Multi-subject data transaction method oriented to data element market

Abstract

The invention relates to a data element market-oriented multi-subject data transaction method, which comprises the following steps of: firstly, constructing a transaction block chain network in a data element market through attribute parameters of a data supplier, a data provider and a data demander in the data element market, and selecting a core data provider; secondly, a revenue function of a data supplier, a benefit function of a data demander and a utility function of a data demander are respectively constructed based on the attribute parameters of all the subjects and the core data demander, the overall benefit of all the subjects in the transaction block chain network is maximized based on a Q learning algorithm, and the quoted price after adjustment of the data demander is generated; then, submitting the adjusted quotation to a transaction block chain network, and triggering a transaction through an intelligent contract; and finally, after the smart contract verifies the transaction compliance, layered payment is started, and the transaction is completed. According to the method, mathematical quantification and balanced solution of a multi-party benefit game are realized, and the flexibility and fairness of data transaction and the resource configuration efficiency are improved.

Description

Technical Field

[0001] The present application relates to the technical field of multi-subject data transactions, and in particular to a multi-subject data transaction method for a data element market. Background Art

[0002] Existing multi-party data transaction methods attempt to introduce blockchain technology to achieve decentralized transactions, but have the following shortcomings: the node role division is single, and it is difficult to fully consider the differentiated needs of data suppliers, data vendors, and demanders; there is a lack of a quantitative interest distribution model, making it difficult to balance the short-term benefits and long-term cooperation of all parties; the dynamic strategy generation mechanism relies on preset rules and cannot adapt to market changes.

[0003] Blockchain technology's distributed ledgers, smart contracts, and consensus mechanisms offer a new paradigm for data transactions. Its decentralized nature enables transparency in data ownership and transfer, automated smart contract execution reduces transaction costs, and consensus mechanisms ensure that data cannot be tampered with. However, existing blockchain data trading platforms primarily focus on evidence storage and settlement, lack dynamic profit distribution mechanisms, and fail to effectively address the ambiguity of data value and the complex demands of multi-party transactions. Summary of the Invention

[0004] Based on this, it is necessary to provide a multi-subject data transaction method for the data element market, which includes: S1: Set attribute parameters for data suppliers, data vendors, and data demanders in the data factor market; build a transaction blockchain network in the data factor market based on the attribute parameters, and select core data vendors; S2: Based on the attribute parameters of each entity and the core data providers, the revenue function of the data supplier, the benefit function of the data provider, and the utility function of the data demander are constructed. Based on the Q-learning algorithm, the overall benefits of each entity in the transaction blockchain network are maximized, and the adjusted quotation of the data provider is generated. The adjusted quotation is submitted to the transaction blockchain network and the transaction is triggered through the smart contract; S3: After the smart contract verifies the transaction compliance, it initiates the layered payment and completes the transaction.

[0005] Beneficial effects: This method first constructs a transaction blockchain network in the data element market through the attribute parameters of data suppliers, data vendors and data demanders in the data element market, and selects core data vendors; secondly, based on the attribute parameters of each subject and the core data vendors, the revenue function of the data supplier, the benefit function of the data vendor and the utility function of the data demander are constructed respectively, and based on the Q-learning algorithm, the overall benefits of each subject in the transaction blockchain network are maximized to generate adjusted quotations from the data vendors; then, the adjusted quotations are submitted to the transaction blockchain network and the transactions are triggered through smart contracts; finally, after the smart contract verifies the compliance of the transaction, the hierarchical payment is initiated to complete the transaction; this method realizes the mathematical quantification and equilibrium solution of the multi-party interest game, realizes the real-time optimization and closed-loop feedback of the transaction strategy, improves the flexibility, fairness and resource allocation efficiency of data transactions, and provides a new solution for building a reliable and efficient data transaction market. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0007] Figure 1 This is a flowchart of the multi-subject data transaction method for the data element market in an embodiment of the present application. DETAILED DESCRIPTION

[0008] To make the above-mentioned objects, features, and advantages of the present application more clearly understood, the specific embodiments of the present application are described in detail below with reference to the accompanying drawings. The following description sets forth many specific details to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar improvements without violating the scope of the present application. Therefore, the present application is not limited to the specific embodiments disclosed below.

[0009] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of such features. Throughout the description of this application, "plurality" means at least two, for example, two, three, etc., unless otherwise specifically defined.

[0010] like Figure 1 As shown, this embodiment provides a multi-agent data transaction method for a data element market, the method comprising: S1: Set attribute parameters for data suppliers, data vendors, and data demanders in the data element market; build a transaction blockchain network in the data element market based on the attribute parameters, and select core data vendors.

[0011] In this embodiment, The attribute parameters of the data supplier include data quality , storage costs ( is the storage unit price, For the i The amount of data from each data provider, Storage time), transmission capacity ( For the i The bandwidth of each data provider, For the i transmission distance of each data provider).

[0012] The attribute parameters of data providers include reputation value (Reputation value is the historical transaction success rate, T (time period), processing delay ( For the j The service rate of each data provider, For the j request arrival rate of each data provider).

[0013] The attribute parameters of the data demander include budget cap and timeliness requirements.

[0014] Furthermore, the transaction blockchain network in the data factor market based on attribute parameters includes: The data suppliers, data vendors, and data demanders in the data factor market are respectively regarded as nodes, and the attribute parameters of the data suppliers, data vendors, and data demanders are respectively regarded as the attributes of the corresponding nodes; Based on the data quality of the data supplier connected to any data provider, as well as the reputation value and processing delay of the corresponding data provider, the node weight of the corresponding data provider is calculated as follows: ; in, Indicates the j The node weight corresponding to each data provider; Indicates the j The reputation value of individual data providers; Indicates the j A collection of data suppliers connected to each data provider; Indicates the j The first data provider connected ithe data quality of individual data suppliers; Indicates the j Processing delays by individual data providers; Traverse all data quotients and calculate the node weights corresponding to all data quotients; A transaction blockchain network is constructed based on all nodes, the attributes of each node, and the node weight corresponding to each data provider.

[0015] Furthermore, the selection of core data providers includes: determining the core data providers based on the node weights corresponding to the data providers and the transmission capacity between the data providers and the data suppliers, and the calculation formula is: ; in, Indicates core data provider; Represents a data quotient set; Indicates the j The node weight corresponding to each data provider; Indicates the j The transmission capacity between a data business and its connected data suppliers.

[0016] S2: Based on the attribute parameters of each subject and the core data provider, the revenue function of the data supplier, the benefit function of the data provider, and the utility function of the data demander are constructed respectively. Based on the Q-learning algorithm, the overall benefits of each subject in the transaction blockchain network are maximized, and the adjusted quotation of the data provider is generated. The adjusted quotation is submitted to the transaction blockchain network and the transaction is triggered through the smart contract.

[0017] In this embodiment, in order to maximize the long-term interests of the data trading market alliance, a benefit function system for data suppliers, data vendors, and data demanders is constructed based on the node data and network structure in the transaction blockchain network. By quantifying the interests of all parties, this system provides a mathematical basis for the dynamic trading strategy optimization of the Q learning algorithm, ensuring that strategy generation meets the global optimal goal.

[0018] Specifically, the revenue function of the data provider is expressed as: ; ; in, Indicates the i The revenue function of each data supplier; Indicates the number of data providers; Indicates the i Data suppliers and j The benchmark transaction price of each data provider; Indicates the j The first data provider connected ithe data quality of individual data suppliers; represents the quality competition factor, ; Represents a collection of data suppliers; Indicates the j The first data provider connected l Data quality representation of individual data suppliers; Indicates the j The data provider and the i The transmission capacity between individual data suppliers; Indicates the j The data provider and the l The transmission capacity between individual data suppliers; Indicates the i Data suppliers and j Incentive pricing between data providers based on transmission capacity; Indicates the i Storage costs of individual data providers; Represents the reward coefficient.

[0019] The benefit function of the data provider is expressed as: ; ; st ; in, Indicates the j The benefit function of each data provider; Indicates the number of data demanders; Indicates the j A data provider to k Quotations from individual data demanders; represents the timeliness penalty function; Indicates the j Data providers and k The actual transmission time between data demanders; Indicates the k The timeliness requirements of individual data demanders; Indicates the number of data suppliers; Indicates the i Storage costs of individual data providers; Indicates the j The first data provider connected i the data quality of individual data suppliers; Indicates the j The reputation value of individual data providers; represents the credibility adjustment factor; represents the budget upper limit of the k-th data demander for a single transaction; represents the demand matching response function; Indicates the j The data provider provides k Data quality of individual data demanders; represents the risk aversion coefficient; Indicates the j The quotation given by a data provider to the first data demander; Indicates the j A data provider to The quotation of a data demander; st means constrained by; represents variance; represents the mass sensitivity coefficient; Indicates the j The data provider provides k The benchmark data quality of each data demander.

[0020] Data quality is calculated across three dimensions, including: Integrity: For data providers Submitted dataset Perform integrity checks ;in, Indicates the i Data providers The integrity test results of Indicates the number of data in the dataset; Represents an indicator function, the value is 1 when the condition is met, otherwise it is 0; Indicates the i Data providers Dataset Middle f individual data; Indicates empty; Timeliness: ;in, Indicates the i The timeliness of individual data providers; is the unit of time; is the data type coefficient; Indicates the time of response; Indicates the time when the request started; Information entropy: , , , ;in, Indicates the i The information entropy of individual data suppliers; Indicates the iThe original information entropy of each data provider; represents the maximum possible information entropy; Indicates the number of intervals into which the data is divided; Indicates that it falls on b The probability of an interval; Indicates the total number of records; Multi-dimensional quality fusion ; The index parameters are determined by objectively assigning weights using the entropy weight method. First, the evaluation matrix is constructed. , calculate the index information entropy: , , and then determine the weight , we get: first index , the second index , the third index ;in, Indicates the m The information entropy of each indicator; Indicates the number of data suppliers; After standardization, i The data provider is m The data value of each indicator; Indicates the i The data provider is m The data value of an indicator.

[0021] The utility function of the data demander is expressed as: ; st ; in, Indicates the k The utility function of each data demander; Indicates the number of data providers; Indicates the j The data provider provides k Data quality of individual data demanders; represents the quality sensitivity index; represents the credit discount factor; Indicates the j The reputation value of individual data providers; represents the price elasticity coefficient; Indicates the j A data provider to k Quotations from individual data demanders; represents the budget upper limit of the k-th data demander for a single transaction; Represents an indicator function, the value is 1 when the condition is met, otherwise it is 0; Indicates thej Data providers and k The actual transmission time between data demanders; Indicates the k The timeliness requirement of a data demander; st represents the constraint.

[0022] Furthermore, relying on the above-mentioned benefit function system, based on the Q-learning algorithm to maximize the overall benefits of each entity in the transaction blockchain network, the adjusted quotes generated by the data providers include: Step 1: Define the state space, where the state is represented as: ; ; ; ; ; in, express t The state of the moment; express t Time-varying mass decay at each moment; express t The average market price at that moment; express t Average reputation value of online data at the moment; express t The average remaining budget of data demanders at each moment; Indicates the number of data suppliers; Indicates the j The first data provider connected i the data quality of individual data suppliers; represents the mass attenuation coefficient, ; Indicates the number of data providers; express t Moment j Quotations from individual data providers; express t Moment j The reputation value of individual data providers; Indicates the number of data demanders; represents the budget upper limit of the k-th data demander for a single transaction; Indicates the j A data provider to k Quotations from individual data demanders; Step 2: Design the action space of the data provider. The actions of the data provider are expressed as: ; in, express t Moment j Actions of individual data providers; express t Moment j The price adjustment amount of each data provider; express t Moment j Bandwidth allocation ratio for each data provider; express t Moment j The service quality commitment threshold of individual data providers; t The service quality commitment thresholds at each moment include: t Data integrity commitment threshold at a given moment , which represents the minimum requirement for the completeness rate of the data field; t The upper limit of response delay at the moment , in seconds, represents the maximum tolerable time from accepting a request to completing delivery; t Data freshness threshold at a certain moment , in seconds, represents the maximum tolerated interval between data generation time and transaction time.

[0023] Step 3: Based on the Q function, and adopt ε- The greedy strategy selects actions, and the action selection is expressed as: ; in, represents the Q function; Represents a set of actions to be selected; Represents the selection probability, which decays with the training rounds. The decay formula is: , represents the selection probability before decay, represents the decay rate; Step 4: Update the Q function. The update formula is: ; ; in, express t The action of momentary choice; represents the learning rate, ; express t The reward function at that moment; represents the discount factor; express t +1 moment status; express t Moment jThe benefit function of each data provider; express t -1 moment j The benefit function of each data provider; represents the penalty coefficient, ; express t -1 moment j Quotations from individual data providers; represents the updated set of candidate actions; in this embodiment, the learning rate can be dynamically adjusted, and the adjustment formula is: ; Step 5: Repeat steps 3-4 until the Q function converges, search for the action with the largest Q value, and add the price adjustment amount of the action with the largest Q value to the data provider's quotation at the previous moment to obtain the adjusted quotation of the data provider.

[0024] Based on the interest relationship between the three parties (the revenue function of the data supplier, the benefit function of the data vendor, and the utility function of the data demander), the optimization goal (maximizing the overall benefit) is provided for the dynamic trading strategy, and the strategy generation is guided by the game equilibrium mechanism. The Q learning model is constructed by market status and action, and the use of ε- The greedy strategy balances exploration and exploitation, updating the strategy based on changes in returns and price fluctuations to form a dynamic optimization closed loop.

[0025] S3: After the smart contract verifies the transaction compliance, it initiates the layered payment and completes the transaction.

[0026] Specifically, this step includes: Transactions are represented as: ;in, Indicates the j Data providers, Indicates the k Data demanders, Indicates the j A data provider to k The price quote adjusted by the data demander, Indicates the j The data provider provides k The data quality of individual data demanders, Indicates the j Data providers and k The actual transmission time between data demanders; The smart contract determines whether the transaction meets the constraints. If so, it distributes the benefits according to the tiered payment model and completes the transaction. The constraints are: st ; Among them, st means constrained to; Indicates the j A data provider to k Quotations from individual data demanders; Indicates the k The minimum data quality required by individual data demanders; represents the budget upper limit of the k-th data demander for a single transaction; Indicates the k The timeliness requirements of individual data demanders; Indicates the k The total budget of each data demander.

[0027] Furthermore, the transaction process is: The total payment amount of the data demander is: ; in, Indicates the j A data provider to k The total bid of the data demander, that is, k The data demander must pay j The total amount of individual data providers; Indicates the j A data provider to k The adjusted quotation of each data demander; represents the quality premium coefficient, ; Indicates the time-based reward amount; Indicates the j The data provider provides k Data quality of individual data demanders; Indicates the k The minimum data quality required by individual data demanders; Indicates the j Data providers and k The actual transmission time between data demanders; Indicates the k The timeliness requirements of individual data demanders.

[0028] The benefits to data providers are: ; in, Indicates the j The income of individual data providers; Indicates the j The commission ratio of individual data providers, , , Data integrity commitment thresholds , response delay threshold , data freshness threshold The weight of represents the delay penalty rate.

[0029] The benefits to data providers are: ; in, Indicates the i The benefits of individual data providers; Indicates the number of data providers; Indicates the j The first data provider connected i the data quality of individual data suppliers; represents the quality competition factor; Represents a collection of data suppliers; Indicates the j The first data provider connected l Data quality representation of individual data suppliers; Indicates the j The data provider and its connected i The transmission capacity between individual data suppliers; Indicates the j The data provider and the p The transmission capacity between data suppliers; Indicates the number of data suppliers.

[0030] In this embodiment, this step also includes: The reputation value of the data provider is updated based on the total number of historical transactions, the number of successful transactions in this round, and the number of transactions in this round. The update formula is: ; in, Indicates the j The updated reputation value of each data provider; Indicates the j The total number of historical transactions of each data provider; Indicates the j The historical reputation value of each data provider; Indicates the transaction success indicator function, the value is 1 if the transaction is successful, otherwise it is 0; Indicates the j The number of new transactions with individual data providers in this round of transactions; The data quality of the data supplier is updated based on the historical data quality and the number of times the data is accessed. The update formula is: ; in, Indicates the i The quality of data after updates from individual data suppliers; Indicates the i The quality of data before the update by the data supplier; represents the natural attenuation coefficient; Indicates a time interval; Indicates the k Data demanders; Indicates the number of times the data has been accessed; represents variance; Indicates the j A data provider to k The adjusted quotation of each data demander; The updated reputation and data quality are fed back to step S2 to calculate the reward function and update the Q value, thereby optimizing the next round of trading strategy decisions. The blockchain ensures state synchronization and transaction immutability, forming an adaptive closed loop of "strategy generation → transaction execution → state feedback → model iteration", ultimately achieving long-term profit maximization and market equilibrium.

[0031] The multi-agent data transaction method for the data element market provided by this embodiment has the following beneficial effects: This method first constructs a transaction blockchain network in the data element market through the attribute parameters of data suppliers, data vendors and data demanders in the data element market, and selects core data vendors; secondly, based on the attribute parameters of each subject and the core data vendors, the revenue function of the data supplier, the benefit function of the data vendor and the utility function of the data demander are constructed respectively, and based on the Q-learning algorithm, the overall benefits of each subject in the transaction blockchain network are maximized to generate adjusted quotations from the data vendors; then, the adjusted quotations are submitted to the transaction blockchain network and the transactions are triggered through smart contracts; finally, after the smart contract verifies the compliance of the transaction, the hierarchical payment is initiated to complete the transaction; this method realizes the mathematical quantification and equilibrium solution of the multi-party interest game, realizes the real-time optimization and closed-loop feedback of the transaction strategy, improves the flexibility, fairness and resource allocation efficiency of data transactions, and provides a new solution for building a reliable and efficient data transaction market.

[0032] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned 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.

[0033] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A multi-agent data transaction method for a data element market, characterized in that: include: S1: Set attribute parameters for data suppliers, data vendors, and data demanders in the data factor market; Build a transaction blockchain network in the data factor market based on attribute parameters and select core data providers; S2: Based on the attribute parameters of each entity and the core data providers, the revenue function of the data supplier, the benefit function of the data provider, and the utility function of the data demander are constructed. Based on the Q-learning algorithm, the overall benefits of each entity in the transaction blockchain network are maximized, and the adjusted quotation of the data provider is generated. The adjusted quotation is submitted to the transaction blockchain network and the transaction is triggered through the smart contract; S3: After the smart contract verifies the transaction compliance, it initiates the layered payment and completes the transaction.

2. The multi-agent data transaction method for the data element market according to claim 1 is characterized in that: The attribute parameters of data suppliers include data quality, storage cost, and transmission capacity; the attribute parameters of data vendors include reputation value and processing delay; the attribute parameters of data demanders include budget cap and timeliness requirements.

3. The multi-agent data transaction method for the data element market according to claim 2 is characterized in that: The transaction blockchain network in the data factor market based on attribute parameters includes: The data suppliers, data vendors, and data demanders in the data factor market are respectively regarded as nodes, and the attribute parameters of the data suppliers, data vendors, and data demanders are respectively regarded as the attributes of the corresponding nodes; Based on the data quality of the data supplier connected to any data provider, as well as the reputation value and processing delay of the corresponding data provider, the node weight of the corresponding data provider is calculated as follows: ； in, Indicates the j The node weight corresponding to each data provider; Indicates the j The reputation value of individual data providers; Indicates j A collection of data suppliers connected to each data provider; Indicates j The first data provider connected i the data quality of individual data suppliers; Indicates the j Processing delays by individual data providers; Traverse all data quotients and calculate the node weights corresponding to all data quotients; A transaction blockchain network is constructed based on all nodes, the attributes of each node, and the node weight corresponding to each data provider.

4. The multi-agent data transaction method for the data element market according to claim 3 is characterized in that: The selection of core data providers includes: determining the core data providers based on the node weights corresponding to the data providers and the transmission capacity between the data providers and the data suppliers. The calculation formula is: ； in, Indicates core data provider; Represents a data quotient set; Indicates the j The node weight corresponding to each data provider; Indicates the j The transmission capacity between a data business and its connected data suppliers.

5. The multi-agent data transaction method for the data element market according to claim 2 is characterized in that: The revenue function of the data provider is expressed as: ； ； in, Indicates the i The revenue function of each data supplier; Indicates the number of data providers; Indicates the i Data suppliers and j The benchmark transaction price of each data provider; Indicates j The first data provider connected i the data quality of individual data suppliers; represents the quality competition factor, ; Represents a collection of data suppliers; Indicates j The first data provider connected l Data quality representation of individual data suppliers; Indicates the j The data provider and its connected i The transmission capacity between data suppliers; Indicates the j The data provider and its connected l The transmission capacity between data suppliers; Indicates the i Data suppliers and j Incentive pricing between data providers based on transmission capacity; Indicates the i Storage costs of individual data providers; Represents the reward coefficient.

6. The multi-agent data transaction method for the data element market according to claim 2 is characterized in that: The benefit function of the data provider is expressed as: ； ； s.t. ; in, Indicates the j The benefit function of each data provider; Indicates the number of data demanders; Indicates the j A data provider to k Quotations from individual data demanders; represents the timeliness penalty function; Indicates the j Data providers and k The actual transmission time between data demanders; Indicates the k The timeliness requirements of individual data demanders; Indicates the number of data suppliers; Indicates the i Storage costs of individual data providers; Indicates the j The first data provider connected i the data quality of individual data suppliers; Indicates the j The reputation value of individual data providers; represents the credibility adjustment factor; represents the budget upper limit of the k-th data demander for a single transaction; represents the demand matching response function; Indicates the j The data provider provides k Data quality of individual data demanders; represents the risk aversion coefficient; Indicates the j The quotation given by a data provider to the first data demander; Indicates the j A data provider to The quotation of a data demander; st means constrained by; represents variance; represents the mass sensitivity coefficient; Indicates the j The data provider provides k The benchmark data quality of each data demander.

7. The multi-agent data transaction method for the data element market according to claim 6 is characterized in that: The utility function of the data demander is expressed as: ； s.t. ; in, Indicates the k The utility function of each data demander; Indicates the number of data providers; Indicates the j The data provider provides k Data quality of individual data demanders; represents the quality sensitivity index; represents the credit discount factor; Indicates the j The reputation value of individual data providers; represents the price elasticity coefficient; Indicates the j A data provider to k Quotations from individual data demanders; represents the budget upper limit of the k-th data demander for a single transaction; Represents an indicator function, the value is 1 when the condition is met, otherwise it is 0; Indicates the j Data providers and k The actual transmission time between data demanders; Indicates the k The timeliness requirement of a data demander; st represents the constraint.

8. The multi-agent data transaction method for the data element market according to claim 1 is characterized in that: Based on the Q-learning algorithm to maximize the overall benefits of each entity in the transaction blockchain network, the generated quotes after data providers adjust the prices include: Step 1: Define the state space. The state is represented as: ； ； ； ； ； in, express t The state of the moment; express t Time-varying mass decay at each moment; express t The average market price at that moment; express t Average reputation value of online data at the moment; express t The average remaining budget of data demanders at each moment; Indicates the number of data suppliers; Indicates the j The first data provider connected i the data quality of individual data suppliers; represents the mass attenuation coefficient, ; Indicates the number of data providers; express t Moment j Quotations from individual data providers; express t Moment j The reputation value of individual data providers; Indicates the number of data demanders; represents the budget upper limit of the k-th data demander for a single transaction; Indicates the j A data provider to k Quotations from individual data demanders; Step 2: Design the action space of the data provider. The actions of the data provider are expressed as: ； in, express t Moment j Actions of individual data providers; express t Moment j The price adjustment amount of each data provider; express t Moment j Bandwidth allocation ratio for individual data providers; express t Moment j The service quality commitment threshold of individual data providers; Step 3: Based on the Q function, and adopt ε- The greedy strategy selects actions, and the action selection is expressed as: ； in, represents the probability of selection; represents the Q function; Represents a set of actions to be selected; Step 4: Update the Q function. The update formula is: ； ； in, express t The action of momentary choice; represents the learning rate; express t The reward function at that moment; represents the discount factor; express t +1 moment status; express t Moment j The benefit function of each data provider; express t -1 moment j The benefit function of each data provider; represents the penalty coefficient, ; express t -1 moment j Quotations from individual data providers; Represents the updated set of candidate actions; Step 5: Repeat steps 3-4 until the Q function converges, search for the action with the largest Q value, and add the price adjustment amount of the action with the largest Q value to the data provider's quotation at the previous moment to obtain the adjusted quotation of the data provider.

9. The multi-agent data transaction method for the data element market according to claim 1 is characterized in that S3 include: Transactions are represented as: ;in, Indicates the j Data providers, Indicates the k Data demanders, Indicates the j A data provider to k The price quote adjusted by the data demander, Indicates the j The data provider provides k The data quality of individual data demanders, Indicates the j Data providers and k The actual transmission time between data demanders; The smart contract determines whether the transaction meets the constraints. If so, it distributes the benefits according to the tiered payment model and completes the transaction. The constraints are: s.t. ; Among them, st means constrained to; Indicates the j A data provider to k Quotations from individual data demanders; Indicates the k The minimum data quality required by individual data demanders; represents the budget upper limit of the k-th data demander for a single transaction; Indicates the k The timeliness requirements of individual data demanders; Indicates the k The total budget of each data demander.

10. The multi-agent data transaction method for the data element market according to claim 1 is characterized in that: S3 also includes: The reputation value of the data provider is updated based on the total number of historical transactions, the number of successful transactions in this round, and the number of transactions in this round. The update formula is: ； in, Indicates the j The updated reputation value of each data provider; Indicates the j The total number of historical transactions of each data provider; Indicates the j The historical reputation value of each data provider; Indicates the transaction success indicator function, the value is 1 if the transaction is successful, otherwise it is 0; Indicates the j The number of new transactions with individual data providers in this round of transactions; The data quality of the data supplier is updated based on the historical data quality and the number of times the data is accessed. The update formula is: ； in, Indicates the i The quality of data after updates from individual data suppliers; Indicates the i The quality of data before the update by the data supplier; represents the natural attenuation coefficient; Indicates a time interval; Indicates the k Data demanders; Indicates the number of times the data is accessed; represents variance; Indicates the j A data provider to k The adjusted quotation of each data demander; The updated reputation value and data quality are fed back to step S2 to calculate the reward function and update the Q value, thereby optimizing the next round of trading strategy decisions.

Images