CN113724096A - Group knowledge sharing method based on public commodity evolution game model - Google Patents
Group knowledge sharing method based on public commodity evolution game model Download PDFInfo
- Publication number
- CN113724096A CN113724096A CN202110942467.8A CN202110942467A CN113724096A CN 113724096 A CN113724096 A CN 113724096A CN 202110942467 A CN202110942467 A CN 202110942467A CN 113724096 A CN113724096 A CN 113724096A
- Authority
- CN
- China
- Prior art keywords
- user
- strategy
- group
- game
- income
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
本发明属于信息资源共享技术领域,公开了一种基于公共品演化博弈模型的群体知识共享方法。本发明基于在线学习社区的特征构建群体知识共享对应的公共品演化博弈模型,结合粒子群优化算法,根据用户的行为特性将用户的策略连续化,调整用户的记忆系数和模仿系数,以动态调整用户的共享策略,能够提高群体共享水平,使得社区能够达到高共享水平稳态,从而促使社区长期稳定的发展。
The invention belongs to the technical field of information resource sharing, and discloses a group knowledge sharing method based on an evolutionary game model of public goods. Based on the characteristics of the online learning community, the invention constructs a public goods evolutionary game model corresponding to group knowledge sharing, combines the particle swarm optimization algorithm, and continuously adjusts the user's strategy according to the user's behavioral characteristics, and adjusts the user's memory coefficient and imitation coefficient to dynamically adjust The user's sharing strategy can improve the level of group sharing, so that the community can achieve a high level of sharing stability, thereby promoting the long-term and stable development of the community.
Description
技术领域technical field
本发明属于信息资源共享技术领域,更具体地,涉及一种基于公共品演化博弈模型的群体知识共享方法。The invention belongs to the technical field of information resource sharing, and more particularly, relates to a group knowledge sharing method based on a public goods evolutionary game model.
背景技术Background technique
在线学习社区中知识共享是非常重要的环节,但是由于进行知识共享行为需要付出一定的成本,以至于社区中很多成员采用不共享的策略,这会导致群体的知识共享程度低下,网络共享数据量较低,不利于社区网络的健康发展,因此,如何促进社区中的成员进行知识共享、提高群体知识共享程度是本领域技术人员迫切需要解决的一个技术问题。Knowledge sharing is a very important link in online learning communities, but because of the cost of knowledge sharing behavior, many members of the community adopt the strategy of not sharing, which will lead to a low degree of knowledge sharing in the group and the amount of data shared on the network. It is not conducive to the healthy development of the community network. Therefore, how to promote knowledge sharing among members of the community and improve the degree of group knowledge sharing is a technical problem that those skilled in the art need to solve urgently.
发明内容SUMMARY OF THE INVENTION
为了克服现有技术中存在的不足,本发明的目的是提供一种基于公共品演化博弈模型的群体知识共享方法,以提高群体共享水平,使得社区能够达到高共享水平稳态,从而促使社区长期稳定的发展。In order to overcome the deficiencies in the prior art, the purpose of the present invention is to provide a group knowledge sharing method based on the evolutionary game model of public goods, so as to improve the level of group sharing, so that the community can reach a steady state of high sharing level, thereby promoting the long-term development of the community. stable development.
本发明提供一种基于公共品演化博弈模型的群体知识共享方法,包括以下步骤:The present invention provides a group knowledge sharing method based on the evolutionary game model of public goods, comprising the following steps:
步骤1、定义在线学习社区的群体为一个包含多个用户节点的集合,将用户进行知识共享的比例作为用户的策略,将用户调整知识共享比例的程度作为用户的策略调整速度,将用户获取的知识总量作为用户的收益,构建群体知识共享对应的公共品演化博弈模型;Step 1. Define the group of the online learning community as a set containing multiple user nodes, take the proportion of users' knowledge sharing as the user's strategy, take the degree to which the user adjusts the proportion of knowledge sharing as the user's strategy adjustment speed, and use the user's acquisition rate as the user's strategy adjustment speed. The total amount of knowledge is used as the income of users, and a public goods evolutionary game model corresponding to group knowledge sharing is constructed;
其中,知识共享比例的取值范围为[0,1];Among them, the value range of the knowledge sharing ratio is [0, 1];
步骤2、初始化所有用户节点,以相同的概率随机设置每个用户的策略,设置每个用户的策略调整速度为0,并对初始化的策略和初始化的策略调整速度进行记录;Step 2. Initialize all user nodes, randomly set each user's policy with the same probability, set the policy adjustment speed of each user to 0, and record the initialized policy and the initialized policy adjustment speed;
步骤3、根据记录的每个用户的策略,通过所述公共品演化博弈模型计算每个用户的收益,并对收益进行记录;Step 3, according to the recorded strategy of each user, calculate the income of each user through the public goods evolutionary game model, and record the income;
步骤4、根据记录的每个用户的收益,得到每个用户的历史最优收益和该用户历史最优收益对应的历史最优策略,以及每个用户所在群体中获得的最优收益和该最优收益对应的最优策略;Step 4. According to the recorded income of each user, obtain the historical optimal income of each user and the historical optimal strategy corresponding to the historical optimal income of the user, as well as the optimal income obtained in the group of each user and the optimal historical income of the user. The optimal strategy corresponding to the optimal income;
步骤5、根据每个用户的历史最优收益和每个用户所在群体中获得的最优收益,对每个用户的记忆系数和模仿系数进行更新;Step 5. Update the memory coefficient and imitation coefficient of each user according to the historical optimal income of each user and the optimal income obtained in the group where each user belongs;
其中,所述记忆系数表示用户根据自身经验改变策略的概率,模仿系数表示用户根据该用户所在群体经验改变策略的概率;Wherein, the memory coefficient represents the probability that the user changes the strategy according to his own experience, and the imitation coefficient represents the probability that the user changes the strategy according to the experience of the group where the user belongs;
步骤6、结合粒子群优化算法对每个用户的策略和策略调整速度进行更新,更新后重复执行步骤3至步骤5,直到群体的知识共享水平达到稳定状态;Step 6, update the strategy and strategy adjustment speed of each user in combination with the particle swarm optimization algorithm, and repeat steps 3 to 5 after the update, until the knowledge sharing level of the group reaches a stable state;
其中,将用户的策略对应于粒子群优化算法中的粒子位置,将用户的策略调整速度对应于粒子群优化算法中的粒子速度,将用户的收益对应于粒子群优化算法中的粒子适应度。Among them, the user's strategy corresponds to the particle position in the particle swarm optimization algorithm, the user's strategy adjustment speed corresponds to the particle speed in the particle swarm optimization algorithm, and the user's income corresponds to the particle fitness in the particle swarm optimization algorithm.
优选的,所述步骤2中,初始化用户的策略采用如下方式:Preferably, in the step 2, the strategy for initializing the user is as follows:
strategy=slimit(1)+(slimit(2)-slimit(1))*rand(1,N)strategy=slimit(1)+(slimit(2)-slimit(1))*rand(1,N)
其中,strategy表示策略,slimit(1)、slimit(2)分别表示策略上限、策略下限,rand函数生成随机数值,N表示用户总人数,rand(1,N)表示生成N个取值范围在[0,1]的数值并赋予N个用户节点作为初始化策略。Among them, strategy represents the strategy, slimit(1) and slimit(2) represent the upper limit of the strategy and the lower limit of the strategy, respectively, the rand function generates random values, N represents the total number of users, and rand(1, N) represents the generation of N values ranging in [ 0,1] and assign N user nodes as the initialization strategy.
优选的,所述步骤3中,计算用户的收益采用如下方式:Preferably, in the step 3, the following methods are used to calculate the user's income:
其中,Ui,j(t)表示第t轮博弈时用户i参与以邻居j为中心的群体中获得的收益,i∈N,j∈N,N表示用户总人数;Qi表示用户i所拥有的知识量,Si,t表示用户i在第t轮博弈时的策略;ki表示用户i的邻居数量,ki+1表示用户i及其邻居的数量之和,r表示增益系数,kj表示用户j的邻居数量,kj+1表示用户j及其邻居数量之和;Ωj表示以用户j为中心的邻居集合;l∈(Ωj∪j)表示用户l属于用户j及其邻居集合,Ql表示用户l所拥有的知识量,Sl,t表示用户l在第t轮博弈时的策略,kl表示用户l的邻居数量,kl+1表示用户l及其邻居的数量之和;Among them, U i,j (t) represents the income obtained by user i participating in the group centered on neighbor j in the t-th round of the game, i∈N, j∈N, N represents the total number of users; Q i represents the total number of users i The amount of knowledge possessed, S i,t represents the strategy of user i in the t-th round of the game; ki represents the number of neighbors of user i , ki +1 represents the sum of the number of user i and its neighbors, r represents the gain coefficient, k j represents the number of neighbors of user j, k j + 1 represents the sum of user j and the number of neighbors; Ω j represents the neighbor set centered on user j; l∈(Ω j ∪j) represents that user l belongs to user j and Its neighbor set, Q l represents the amount of knowledge possessed by user l, S l, t represents the strategy of user l in the t-th round of the game, k l represents the number of neighbors of user l, k l +1 represents user l and its neighbors the sum of the quantities;
其中,Ui(t)表示第t轮博弈时用户i获得的累积收益,Ωi表示以用户i为中心的邻居集合,j∈(Ωi∪i)表示用户j属于用户i及其邻居集合;C表示成本系数。Among them, U i (t) represents the cumulative income obtained by user i in the t-th round of the game, Ω i represents the set of neighbors centered on user i, and j∈(Ω i ∪i) represents that user j belongs to user i and its neighbor set ; C is the cost factor.
优选的,所述步骤5中,记忆系数记为w,模仿系数记为1-w,对用户的记忆系数和模仿系数进行更新采用如下方式:Preferably, in the step 5, the memory coefficient is recorded as w, the imitation coefficient is recorded as 1-w, and the user's memory coefficient and imitation coefficient are updated as follows:
其中,表示用户i在开始第t+1轮博弈前,获得的历史最优收益;表示在第t轮博弈时用户i所在群体中获得的最优收益,用户i所在群体包括用户i及其邻居集合;P表示用户i在第t轮博弈中获得的收益。in, Represents the historical optimal profit obtained by user i before starting the t+1th round of the game; Indicates the optimal profit obtained by user i in the group of user i in the t-th round of the game, and the group of user i includes user i and its neighbors; P represents the profit obtained by user i in the t-th round of the game.
优选的,所述步骤6中,结合粒子群优化算法对用户的策略和策略调整速度进行更新采用如下方式:Preferably, in the step 6, the following methods are used to update the user's strategy and strategy adjustment speed in combination with the particle swarm optimization algorithm:
Si,t+1=Si,t+vi,t+1 S i,t+1 =S i,t +v i,t+1
其中,vi,t+1表示第t+1轮博弈时用户i的策略调整速度;vi,t表示第t轮博弈时用户i的策略调整速度;w表示记忆系数,表示用户i在开始第t+1轮博弈前,获得的历史最优策略;Si,t表示用户i在第t轮博弈时的策略;表示在第t轮博弈时用户i所在群体中获得最高收益的用户采用的策略,用户i所在群体包括用户i及其邻居集合;Sj,t表示表示用户j在第t轮博弈时的策略;Si,t+1表示用户i在第t+1轮博弈时的策略。Among them, vi ,t+1 represents the strategy adjustment speed of user i in the t+1th round of the game; vi ,t represents the strategy adjustment speed of user i in the tth round of the game; w represents the memory coefficient, Represents the historical optimal strategy obtained by user i before starting the t+1th round of the game; S i,t represents the strategy of user i in the tth round of the game; Represents the strategy adopted by the user who obtains the highest profit in the group of user i in the t-th round of the game, and the group of user i includes the user i and its neighbors; S j,t represents the strategy of user j in the t-th round of the game; S i,t+1 represents the strategy of user i in the t+1th round of the game.
本发明中提供的一个或多个技术方案,至少具有如下技术效果或优点:One or more technical solutions provided in the present invention have at least the following technical effects or advantages:
在发明中,定义在线学习社区的群体为一个包含多个用户节点的集合,将用户进行知识共享的比例作为用户的策略,将用户调整知识共享比例的程度作为用户的策略调整速度,将用户获取的知识总量作为用户的收益,构建群体知识共享对应的公共品演化博弈模型;对所有用户对应的策略和策略调整速度进行初始化;基于用户的策略,通过公共品演化博弈模型计算每个用户的收益;基于用户的收益,得到每个用户的历史最优收益和该用户历史最优收益对应的历史最优策略,以及每个用户所在群体中获得的最优收益和该最优收益对应的最优策略;根据每个用户的历史最优收益和每个用户所在群体中获得的最优收益,对每个用户的记忆系数和模仿系数进行更新;将用户的策略对应于粒子群优化算法中的粒子位置,将用户的策略调整速度对应于粒子群优化算法中的粒子速度,将用户的收益对应于粒子群优化算法中的粒子适应度,结合粒子群优化算法对每个用户的策略和策略调整速度进行更新,更新后重复执行上述步骤直到群体达到稳态。即本发明基于在线学习社区的特征构建群体知识共享对应的公共品演化博弈模型,结合粒子群优化算法,根据用户的行为特性将用户的策略连续化,调整用户的记忆系数和模仿系数,以动态调整用户的共享策略,能够提高群体共享水平,使得社区能够达到高共享水平稳态,从而促使社区长期稳定的发展。In the invention, the group of the online learning community is defined as a set containing multiple user nodes, the proportion of users' knowledge sharing is taken as the user's strategy, the degree to which users adjust the proportion of knowledge sharing is taken as the user's strategy adjustment speed, and the user acquisition The total amount of knowledge is used as the income of users, and the public goods evolutionary game model corresponding to group knowledge sharing is constructed; the strategies and strategy adjustment speeds corresponding to all users are initialized; based on the user's strategy, the public goods evolutionary game model is used to calculate each user's Revenue; based on the user's revenue, obtain the historical optimal revenue of each user and the historical optimal strategy corresponding to the user's historical optimal revenue, as well as the optimal revenue obtained in each user's group and the optimal revenue corresponding to the optimal revenue. The optimal strategy; update the memory coefficient and imitation coefficient of each user according to the historical optimal income of each user and the optimal income obtained in the group of each user; the user's strategy corresponds to the particle swarm optimization algorithm in the Particle position, the user's strategy adjustment speed corresponds to the particle speed in the particle swarm optimization algorithm, and the user's income corresponds to the particle fitness in the particle swarm optimization algorithm, combined with the particle swarm optimization algorithm to adjust each user's strategy and strategy The speed is updated, and after the update, the above steps are repeated until the population reaches a steady state. That is, the present invention builds a public goods evolutionary game model corresponding to group knowledge sharing based on the characteristics of the online learning community, combines with the particle swarm optimization algorithm, and continuously adjusts the user's strategy according to the user's behavioral characteristics, and adjusts the user's memory coefficient and imitation coefficient to dynamically Adjusting the user's sharing strategy can improve the level of group sharing, so that the community can achieve a high level of sharing stability, thus promoting the long-term and stable development of the community.
附图说明Description of drawings
图1为本发明实施例提供的一种基于公共品演化博弈模型的群体知识共享方法的流程图。FIG. 1 is a flowchart of a group knowledge sharing method based on a public goods evolutionary game model provided by an embodiment of the present invention.
具体实施方式Detailed ways
为了更好的理解上述技术方案,下面将结合说明书附图以及具体的实施方式对上述技术方案进行详细的说明。In order to better understand the above technical solutions, the above technical solutions will be described in detail below with reference to the accompanying drawings and specific embodiments.
对于知识共享,传统的博弈模型采用的是纯策略,即共享或者不共享。而在现实情况下,对于具有独立行动、思考、判断和决策能力的用户而言,其决策行为过程较为复杂并具有一定程度的不确定性,因此相对于纯策略来说,混合策略会更加的合理。因此,根据在线学习社区的特征,本发明采用公共品博弈模型对其进行模拟,该模型属于多人博弈类型,与在线学习社区交互行为的特性更加的契合。For knowledge sharing, the traditional game model adopts pure strategy, that is, sharing or not sharing. In reality, for users with independent action, thinking, judgment and decision-making abilities, their decision-making behavior process is more complex and has a certain degree of uncertainty. Therefore, compared with pure strategies, mixed strategies will be more effective. Reasonable. Therefore, according to the characteristics of the online learning community, the present invention uses a public goods game model to simulate it, which belongs to the multiplayer game type and is more in line with the characteristics of the online learning community's interactive behavior.
本发明定义每个用户可以决定自己共享的比例,该比例就是用户进行知识共享选择的策略。将策略定义在[0,1]这一区间,以刻画用户共享知识的程度和意愿,0表示用户未共享任何知识,但是根据社区特性,其也可以从别人的共享中获的收益,并且无需承担任何损失;而1表示用户共享所有的知识量,付出的成本是因分享而失去知识独占性所带来的损失,以及在知识共享过程中花费的时间、精力等。如此,用户的不同共享比例对应不同的策略。The present invention defines the ratio that each user can decide to share by himself, and the ratio is the strategy for the user to choose knowledge sharing. The strategy is defined in the interval [0,1] to describe the degree and willingness of users to share knowledge. 0 means that users do not share any knowledge, but according to the characteristics of the community, they can also benefit from the sharing of others, and there is no need to Bear any loss; and 1 means that the user shares all the knowledge, and the cost is the loss caused by the loss of knowledge exclusivity due to sharing, as well as the time and energy spent in the process of knowledge sharing. In this way, different sharing ratios of users correspond to different strategies.
粒子群优化算法(PSO)具有很好的演化寻优特性,大多是用于“连续变量”中寻优,而连续策略空间公共品演化博弈中个体的策略更新过程与PSO算法中粒子的行为有很多相似的特征,由此启发,本发明引入PSO算法。Particle swarm optimization (PSO) has good evolutionary optimization characteristics, and it is mostly used for optimization in "continuous variables", while the strategy update process of individuals in the evolutionary game of public goods in continuous strategy space is similar to the behavior of particles in the PSO algorithm. Many similar features, inspired by this, the present invention introduces the PSO algorithm.
PSO算法初始化生成一群随机例子,算法的核心是每一次迭代,粒子通过追随自身历史最优解和种群历史最优解两个“极值”来调整速度和位置,该动态调整的速度是更新粒子位置的关键,而粒子位置象征着离最优解的距离,适应度是评价该粒子解的唯一标准。从生物和社会角度而言,用户都希望自己通过交互能够获得较高的收益,因此收益对用户的策略选择具有重要的影响。因此,本发明在用户策略更新的过程中引入PSO算法,将用户的策略(贡献比例)对应于粒子位置,把用户的收益对应于粒子的适应度,用户结合个体经验及社会群体经验调整贡献比例,进行策略的更新和演化。The PSO algorithm is initialized to generate a group of random examples. The core of the algorithm is each iteration. The particle adjusts the speed and position by following the two "extreme values" of its own historical optimal solution and the population historical optimal solution. The speed of this dynamic adjustment is to update the particle The key to the position, and the particle position symbolizes the distance from the optimal solution, the fitness is the only criterion for evaluating the particle solution. From a biological and social point of view, users hope that they can obtain higher benefits through interaction, so the benefits have an important impact on the user's strategy choice. Therefore, the present invention introduces the PSO algorithm in the process of user policy update, which corresponds the user's policy (contribution ratio) to the particle position, and the user's income corresponds to the particle's fitness, and the user adjusts the contribution ratio based on individual experience and social group experience. , to update and evolve the strategy.
本发明将粒子群优化算法与公共品博弈技术相结合,根据用户的行为特性将用户的策略连续化,设置动态函数调整用户的记忆系数和模仿系数,以动态调整用户的共享策略,能够提高群体共享水平,使得社区能够达到高共享水平稳态,从而促使社区长期稳定的发展。The invention combines the particle swarm optimization algorithm with the game technology of public goods, and the user's strategy is continuous according to the user's behavioral characteristics, and the dynamic function is set to adjust the user's memory coefficient and imitation coefficient, so as to dynamically adjust the user's sharing strategy, and can improve the group. The sharing level enables the community to achieve a steady state of high sharing level, thereby promoting the long-term and stable development of the community.
本实施例提供了一种基于公共品演化博弈模型的群体知识共享方法,参见图1,包括以下步骤:This embodiment provides a group knowledge sharing method based on the evolutionary game model of public goods, see FIG. 1 , and includes the following steps:
步骤1:定义在线学习社区的群体为一个包含多个用户节点的集合,将用户进行知识共享的比例作为用户的策略,将用户调整知识共享比例的程度作为用户的策略调整速度,将用户获取的知识总量作为用户的收益,构建群体知识共享对应的公共品演化博弈模型。Step 1: Define the group of the online learning community as a set containing multiple user nodes, take the proportion of users' knowledge sharing as the user's strategy, take the user's adjustment of the proportion of knowledge sharing as the user's strategy adjustment speed, The total amount of knowledge is used as the income of users, and a public goods evolutionary game model corresponding to group knowledge sharing is constructed.
其中,知识共享比例的取值范围为[0,1]。Among them, the value range of the knowledge sharing ratio is [0, 1].
即根据在线学习社区和社区中的用户学习特征,构建相应的公共品演化博弈模型。特征主要指:社区用户具有一定的知识水平(即所拥有的知识量),用户的共享意愿,用户进行知识共享需要花费的成本,用户能够获得的知识收益等。That is, according to the online learning community and the user learning characteristics in the community, the corresponding public goods evolutionary game model is constructed. Features mainly refer to: community users have a certain level of knowledge (that is, the amount of knowledge they possess), users' willingness to share, the cost of knowledge sharing by users, and the knowledge benefits that users can obtain.
设置了一个由N个用户节点组成的群体,由集合N={1,2,...,N}表示;根据在线学习社区用户的学习特征,对每一个用户节点i∈N进行了“知识量”“共享成本”“策略”“策略调整速度”“博弈收益”的定义。A group consisting of N user nodes is set up, which is represented by the set N={1, 2, ..., N}; according to the learning characteristics of users in the online learning community, each user node i∈N is subjected to "knowledge". Definition of “quantity”, “shared cost”, “strategy”, “speed of strategy adjustment” and “game benefit”.
知识量Q指:用户自身所拥有的知识量。共享成本C(成本系数)指:用户进行知识共享需要花费的时间精力等成本。策略S指:用户进行知识共享的比例。策略调整速度v指:用户调整知识共享比例的程度,具体的,用户根据自身惯性速度(即用户在上一次博弈中的策略调整速度,初次为0)、个体经验(即用户学习自身最优策略的程度)、社会经验(即用户向群体中收益最高的用户学习策略的程度)调整策略的方向(即程度)。收益P指:用户博弈后获得的知识量。The amount of knowledge Q refers to: the amount of knowledge the user owns. Sharing cost C (cost coefficient) refers to the time and energy that users need to spend to share knowledge. Strategy S refers to the proportion of users who share knowledge. The strategy adjustment speed v refers to: the degree to which the user adjusts the knowledge sharing ratio. Specifically, the user is based on his own inertial speed (that is, the user's strategy adjustment speed in the last game, which is 0 for the first time), individual experience (that is, the user learns his own optimal strategy) degree), social experience (that is, the degree to which the user learns the strategy from the users with the highest profit in the group) to adjust the direction (that is, the degree) of the strategy. Profit P refers to the amount of knowledge gained by users after gaming.
步骤2:初始化所有用户节点,以相同的概率随机设置每个用户的策略,设置每个用户的策略调整速度为0,并对初始化的策略和初始化的策略调整速度进行记录。Step 2: Initialize all user nodes, randomly set each user's policy with the same probability, set each user's policy adjustment speed to 0, and record the initialized policy and the initialized policy adjustment speed.
一方面,对策略进行初始化时,设strategy表示各个用户的策略集合,不同的策略代表着不同的贡献比例,其为连续值,此处可以用以下方式进行初始化:On the one hand, when initializing the strategy, let strategy represent the strategy set of each user, and different strategies represent different contribution ratios, which are continuous values, and can be initialized in the following ways:
strategy=slimit(1)+(slimit(2)-slimit(1))*rand(1,N)strategy=slimit(1)+(slimit(2)-slimit(1))*rand(1,N)
其中,slimit(1)和slimit(2)分别表示策略上限、策略下限,rand函数生成随机数值,N表示用户总人数,rand(1,N)表示生成N个取值范围在[0,1]的数值并赋予N个用户节点作为初始化策略。Among them, slimit(1) and slimit(2) represent the upper limit of the policy and the lower limit of the policy respectively, the rand function generates random values, N represents the total number of users, and rand(1,N) represents the generation of N values ranging from [0,1] and assign N user nodes as the initialization strategy.
例如,slimit(1)=0,slimit(2)=1。For example, slimit(1)=0, slimit(2)=1.
另一方面,将各个用户的策略调整速度均初始化为0。On the other hand, the policy adjustment speed of each user is initialized to 0.
在接下来的博弈过程中,每个用户会根据自身历史收益和其他邻居的收益来调整自己的策略调整速度,而后进一步的更新策略。In the next game process, each user will adjust his strategy adjustment speed according to his own historical income and the income of other neighbors, and then further update the strategy.
步骤3:根据记录的每个用户的策略,通过所述公共品演化博弈模型计算每个用户的收益,并对收益进行记录。Step 3: Calculate the income of each user through the public goods evolutionary game model according to the recorded strategy of each user, and record the income.
用户i参与以其自身为中心及以其邻居为中心的共ki+1个群体的博弈,其中ki是用户i的邻居数量。User i participates in a game of ki + 1 groups centered on itself and its neighbors, where ki is the number of neighbors of user i.
所述步骤3包括以下子步骤:The step 3 includes the following sub-steps:
步骤3.1:计算每个用户在每轮博弈中参与以某一邻居为中心的群体(包括该用户)时获得的收益。Step 3.1: Calculate the income obtained by each user when participating in a group centered on a neighbor (including the user) in each round of the game.
第t轮博弈时,用户i参与以邻居j为中心的群体中获得的收益表示为:In the t-th round of the game, the benefits obtained by user i participating in the group centered on neighbor j are expressed as:
其中,Ui,j(t)表示第t轮博弈时用户i参与以邻居j为中心的群体中获得的收益,i∈N,j∈N,N表示用户总人数;Qi表示用户i所拥有的知识量,Si,t表示用户i在第t轮博弈时的策略(分享比例),0≤Si,t≤1;ki表示用户i的邻居数量,ki+1表示用户i及其邻居的数量之和,r表示的是增益系数,kj表示用户j的邻居数量,kj+1表示用户j及其邻居数量之和;Ωj表示以用户j为中心的邻居集合;l∈(Ωj∪j)表示用户l属于用户j及其邻居集合,Ql表示用户l所拥有的知识量,Sl,t表示用户l在第t轮博弈时的策略,kl表示用户l的邻居数量,kl+1表示用户l及其邻居的数量之和。Among them, U i,j (t) represents the income obtained by user i participating in the group centered on neighbor j in the t-th round of the game, i∈N, j∈N, N represents the total number of users; Q i represents the total number of users i The amount of knowledge possessed, S i,t represents the strategy (sharing ratio) of user i in the t-th round of the game, 0≤S i,t ≤1; ki represents the number of neighbors of user i , and ki +1 represents user i The sum of the number of its neighbors, r represents the gain coefficient, k j represents the number of neighbors of user j, k j + 1 represents the sum of the number of user j and its neighbors; Ω j represents the set of neighbors centered on user j; l∈(Ω j ∪j) indicates that user l belongs to user j and its neighbor set, Q l indicates the amount of knowledge possessed by user l, S l, t indicates user l’s strategy in the t-th round of the game, k l indicates user l The number of neighbors of l, k l + 1 represents the sum of the number of user l and its neighbors.
例如,增益系数为1.56,将群体中共享的知识量翻1.56倍后平均分给群体中的成员,再减去成员的共享量,这就是成员在这个群体中获得的收益。For example, the gain coefficient is 1.56. After multiplying the amount of knowledge shared in the group by 1.56 times, it is equally distributed to the members of the group, and then the amount of sharing by the members is subtracted, which is the benefit that members get in this group.
步骤3.2:计算每个用户在每轮博弈中获得的总收益。Step 3.2: Calculate the total revenue that each user gets in each round of the game.
第t轮博弈时用户i获得的累积收益Ui(t)表示为:The cumulative income U i (t) obtained by user i in the t-th round of the game is expressed as:
其中,Ωi表示以用户i为中心的邻居集合,j∈(Ωi∪i)表示用户j属于用户i及其邻居集合;C表示成本系数。Among them, Ω i represents the neighbor set centered on user i, j∈(Ωi∪i) represents that user j belongs to user i and its neighbor set; C represents the cost coefficient.
例如,成本系数为0.01,将用户共享的知识量乘以成本系数0.01,即为用户进行知识共享所花费的成本。花费的成本应远小于用户的共享知识量,所以成本系数也应足够小。For example, if the cost coefficient is 0.01, multiply the amount of knowledge shared by users by the cost coefficient 0.01, which is the cost of knowledge sharing by users. The cost of spending should be much less than the amount of shared knowledge of users, so the cost factor should also be small enough.
步骤4:根据记录的每个用户的收益,得到每个用户的历史最优收益和该用户历史最优收益对应的历史最优策略,以及每个用户所在群体中获得的最优收益和该最优收益对应的最优策略。Step 4: According to the recorded income of each user, obtain the historical optimal income of each user and the historical optimal strategy corresponding to the historical optimal income of the user, as well as the optimal income obtained in the group of each user and the optimal historical income of the user. The optimal strategy corresponding to the optimal return.
步骤5:根据每个用户的历史最优收益和每个用户所在群体中获得的最优收益,对每个用户的记忆系数和模仿系数进行更新。Step 5: Update the memory coefficient and imitation coefficient of each user according to the historical optimal income of each user and the optimal income obtained in each user's group.
其中,所述记忆系数表示用户根据自身经验改变策略的概率,模仿系数表示用户根据该用户所在群体经验改变策略的概率。Wherein, the memory coefficient represents the probability that the user changes the strategy according to his own experience, and the imitation coefficient represents the probability that the user changes the strategy according to the experience of the group where the user belongs.
即根据步骤4得出的相关数据,对用户的记忆系数w和模仿系数(1-w)进行更新。更新公式如下:That is, according to the relevant data obtained in step 4, the user's memory coefficient w and imitation coefficient (1-w) are updated. The update formula is as follows:
其中,表示用户i在开始第t+1轮博弈前,获得的历史最优收益,表示在第t轮博弈时用户i所在群体中(包括自身)获得的最优收益,P表示用户i在本次(第t轮)博弈中获得的收益。in, represents the historical optimal profit obtained by user i before starting the t+1th round of the game, Indicates the optimal income obtained by user i in the group (including itself) in the t-th round of the game, and P represents the income obtained by user i in this (t-th round) game.
步骤6:结合粒子群优化算法对每个用户的策略和策略调整速度进行更新,更新后重复执行步骤3至步骤5,直到群体的知识共享水平达到稳定状态。Step 6: Combine the particle swarm optimization algorithm to update the strategy and strategy adjustment speed of each user. After the update, repeat steps 3 to 5 until the knowledge sharing level of the group reaches a stable state.
其中,将用户的策略对应于粒子群优化算法中的粒子位置,将用户的策略调整速度对应于粒子群优化算法中的粒子速度,将用户的收益对应于粒子群优化算法中的粒子适应度。Among them, the user's strategy corresponds to the particle position in the particle swarm optimization algorithm, the user's strategy adjustment speed corresponds to the particle speed in the particle swarm optimization algorithm, and the user's income corresponds to the particle fitness in the particle swarm optimization algorithm.
即本发明利用PSO算法的核心思想对策略进行更新,此处策略这一属性对应着PSO算法中的“位置”,收益代表PSO算法中的“适应度”,策略调整速度对应PSO算法中的“速度”。That is, the present invention uses the core idea of the PSO algorithm to update the strategy. Here, the attribute of the strategy corresponds to the "position" in the PSO algorithm, the income represents the "fitness" in the PSO algorithm, and the strategy adjustment speed corresponds to the "position" in the PSO algorithm. speed".
结合粒子群优化算法对用户的策略和策略调整速度进行更新采用如下方式:Combined with the particle swarm optimization algorithm, the following methods are used to update the user's strategy and strategy adjustment speed:
Si,t+1=Si,t+vi,t+1 S i,t+1 =S i,t +v i,t+1
其中,vi,t+1表示第t+1轮博弈时用户i的策略调整速度;vi,t表示第t轮博弈时用户i的策略调整速度;w表示记忆系数,表示用户i在开始第t+1轮博弈前,获得的历史最优策略;Si,t表示用户i在第t轮博弈时的策略;表示在第t轮博弈时用户i所在群体中获得最高收益的用户采用的策略,用户i所在群体包括用户i及其邻居集合;Sj,t表示表示用户j在第t轮博弈时的策略;Si,t+1表示用户i在第t+1轮博弈时的策略。Among them, vi ,t+1 represents the strategy adjustment speed of user i in the t+1th round of the game; vi ,t represents the strategy adjustment speed of user i in the tth round of the game; w represents the memory coefficient, Represents the historical optimal strategy obtained by user i before starting the t+1th round of the game; S i,t represents the strategy of user i in the tth round of the game; Represents the strategy adopted by the user who obtains the highest profit in the group of user i in the t-th round of the game, and the group of user i includes the user i and its neighbors; S j,t represents the strategy of user j in the t-th round of the game; S i,t+1 represents the strategy of user i in the t+1th round of the game.
当群体中知识共享水平达到稳定时,所有用户都会选择对社区的健康发展有利的策略,能够达到提高群体的知识共享程度,提高网络共享数据量,促进社区网络健康发展的技术效果。When the level of knowledge sharing in the group is stable, all users will choose strategies that are beneficial to the healthy development of the community, which can achieve the technical effect of improving the level of knowledge sharing in the group, increasing the amount of data shared on the network, and promoting the healthy development of the community network.
综上,采用本发明提供的基于公共品演化博弈模型的群体知识共享方法能够促使用户向高收益的用户学习,能够使得群体能够达到较高的合作水平。To sum up, using the group knowledge sharing method based on the evolutionary game model of public goods provided by the present invention can promote users to learn from high-income users, and can enable the group to achieve a higher level of cooperation.
最后所应说明的是,以上具体实施方式仅用以说明本发明的技术方案而非限制,尽管参照实例对本发明进行了详细说明,本领域的普通技术人员应当理解,可以对本发明的技术方案进行修改或者等同替换,而不脱离本发明技术方案的精神和范围,其均应涵盖在本发明的权利要求范围当中。Finally, it should be noted that the above specific embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to examples, those of ordinary skill in the art should understand that the technical solutions of the present invention can be Modifications or equivalent substitutions without departing from the spirit and scope of the technical solutions of the present invention should be included in the scope of the claims of the present invention.
Claims (5)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110942467.8A CN113724096B (en) | 2021-08-17 | 2021-08-17 | Group knowledge sharing method based on public evolution game model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110942467.8A CN113724096B (en) | 2021-08-17 | 2021-08-17 | Group knowledge sharing method based on public evolution game model |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113724096A true CN113724096A (en) | 2021-11-30 |
CN113724096B CN113724096B (en) | 2023-09-15 |
Family
ID=78676043
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110942467.8A Active CN113724096B (en) | 2021-08-17 | 2021-08-17 | Group knowledge sharing method based on public evolution game model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113724096B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114257507A (en) * | 2021-12-22 | 2022-03-29 | 中国人民解放军国防科技大学 | A method to improve the level of network information sharing based on evolutionary game theory |
CN114844789A (en) * | 2022-04-20 | 2022-08-02 | 华中师范大学 | Community knowledge sharing evaluation method based on evolutionary game model |
CN117592236A (en) * | 2023-12-05 | 2024-02-23 | 北京大数据先进技术研究院 | Data sharing network strategy evolution prediction method, device and product |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090112678A1 (en) * | 2007-10-26 | 2009-04-30 | Ingram Micro Inc. | System and method for knowledge management |
EP2296395A1 (en) * | 2009-09-09 | 2011-03-16 | Deutsche Telekom AG | System and method to derive deployment strategies for metropolitan wireless networks using game theory |
CN111582429A (en) * | 2020-05-12 | 2020-08-25 | 陕西师范大学 | Method for solving evolutionary game problem based on brain storm optimization algorithm |
CN112182485A (en) * | 2020-09-22 | 2021-01-05 | 华中师范大学 | Online knowledge sharing dynamic rewarding method based on evolutionary game |
-
2021
- 2021-08-17 CN CN202110942467.8A patent/CN113724096B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090112678A1 (en) * | 2007-10-26 | 2009-04-30 | Ingram Micro Inc. | System and method for knowledge management |
EP2296395A1 (en) * | 2009-09-09 | 2011-03-16 | Deutsche Telekom AG | System and method to derive deployment strategies for metropolitan wireless networks using game theory |
CN111582429A (en) * | 2020-05-12 | 2020-08-25 | 陕西师范大学 | Method for solving evolutionary game problem based on brain storm optimization algorithm |
CN112182485A (en) * | 2020-09-22 | 2021-01-05 | 华中师范大学 | Online knowledge sharing dynamic rewarding method based on evolutionary game |
Non-Patent Citations (1)
Title |
---|
王鹏民;侯贵生;杨磊;: "基于知识质量的社会化问答社区用户知识共享的演化博弈分析", 现代情报, no. 04 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114257507A (en) * | 2021-12-22 | 2022-03-29 | 中国人民解放军国防科技大学 | A method to improve the level of network information sharing based on evolutionary game theory |
CN114257507B (en) * | 2021-12-22 | 2024-02-13 | 中国人民解放军国防科技大学 | Method for improving network information sharing level based on evolution game theory |
CN114844789A (en) * | 2022-04-20 | 2022-08-02 | 华中师范大学 | Community knowledge sharing evaluation method based on evolutionary game model |
CN114844789B (en) * | 2022-04-20 | 2023-05-26 | 华中师范大学 | Community knowledge sharing evaluation method based on evolution game model |
CN117592236A (en) * | 2023-12-05 | 2024-02-23 | 北京大数据先进技术研究院 | Data sharing network strategy evolution prediction method, device and product |
CN117592236B (en) * | 2023-12-05 | 2024-07-26 | 北京大数据先进技术研究院 | Data sharing network strategy evolution prediction method, device and product |
Also Published As
Publication number | Publication date |
---|---|
CN113724096B (en) | 2023-09-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113724096B (en) | Group knowledge sharing method based on public evolution game model | |
CN105654388B (en) | A kind of modeling method of dynamic social network Information Propagation Model | |
He et al. | Three-stage Stackelberg game enabled clustered federated learning in heterogeneous UAV swarms | |
CN111275174A (en) | A Game-Oriented Radar Countermeasure Strategy Generation Method | |
CN113191804A (en) | Optimal bidding strategy solving method | |
Hao et al. | Efficient and robust emergence of norms through heuristic collective learning | |
CN111445291B (en) | A method to provide dynamic decision-making for social network influence maximization problem | |
CN109858966A (en) | Cooperative method for promoting network community based on evolutionary game | |
Hao et al. | Achieving socially optimal outcomes in multiagent systems with reinforcement social learning | |
CN116228449A (en) | Method for analyzing online social network information propagation dynamics based on evolution game | |
Jin et al. | The convergence analysis of evolutionary dynamics for continuous action iterated dilemma in information loss networks | |
CN114048833B (en) | Multiplayer, large-scale incomplete information game method and device based on neural network virtual self-play | |
CN114792187B (en) | Crowd sensing team recruitment method and system based on willingness and trust double constraint | |
Deng et al. | The choice-decision based on memory and payoff favors cooperation in stag hunt game on interdependent networks | |
Chang et al. | Learning and evaluating human-like NPC behaviors in dynamic games | |
Li et al. | Two-level Q-learning: learning from conflict demonstrations | |
Luo et al. | Improved learning efficiency of deep Monte-Carlo for complex imperfect-information card games | |
Chen et al. | Using rewards reasonably: The effects of stratified-rewards in public goods game | |
Yang et al. | Effects of an update mechanism based on combinatorial memory and high-reputation learning objects on the evolution of cooperation | |
CN114844789B (en) | Community knowledge sharing evaluation method based on evolution game model | |
Kreminski | Toward better gossip simulation in emergent narrative systems | |
CN117217329B (en) | A federated learning method with drift correction and convergence acceleration | |
Li et al. | NFSP-PER: An efficient sampling NFSP-based method with prioritized experience replay | |
Lv et al. | MODELING and simulation of live streaming e-commerce information dissemination considering opinion leader in social networks | |
Bhattacharjee et al. | Optimizing Fantasy Sports Team Selection with Deep Reinforcement Learning |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |