Data sharing incentivizing method based on blockchain and evolutionary game theory

Abstract

A data sharing incentivizing method based on a blockchain and an evolutionary game theory, relating to the technical field of data sharing incentivizing mechanisms. The method comprises: constructing a data sharing incentivizing model; on the basis of the data sharing incentivizing model, computing expected gains of users adopting different strategies; computing a replicator dynamics equation on the basis of the expected gains of the users adopting different strategies; and solving evolutionarily stable state points on the basis of the replicator dynamics equation, and analyzing a stable state condition on the basis of the evolutionarily stable state points to obtain a sharing incentivizing scheme. The present invention can improve the user participation rate.

Description

A data sharing incentive method based on blockchain and evolutionary game

[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on June 15, 2023, with application number 202310713465.0 and invention name “A data sharing incentive method based on blockchain and evolutionary game”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present invention belongs to the technical field of data sharing incentive mechanisms, and specifically relates to a data sharing incentive method based on blockchain and evolutionary game. Background Art

[0003] In recent years, numerous researchers have conducted extensive research on computation offloading and resource allocation in MEC systems. Emerging intelligent optimization algorithms, such as reinforcement learning (RL) and deep learning (DRL), have enabled the integration of MEC with intelligent algorithms to achieve real-time data processing and provide efficient and intelligent services in dynamic environments. Furthermore, many studies have proposed cloud-edge collaborative computing architectures to improve computing power and alleviate the computational load on MEC servers. Consequently, many studies have proposed combining cloud-edge collaboration with deep reinforcement learning in edge computing. One such approach is the integration of collective reinforcement learning (CRL) algorithms with blockchain. CRL improves the training efficiency of traditional machine learning by sharing training results among different intelligent agents. However, from a practical implementation perspective, the willingness of individual intelligent nodes (MEC servers) to share data is currently weak due to a lack of mutual trust and a lack of incentive mechanisms and incentives. Currently, blockchain technology can be used to address security and reliability issues in data sharing. Leveraging blockchain's decentralized and tamper-resistant properties offers a new approach to addressing these issues.

[0004] In summary, there is an urgent need for a safe and reliable training data sharing incentive mechanism in distributed scenarios.

[0005] Summary of the Invention

[0006] In response to the shortcomings of the existing technology, the present invention proposes a data sharing incentive method based on blockchain and evolutionary game, which includes:

[0007] S1: Build a data sharing incentive model;

[0008] S2: Calculate the expected benefits of users adopting different strategies based on the data sharing incentive model;

[0009] S3: Calculate the replication dynamic mechanics equation based on the expected returns of users adopting different strategies;

[0010] S4: Solve the evolutionary steady-state point based on the replication dynamic mechanics equation and analyze the steady-state conditions based on the evolutionary steady-state point to obtain the sharing strategy.

[0011] Preferably, the data sharing incentive model includes: a set P consisting of multiple users participating in data sharing, a set N consisting of all users, the user's strategy space S = (s1, s2) and the user's profit strategy U; where s1 indicates that the user participates in sharing, and s2 indicates that the user does not participate in sharing.

[0012] Preferably, the process of calculating the expected benefits of users adopting different strategies includes: setting a direct benefit function and a collaborative benefit function; calculating the expected benefits of users choosing parameter sharing and the expected benefits of users choosing not to share parameters based on the direct benefit function and the collaborative benefit function when the user plays a game with other users.

[0013] Furthermore, the direct benefit function and the collaborative benefit function are: Z = μln(1+R) Y = kμln(1+R)

[0014] Among them, Z represents direct benefit, Y represents collaborative benefit, k represents collaborative benefit coefficient, μ represents scaling coefficient, and R represents direct benefit parameter.

[0015] Furthermore, the expected benefits of the user choosing parameter sharing and the expected benefits of the user choosing not to share parameters are: u(s1,x)=x[kμln(1+R)+I i -τ]+(1-x)[μln(1+R)+I i -τ] u(s2,x)=x(μln(1+R))+(1-x)μln(1+R)

[0016] Among them, u(s1,x) represents the expected benefit of the user choosing parameter sharing, u(s2,x) represents the expected benefit of the user choosing not to share parameters, x represents the probability of the user choosing to participate in data sharing, I i represents the sharing incentive, and τ represents the data sharing cost.

[0017] Preferably, the replicated dynamic mechanics equation is expressed as: F(x) = x(1-x) [u(s1,x)-u(s2,x)]

[0018] Among them, x represents the probability that the user chooses to participate in data sharing, u(s1,x) represents the expected benefit of the user choosing parameter sharing, and u(s2,x) represents the expected benefit of the user choosing not to share parameters.

[0019] Preferably, the evolutionary steady-state points include a first evolutionary steady-state point, a second evolutionary steady-state point and a third evolutionary steady-state point; the three evolutionary steady-state points are: x1=0, x2=1 and Among them, I i represents sharing incentives, τ represents data sharing costs, k represents the collaborative benefit coefficient, μ represents the scaling factor, and R represents the direct benefit parameter.

[0020] Preferably, the sharing incentive scheme is: judging whether the initial proportion of users participating in data sharing is less than the third evolutionary steady-state point, if so, increasing the sharing incentive I i Until I i >τ-(k-1)x * μln(1+R); where τ represents the data sharing cost, k represents the collaborative benefit coefficient, μ represents the scaling coefficient, R represents the direct benefit parameter, and x * Indicates the initial proportion of users participating in data sharing.

[0021] The beneficial effects of the present invention are as follows: the present invention ensures data security by adopting an evolutionary game incentive method through the smart contract technology of the blockchain; the present invention incentivizes nodes to share training results, which is conducive to distributed reinforcement learning to accelerate the convergence of intelligent algorithm models, reduce the waste of computing resources of a single intelligent node, and maximize data value and resource utilization; the evolutionary game incentive method of the present invention can achieve an increase in user participation rate while ensuring the stability of data sharing. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] FIG1 is a flow chart of the data sharing incentive method based on blockchain and evolutionary game in the present invention;

[0023] FIG2 is a diagram of a cloud-edge collaborative distributed scenario model of multiple MEC servers in the present invention;

[0024] FIG3 is a diagram showing the evolutionary stability strategy trend of the data sharing incentive model in the present invention. DETAILED DESCRIPTION

[0025] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0026] The present invention proposes a data sharing incentive method based on blockchain and evolutionary game, as shown in FIG1 . The method includes the following contents:

[0027] S1: Build a data sharing incentive model.

[0028] As shown in Figure 2, the present invention is applicable to cloud-edge collaborative distributed scenarios with multiple MEC servers. The present invention uses an evolutionary game algorithm to establish an incentive mechanism and relies on blockchain technology to encapsulate the game model, enabling the intelligent cluster to share its own training data in the blockchain network. At the same time, the decentralized characteristics of the blockchain network also ensure the privacy of the shared data.

[0029] In a multi-MEC server environment scenario, a data sharing incentive model is constructed. The model can be expressed as a four-element array G = (P, N, S, U); P represents the set P consisting of multiple users participating in data sharing, N represents the set consisting of all users, S = (s1, s2) represents the user's policy space, and U represents the user's profit strategy (profit matrix); among them, s1 represents the user participating in sharing, s2 represents the user not participating in sharing, and the user is the intelligent edge node in the system.

[0030] The relevant parameters of this model are described below:

[0031] (1) The two participants in the evolutionary game belong to the same entity, namely, users of the data sharing community. These users have the same strategy space S = (s1, s2).

[0032] (2) During the data sharing phase, user nodes need to spend a certain amount of data sharing cost, denoted as τ, which is related to the number of CPU cycles consumed in the blockchain consensus phase.

[0033] (3) Assume that in the initial stage of the game, in the set N of all user nodes, the probability that each user chooses to participate in the data sharing strategy is x; therefore, the probability that a user chooses not to participate in data sharing is 1-x; this is equivalent to the ratio of people who choose the sharing strategy being x(t), and the ratio of users who choose not to share being 1-x(t). In the following description, x(t) is abbreviated as x; where x is a function of time t, x∈[0,1].

[0034] (4) The goal of the incentive mechanism is to encourage more users to participate in data sharing. In the data sharing community, in order to encourage users to share data, the data sharing blockchain platform can provide all users with certain external incentive rewards, that is, each node user who participates in sharing its own training data can get a reward from the blockchain system, which is denoted as I.

[0035] (5) User nodes will benefit from uploading data. During the game, relevant data providers decide whether to participate in data sharing by calculating benefits and costs. User benefits can be divided into two categories: direct benefits Z and collaborative benefits Y.

[0036] In the model building process of the present invention, all participating entities are considered to be boundedly rational, which means that they cannot directly reach the Nash equilibrium state at the beginning of the game. Therefore, it is necessary to use an evolutionary simulation to learn the strategies of others, dynamically adjust and try repeatedly to achieve the final optimal balance, thereby maximizing overall benefits. The user's profit strategy is formed in the two-party game between the two users. The game process conforms to the "general two-person symmetric game", that is, members of the group are randomly paired and repeatedly play the game. The user's profit strategy under the game is as follows:

[0037] 1) Case 1: Users A and B both participate in data sharing. All members share data. Both parties not only gain the original benefits, but also gain the shared benefits brought by data sharing. The total benefit is the collaborative benefit Y.

[0038] 2) Case 2: Users A and B do not participate in data sharing. In this case, the data sharing community faces a scenario where no one is willing to share data, so both parties can only obtain the original direct benefit Z;

[0039] Scenario 3: Only one of users A and B chooses to participate in data sharing. The user who chooses this strategy joins the data-sharing blockchain platform and shares their own data, but cannot access data from other users, incurring additional costs without any additional sharing benefits. Non-participating users pay nothing and receive no benefits.

[0040] S2: Calculate the expected benefits of users adopting different strategies based on the data sharing incentive model.

[0041] Set the direct benefit function and the collaborative benefit function; specifically:

[0042] Direct benefits refer to the benefits that data providers can obtain. For the i-th node, its direct benefits only represent the training parameters obtained through training on its own node; the direct benefit function is expressed as: Z = μln(1 + R)

[0043] Among them, Z represents direct benefit; μ represents the scaling factor, which can be a suitable fixed value; R represents the direct benefit parameter.

[0044] Collaboration benefit refers to when nodes i and j participating in decision-making both choose to participate in data sharing. Each node can share data and also receive data shared by other nodes, helping to increase benefits. The collaboration benefit function is expressed as: Y = kμln(1+R)

[0045] Where Y represents the collaborative benefit and k represents the collaborative benefit coefficient. Based on the rationality analysis of the actual situation, the collaborative benefit coefficient k>1, otherwise the node will have no motivation to share.

[0046] The present invention uses a logarithmic benefit function ln (a logarithmic function with e as the base) to calculate the benefit. The characteristics of this function are consistent with the characteristics of real social economics, that is, the benefit has a certain growth inhibition.

[0047] The benefits of users adopting different sharing strategies are expressed in the benefit matrix as follows:

[0048] Among them, I A represents the sharing incentive of user A, I B Represents the sharing incentive of user B.

[0049] When a user plays a game with other users, the expected benefits of the user choosing parameter sharing and the expected benefits of the user choosing not to share parameters are calculated based on the direct benefit function and the collaborative benefit function. Specifically:

[0050] If user A wants to gain more benefits, he must participate. Then, A's income is x[kμln(1+R)+I A -τ]+(1-x)[μln(1+R)+I A -τ]; where x[kμln(1+R)+I A -τ] is the payoff of user A when B participates with probability x; (1-x)[μln(1+R)+I A -τ] is the payoff of user A when B does not participate with probability 1-x. The case of user B can be analyzed in the same way as user A; the expected payoff of the user who chooses parameter sharing is expressed as: u(s1,x)=x[kμln(1+R)+I i -τ]+(1-x)[μln(1+R)+I i -τ]

[0051] The expected benefit of the user choosing not to share parameters is expressed as: u(s2,x)=x(μln(1+R))+(1-x)μln(1+R)

[0052] Among them, u(s1,x) represents the expected benefit of the user choosing parameter sharing, u(s2,x) represents the expected benefit of the user choosing not to share parameters, I i represents the sharing incentive of the i-th user.

[0053] S3: Calculate the replication dynamic mechanics equation based on the expected returns of users adopting different strategies.

[0054] The average expected benefit of data sharing is:

[0055] The replication dynamics equation is essentially a differential equation describing how often a particular strategy is adopted within a population. If the outcomes of adopting a strategy are better than average, then the proportion of those who choose that strategy in the population will increase. The replication dynamics equation is expressed as follows:

[0056] S4: Solve the evolutionary steady-state point based on the replication dynamic mechanics equation and analyze the steady-state conditions based on the evolutionary steady-state point to obtain the sharing strategy.

[0057] The stability of the model is analyzed using the local stability analysis method. Specifically:

[0058] According to the stability theorem of the replication dynamic equation, by bringing the equilibrium point x into the replication dynamic equation, the user's stable strategy can be judged and analyzed. Assuming F(x) = 0, the three stable states of the user replication dynamic equation F(x) can be obtained: the first evolutionary stable point x1 = 0, the second evolutionary stable point x2 = 1, and the third evolutionary stable point x2 = 1. The above-mentioned evolutionary steady-state point is the stable state of the evolutionary stable strategy when the conditions are met.

[0059] In the analysis based on the Evolutionary Stable Strategy (ESS) and the stability theorem, when there are small perturbations in the dynamic system, the steady state should be stable. Therefore, if a point is a stable state, it should satisfy F′(x)<0, that is, F′(x)=(1-2x)[x(k-1)μln(1+R)+I i -τ]+x(1-x)(k-1)μln(1+R)<0. Substituting the three stable points x1, x2, and x3 into the equation, we can get the following formula:

[0060] As shown in Figure 3, ESS is analyzed from the following situations:

[0061] 1) Case 1: (τ-I i )>(k-1)μln(1+R). This condition means that the cost of users participating in data sharing is higher than the benefits they can obtain. At this point, F′(0)<0, F′(1)>0, and x3 does not exist. Therefore, x1 is the only evolutionary equilibrium point of the system. The final ESS of system users, i.e., the sharing incentive scheme, is that they do not participate in data sharing.

[0062] 2) Case 2: 0<(τ-I i )<(k-1)μln(1+R). This condition means that the collaborative benefits that users can obtain from participating in data sharing are higher than the participation costs. At this time, F′(x2)<0, F′(x1)<0, F′(x3)>0; therefore, x1 and x2 are the evolutionary equilibrium points of the system. The final decision of the system users will depend on the initial proportion of users x participating in data sharing in the model.* ; When x * When <x3, the strategy tends not to participate in data sharing; when x * When >x3, the strategy tends to participate in data sharing.

[0063] 3) Case 3: (τ-I i )<(k-1)μln(1+R) and I i -τ>0. Under this condition, F′(x2)<0, F′(x1)>0, F′(x3)<0, and x3<0 does not exist. Therefore, x2 is the evolutionary equilibrium point of the system. Users will eventually choose to participate in data sharing.

[0064] The above analysis analyzes the evolution and stability of data sharing. User choice of strategy influences evolutionary stability. ESS depends on the benefits gained from user data sharing and the proportion of users involved in the system's initial state. The total benefits of participants are related to the participation cost τ, the direct benefit parameter R, and the dynamic incentive benefit I. When users can allocate fixed benefits, dynamic incentive benefits are a key factor in increasing the proportion of system users participating.

[0065] Based on the above stability analysis, when the initial proportion of users participating in data sharing When , users tend to participate in data sharing; I can be used to motivate users to participate in data sharing. i The value range of I i >τ-(k-1)x * μln(1+R)

[0066] Therefore, in order to encourage users to participate in data sharing, the following sharing incentive schemes can be used:

[0067] When the initial proportion of users participating in data sharing is less than the threshold x3, the sharing incentive income is increased until the sharing incentive meets the range I i >τ-(k-1)x * μln(1+R) is used to increase the proportion of users participating in data sharing. When the initial proportion of users participating in data sharing exceeds the threshold, there is no need to increase it. The system can achieve an increase in user participation rate by using the above-mentioned sharing incentive scheme.

[0068] The above embodiments further illustrate the purpose, technical solutions and advantages of the present invention in detail. It should be understood that the above embodiments are only preferred implementation plans of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made to the present invention within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A data sharing incentive method based on blockchain and evolutionary game, characterized in that: include: S1: Build a data sharing incentive model; S2: Calculate the expected benefits of users adopting different strategies based on the data sharing incentive model; S3: Calculate the replication dynamic mechanics equation based on the expected returns of users adopting different strategies; S4: Solve the evolutionary steady-state point based on the replicated dynamic mechanics equation and analyze the steady-state conditions based on the evolutionary steady-state point to obtain a shared incentive scheme.

2. According to claim 1, a data sharing incentive method based on blockchain and evolutionary game, characterized in that: The data sharing incentive model includes: a set P consisting of multiple users participating in data sharing, a set N consisting of all users, the user's strategy space S = (s1, s2) and the user's profit strategy U; wherein s1 indicates that the user participates in sharing, and s2 indicates that the user does not participate in sharing.

3. According to claim 1, a data sharing incentive method based on blockchain and evolutionary game, characterized in that: The process of calculating the expected benefits of users adopting different strategies includes: setting a direct benefit function and a collaborative benefit function; calculating the expected benefits of users choosing parameter sharing and the expected benefits of users choosing not to share parameters based on the direct benefit function and the collaborative benefit function when the user plays a game with other users.

4. According to claim 3, a data sharing incentive method based on blockchain and evolutionary game is characterized in that: The direct benefit function and the collaborative benefit function are: Z = μln (1 + R) Y = kμln (1 + R) Among them, Z represents direct benefit, Y represents collaborative benefit, k represents collaborative benefit coefficient, μ represents scaling factor, and R represents direct benefit parameter.

5. According to claim 3, a data sharing incentive method based on blockchain and evolutionary game is characterized in that: The expected benefits of the user choosing parameter sharing and the expected benefits of the user choosing not to share parameters are: u(s1,x)=x[kμln(1+R)+I i -τ]+(1-x)[μln(1+R)+I i -τ] u(s2,x)=x(μln(1+R))+(1-x)μln(1+R) Among them, u(s1,x) represents the expected benefit of the user choosing parameter sharing, u(s2,x) represents the expected benefit of the user choosing not to share parameters, x represents the probability of the user choosing to participate in data sharing, I i represents the sharing incentive, and τ represents the data sharing cost.

6. According to claim 1, a data sharing incentive method based on blockchain and evolutionary game is characterized in that: The replicated dynamic mechanical equation is expressed as: F(x)=x(1-x)·[u(s1,x)-u(s2,x)] Among them, x represents the probability that the user chooses to participate in data sharing, u(s1,x) represents the expected benefit of the user choosing parameter sharing, and u(s2,x) represents the expected benefit of the user choosing not to share parameters.

7. According to claim 1, a data sharing incentive method based on blockchain and evolutionary game, characterized in that: The evolutionary steady-state points include a first evolutionary steady-state point, a second evolutionary steady-state point and a third evolutionary steady-state point; the three evolutionary steady-state points are: x1=0, x2=1 and Among them, I i represents sharing incentive, τ represents data sharing cost, k represents collaboration benefit coefficient, μ represents scaling factor, and R represents direct benefit parameter.

8. According to claim 1, a data sharing incentive method based on blockchain and evolutionary game, characterized in that: The sharing incentive scheme is: determine whether the initial proportion of users participating in data sharing is less than the third evolutionary steady-state point. If so, increase the sharing incentive I i Until I i >τ-(k-1)x * μln(1+R); where τ represents the data sharing cost, k represents the collaborative benefit coefficient, μ represents the scaling factor, R represents the direct benefit parameter, and x * Indicates the initial proportion of users participating in data sharing.

Images