribution of each subject in the data model;

[0023] Adjustment and balance module, comprehensive adjustment and balance of multi-subject feature contribution;

[0024] Multi-agent value calculation module estimates the value of multi-agent data based on feature contribution;

[0025] The allocation module determines the allocation ratio based on the contribution of each entity and other adjustment factors, and allocates the distributable service performance value to each entity based on the allocation ratio.

[0026] A computer device includes a memory and a processor, wherein the memory stores a computer program, and is characterized in that when the processor executes the computer program, it implements the steps of any one of the methods for market-based distribution of multi-subject data value based on a vehicle-pile network scenario.

[0027] A computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps of any one of the methods for market-based distribution of multi-agent data value based on a vehicle-pile network scenario are implemented.

[0028] Beneficial effects of the invention: The invention can fully explore and utilize the multi-dimensional value of data, and at the same time, combined with the framework of multi-host value co-creation, realize the fair, transparent and efficient distribution of the effectiveness value of data elements by multiple participants. Compared with the existing technology, this patent has the following advantages:

[0029] 1. Comprehensive evaluation: This patent combines the multi-subject value co-creation model framework to evaluate the value of each subject's data in multiple dimensions, so as to more comprehensively explore and utilize the value of data, and adjust the value contribution of each party on this basis to meet the interests of the value co-creation subject.

[0030] 2. Fair distribution: This patent realizes fair data trading and distribution in the vehicle-pile network scenario, avoiding the unfairness and opacity of traditional distribution methods and improving the rationality and scientificity of benefit distribution.

[0031] 3. Dynamic adjustment: This patent makes timely adjustments and optimizations based on the actual situation of the vehicle-pile network scenario and market feedback, making the method more adaptable and dynamic, meeting the needs of multiple parties for market-oriented distribution of data value. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0033] Figure 1 A schematic diagram of the data acquisition equipment connection and technical flow of a multi-agent data value market-based allocation method based on a vehicle-pile network scenario provided for one embodiment of the present invention.

[0034] Figure 2 A signal matrix diagram of a rule file for writing a method for market-based allocation of multi-agent data value based on a vehicle-pile network scenario provided by an embodiment of the present invention.

[0035] Figure 3 A schematic diagram of a method for market-based distribution of multi-agent data value based on a vehicle-pile network scenario provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0036] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0037] Example 1

[0038] Reference Figure 1-Figure 3 , which is the first embodiment of the present invention, and provides a multi-agent data value market allocation method based on a vehicle-pile network scenario, including:

[0039] Step 1: Identify the assignable primary data outcome value.

[0040] In the embodiment of the present application, the allocable primary data effectiveness value includes statistics collected by the operator of the data circulation, transaction and sharing system (for example, the vehicle charging network platform) on various fees obtained during the external service of a data service product ① (for example, battery health), including direct charges (for example, the cost of purchasing vehicle charging data), indirect charges (for example, data API call fees) and other charges, and calculations are performed to calculate the allocable service effectiveness value V②.

[0041] In an optional embodiment, the clearly allocable primary data effectiveness value is as follows: Vehicle remote diagnostic data analysis: providing real-time diagnostic data of the vehicle, including fault codes, performance indicators, maintenance reminders, etc. Effectiveness value V② calculation: Through the analysis of vehicle remote diagnostic data, valuable information can be provided to automobile manufacturers or maintenance service providers, thereby generating direct charges, such as diagnostic service fees. Charging user satisfaction survey data: Collect and analyze survey results on user satisfaction with charging facilities and services. Effectiveness value V② calculation: These data can help charging facility operators improve their services and increase the frequency of use of charging services by improving user satisfaction, thereby indirectly increasing revenue.

[0042] In an optional embodiment, the assignable primary data effectiveness value is clearly defined as vehicle energy consumption analysis data; analyzing the vehicle's energy consumption, including power consumption rate, cruising range, etc. Effectiveness value V② calculation: Energy consumption analysis data can be provided to electric vehicle manufacturers to help optimize vehicle design and improve energy efficiency, and operators can generate direct charges through data sales. Charging facility market trend analysis data: Provide trend analysis of the charging facility market, including market demand forecasts, competitive landscape, and technological development trends. Effectiveness value V② calculation: These market trend analysis reports can be sold to investors, research institutions, or charging facility manufacturers to generate direct charges, and can also provide market insights to help customers make better business decisions and achieve indirect benefits.

[0043] Step 2: Calculate the contribution of the multi-agent model

[0044] It should be noted that the multi-agent model contribution calculation includes:

[0045] When the value of the performance available for distribution is clearly defined, the data contribution of each entity is measured based on the primary model.

[0046] In the embodiment of the present application, after determining the effectiveness value available for distribution, it is necessary to determine the contribution of each subject (for example, car companies, charging pile operators, charging pile companies, etc.). In order to more accurately measure the contribution of each subject, the first-level model adopts a profit distribution mechanism based on the Shapley value method ③ to measure the data contribution of each subject.

[0047] Specifically, the calculation of multi-subject data contribution can be divided into the following steps:

[0048] In the first step, randomly select another vehicle from the data (the vehicle hash = cfa2870ab07888b21eeba8, frequency = 10, date and mileage can be ignored), use the vehicle hash feature cfa2870ab07888b21eeba8, simulate the alliance of date = 2021, mileage = 200 and frequency = 15, and then use this combination (hash = cfa2870ab07888b21eeba8, frequency = 15, date = 2021, mileage = 200) to predict the PAYD score of the vehicle as 8.0 points.

[0049] In the second step, frequency = 15 is deleted from the alliance, and then replaced with the frequency feature value 10 of the model. The PAYD score of the vehicle predicted by this combination (hash = cfa2870ab07888b21eeba8, frequency = 10, date = 2021, mileage = 200) is 8.5 points. It can be seen that based on the randomly selected vehicles, the predicted score of the vehicle with frequency = 15, date = 2021, mileage = 200 is 8.0 points, and the predicted score of the vehicle with date = 2021, mileage = 200 is 8.5 points. Then the contribution value of frequency = 15 is 8.0 points - 8.5 points = -0.5 points (such as Figure 2 The comparison diagram of the frequency characteristics of the electric vehicle PAYD model is shown in the figure)

[0050] Since the vehicle acts as a contributor to the frequency and hash features, this estimate depends on the value of randomly sampling a vehicle. If we repeat this sampling step and take the average of the contributions, we will get a better estimate.

[0051] In an optional embodiment, the primary model can also use a centrality measurement algorithm; the contribution of each subject is measured by calculating the number of connections in the network. The more other subjects a subject is connected to, the higher its degree centrality is, and it is generally believed that its contribution is greater. Measure the frequency of a subject appearing on all shortest paths in the network. If a subject is on the shortest path between many other subjects, then it has a greater influence in the network and a higher contribution. Measure the average shortest path length from a subject to all other subjects in the network. The shorter the path, the more central the subject is in the network and the higher its contribution.

[0052] In an optional embodiment, the first-level model can also adopt other profit distribution mechanisms in game theory; the nucleus is an equilibrium point in the cooperative game, which meets the fairness requirements of all alliances and is the profit distribution scheme with the least alliance motivation deviation. Define the structure of the cooperative game, including the set of alliances and the profit of each alliance; determine all possible alliances and their corresponding profits. Calculate the nucleus, which usually involves solving a series of linear programming problems. According to the nucleus allocation principle, distribute the profit to each subject;

[0053] Weighted Shapley is an extension of traditional Shapley that allows for different entities to have different importance in the alliance. Assign a weight to each entity that reflects its relative importance in the alliance. Calculate the marginal contribution of each entity in all possible alliances while taking the weight into account. Calculate the Shapley value of each entity based on the weighted marginal contribution. Distribute the benefits based on the weighted Shapley value.

[0054] It should be noted that the selected algorithm must be able to adapt to the complexity of multi-agent systems and be able to process data of different types and formats. The computational complexity of the algorithm should be within an acceptable range so that it can be efficiently executed in practical applications. The algorithm should ensure the fairness and rationality of the distribution of benefits so that each subject can accept the distribution results. The algorithm should have a certain degree of robustness to cope with noise or incompleteness in the data. The results of the algorithm need to be explainable so that the subject can understand how its contribution is measured.

[0055] Step 3: Evaluate the feature contribution of each subject in the data model

[0056] A feature may have different effects on the prediction results under different feature combinations. The present invention takes into account the marginal contribution of this feature under all possible feature sequences. The above only introduces the contribution of the alliance of date=2021 and mileage=200, and the Shapely value is the average of all marginal contributions of all possible alliances, so the present invention repeats this calculation for all possible multi-agent alliances. The calculation time grows exponentially with the number of features and the number of instances sampled in each alliance.

[0057] Below is the union of all eigenvalues ​​for the Shapely value of the target vehicle with frequency = 15:

[0058] {}

[0059] {x hash =3bf03b40f54a9fe398c79a6}

[0060] {x date =2021}

[0061] {x mileage =200}

[0062] {x hash =3bf03b40f54a9fe398c79a6,x date =2021}

[0063] {x hash =3bf03b40f54a9fe398c79a6,x mileage =200}

[0064] {x date =2021,x mileage =200}

[0065] {x hash =3bf03b40f54a9fe398c79a6,x date =2021,x mileage =200}

[0066] For each of the multi-agent alliance, the present invention calculates the PAYD score of the vehicle with and without the frequency=15 feature value, and calculates the difference to obtain the marginal contribution, the Shapely value is the (weighted) average of the marginal contribution, and replaces the feature values ​​of the features not in the alliance with the feature values ​​of random vehicles from the dataset to obtain predictions from the machine learning model.

[0067] Step 4: Comprehensive adjustment and balance of multi-agent feature contribution

[0068] In order to adjust the overall feature contribution of each data subject④, the present invention calculates the overall feature contribution of the data subject. It is necessary to know how each feature affects the prediction. The contribution of each feature can be easily calculated in the linear model. The following is a linear model prediction of a data instance:

[0069]

[0070] Where x is the instance for which we want to calculate the contribution, and each x j (j=1,…,p) is the eigenvalue of the instance, β j is the weight corresponding to feature j. Define the jth feature pair prediction The contribution is

[0071] φ j =β j x j -E(β j X j )=β j x j -β j E(X j

[0072] Where E(β j X j ) is the average estimated impact of feature j, and the contribution is the difference between the feature impact and the average impact. Now we know the contribution of each feature to the prediction. The present invention adds up the contributions of all features of an instance, and the result is as follows:

[0073]

[0074] Based on the above formula, the sum of the contributions of data point x is equal to the predicted value minus the average predicted value. Since there is no similar weight in other nonlinear models (such as ensemble models), a different solution is needed to obtain the feature contribution of a single prediction of the machine learning model through Shapley in cooperative game theory. The Shapley value of each eigenvalue is the contribution of the eigenvalue to the prediction, which is obtained by weighting and summing all possible combinations of eigenvalues:

[0075]

[0076] where S is the subset (union) of features used in the model, x is the vector of feature values ​​of the instances to be explained, and p is the number of features. is the weight of the subset S, which is related to the order. x (S) is the prediction for subset S, marginalized over the features not included in set S:

[0077]

[0078] In fact, multiple integrations are performed for each feature not included. For example: 4 features x1, x2, x3 and x4 are used in the PAYD model, and the present invention estimates the prediction of the alliance S composed of feature values ​​x1 and x3:

[0079]

[0080] The calculated Shapley value is the only attribution method that meets the four properties of efficiency, symmetry, dummy and additivity. They can be used together as the basis for adjusting the contribution of the overall characteristics of each data subject in multiple subjects.

[0081] Step 5: Estimate the value of multi-subject data based on feature contribution

[0082] The precise Shapley value must be estimated by using all possible feature alliances (sets) with and without the jth feature. In the vehicle-pile network scenario, due to the large number of features of each data subject, the number of multi-subject alliances will increase exponentially with the increase of features. Therefore, an approximate valuation method (Shapley sampling value) based on the vehicle-pile network scenario is proposed to estimate the value of multi-subject data according to the overall feature contribution:

[0083]

[0084] in is the prediction of x, except for the value of feature j, the other feature values ​​that are not in the league are replaced by the feature values ​​from the random data point z. The x vector Almost The features that are the same but replaced by the feature values ​​of a random data point z include feature j.

[0085] Approximate estimate of the value of a single data subject’s data:

[0086] Input: number of iterations M, instance of interest x, feature index j, data matrix X, and machine learning model f

[0087] Output: Shapley value of the j-th eigenvalue of instance x

[0088] For all m=1,...,M:

[0089] Randomly sample instances z from the data matrix X

[0090] Generate a random order of features o

[0091] Random order instance x:x o =(x (1) ,…,x (j) ,…,x (p) )

[0092] Random order instance z: z o =(z (1) ,…,z (j) ,…,z (p) )

[0093] Construct two new instances

[0094] With feature j: x +j =(x (1) ,…,x (j-1) ,x (j) ,z (j+1) ,…,z (p) )

[0095] No feature j:x -j =(x (1),…,x (j-1) ,z (j) ,z (j+1) ,…,z (p) )

[0096] Calculate the contribution margin:

[0097] The method for market-based allocation of effectiveness value in the multi-agent value co-creation model of the vehicle-pile network of the present invention implicitly measures the sample weight with the average value through the probability distribution of X, so the average value of multi-agent data can be estimated according to the overall feature contribution as the effectiveness value in the subject value co-creation model:

[0098]

[0099] Step 6: Calculate the market allocation ratio of each entity

[0100] According to each subject's contribution Si and other adjustment factors (such as value contribution, service contribution and resource contribution, etc.), calculate the distribution amount of the effectiveness value that each subject should obtain. Specifically, a distribution ratio⑤ can be determined according to each subject's contribution and other adjustment factors, and then the service effectiveness value V available for distribution can be distributed to each subject according to the ratio. Specifically, it can be achieved through the following steps:

[0101] Interpret the predictions of each subject model as the sum of the attribution values ​​for each input feature:

[0102]

[0103] in:

[0104] g is the explanatory model,

[0105] z′∈{0,1} M Indicates whether the corresponding feature can be observed (1 or 0)

[0106] M is the number of input features,

[0107] is the attributed value (Shapley value) of each feature,

[0108] φ0 is the constant of the explanatory model (this value is actually the same as the Shapley value introduced earlier). That is, the predicted mean of all training samples).

[0109] Substitute scenarios to simplify model predictions

[0110] Since the input of the model in the vehicle-pile network scenario must be structured data, for instance x, z' should be a vector with all values ​​1, that is, all features can be observed, so the formula is simplified to:

[0111]

[0112] Calculate the characteristic attribution value of each subject separately

[0113] In order to calculate the characteristic attribution value φ of each subject i , is defined using the following mathematical method:

[0114] f x (S) = E[f(x)|x S ],

[0115] in

[0116] S is a possible subset of input features;

[0117] pass( Figure 3 E[f(x)] obtains the predicted value example diagram) It can be seen that each subject model assigns the attribution value of each feature to the expected change in the model prediction when adjusting the feature, and interprets the prediction of model f for the sample {x1=a1,x2=a2,x3=a3,x4=a4} as the impact of each feature introduced by the conditional expectation φ j The sum of .

[0118] Substitute the scenario to calculate the market allocation ratio of each entity

[0119] In the multi-agent value co-creation model of the vehicle-pile network scenario, the above step 3 only calculates the ranking of a single subject, and it is also necessary to use the multi-agent data value marketization allocation method and system of the present invention to perform comprehensive sorting and allocation:

[0120] When S is an empty set, Where E[f(x)] is the expectation of the car company's model prediction value, which can be approximated by the average value of the model prediction value of the training samples.

[0121] S sequentially add the pile enterprise model feature x1, at this time That is, the expected value predicted by the model when {x1=a1} - the expected value predicted by the model.

[0122] S sequentially add the Yunnan power grid model feature x2, at this time φ2=f x ({x1,x2})-f x ({x1})=E[f(x)|x1,x2]-E[f(x)|x1], that is, the expected model prediction value when {x1=a1,x2=a2} - the expected model prediction value when {x1=a1}.

[0123] Repeat the above three steps.

[0124] Until the last insurance company model feature x4 is added,

[0125] φ4=f x ({x1,x2,x3,x4}0-f x ({x1,x2,x3}0=E[f(x)|x1,x2,x3,x4]-E[f(x)|x1,x2,x3]

[0126] That is, the expected predicted value of the model when {x1=a1,x2=a2,x3=a3,x4=a4} - the expected predicted value of the model when {x1=a1,x2=a2,x3=a3}. At this time, it is the predicted value under a single ranking of the four features, which is actually the market-based distribution ratio of each entity.

[0127] Through the above steps, a market-based distribution method of effectiveness value based on the multi-subject value co-creation model of the vehicle-pile network can be realized. This distribution method can meet the needs of fair distribution of interests of various subjects and promote the optimization and development of the data factor market.

[0128] Example 2

[0129] The second embodiment of the present invention is different from the previous embodiment in that:

[0130] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0131] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0132] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0133] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0134] Example 3

[0135] The third embodiment of the present invention provides a multi-agent data value market distribution system based on a vehicle-pile network scenario, which is characterized by: comprising a contribution degree measurement module, an adjustment and balance module, a multi-agent value calculation module, and a distribution module;

[0136] The contribution calculation module clarifies the distributable data effectiveness value, performs multi-agent model contribution calculation, and evaluates the characteristic contribution of each subject in the data model;

[0137] Adjustment and balance module, comprehensive adjustment and balance of multi-subject feature contribution;

[0138] Multi-agent value calculation module estimates the value of multi-agent data based on feature contribution;

[0139] The allocation module determines the allocation ratio based on the contribution of each entity and other adjustment factors, and allocates the distributable service performance value to each entity based on the allocation ratio.

[0140] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A multi-agent data value market allocation method based on the vehicle-pile network scenario, characterized by: include, Clarify the allocable primary data effectiveness value, calculate the contribution of the multi-subject model, and evaluate the characteristic contribution of each subject in the data model; Comprehensive adjustment and balance of multi-subject feature contribution; Estimate the value of multi-subject data based on feature contribution; The distribution ratio is determined based on the contribution of each entity and other adjustment factors, and the distributable service performance value is distributed to each entity based on the distribution ratio.

2. The method for market-based distribution of multi-agent data value based on the vehicle-pile network scenario as claimed in claim 1 is characterized by: The clearly allocable first-level data effectiveness value includes the various fees obtained by the operator of the data circulation, transaction and sharing system in the process of providing external services with statistical data service products, and the service effectiveness value that can be allocated is calculated through accounting.

3. The method for market-based distribution of multi-agent data value based on the vehicle-pile network scenario as claimed in claim 2 is characterized by: The multi-agent model contribution calculation includes: When the value of the performance available for distribution is clearly defined, the data contribution of each entity is measured based on the primary model.

4. The method for market-based distribution of multi-agent data value based on the vehicle-pile network scenario as claimed in claim 3 is characterized by: The feature contribution of each subject in the data model includes: The average of editorial contributions was calculated repeatedly for all multi-agent coalitions using the first-level model.

5. The method for market-based distribution of multi-agent data value based on the vehicle-pile network scenario as claimed in claim 4 is characterized by: The comprehensive adjustment and balance of multi-subject feature contribution includes: By calculating the overall characteristic contribution of the data subject, the overall characteristic contribution of each data subject is adjusted.

6. The method for market-based distribution of multi-agent data value based on the vehicle-pile network scenario as claimed in claim 5 is characterized by: The method of estimating the value of multi-subject data based on feature contribution includes estimating through all possible feature alliances using and not using features. In the vehicle-pile network scenario, due to the large number of features of each data subject, the number of multi-subject alliances will increase exponentially with the increase of features. This is an approximate valuation method for estimating the value of multi-subject data based on the overall feature contribution in the vehicle-pile network scenario.

7. The method for market-based distribution of multi-agent data value based on the vehicle-pile network scenario as claimed in claim 6 is characterized by: The distribution ratio determined based on the contribution of each subject and other adjustment factors includes: Based on the contribution of each entity and other adjustment factors, calculate the performance value distribution that each entity should receive.

8. A system based on the multi-agent data value market allocation method based on the vehicle-pile network scenario according to any one of claims 1 to 7, characterized in that: Including contribution measurement module, adjustment and balance module, multi-subject value calculation module, and allocation module; The contribution calculation module clarifies the distributable data effectiveness value, performs multi-agent model contribution calculation, and evaluates the characteristic contribution of each subject in the data model; Adjustment and balance module, comprehensive adjustment and balance of multi-subject feature contribution; Multi-agent value calculation module estimates the value of multi-agent data based on feature contribution; The allocation module determines the allocation ratio based on the contribution of each entity and other adjustment factors, and allocates the distributable service performance value to each entity based on the allocation ratio.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images