Open Data in the Digital Economy: An Evolutionary Game Theory Perspective

TL;DR

Problem: open data governance with regulators; Approach: a three-party evolutionary game with replicator dynamics; Payoffs incorporate costs $C_1$, $C_2$, $C_3$, values $V_1$, $V_2$, benefits $R_1$, $R_2$, losses $L_1$, $L_2$, rewards $F$, regulator income $P$, and data-mining capability $\alpha$. Findings: eight fixed points exist; the internal equilibrium is not ESS; four pure ESS emerge under parameter regimes defined by inequalities such as $C_1 < 2F + R_1 + L_1$, $C_2 < 2F + R_1 + L_2 + \alpha V_1$, and $C_3 < P$; and numerical simulations confirm analytic predictions and show $\alpha$ enhances cooperation. Significance: informs governance strategies to promote open data quality and sustainable digital economy.

Abstract

Open data, as an essential element in the sustainable development of the digital economy, is highly valued by many relevant sectors in the implementation process. However, most studies suppose that there are only data providers and users in the open data process and ignore the existence of data regulators. In order to establish long-term green supply relationships between multi-stakeholders, we hereby introduce data regulators and propose an evolutionary game model to observe the cooperation tendency of multi-stakeholders (data providers, users, and regulators). The newly proposed game model enables us to intensively study the trading behavior which can be realized as strategies and payoff functions of the data providers, users, and regulators. Besides, a replicator dynamic system is built to study evolutionary stable strategies of multi-stakeholders. In simulations, we investigate the evolution of the cooperation ratio as time progresses under different parameters, which is proved to be in agreement with our theoretical analysis. Furthermore, we explore the influence of the cost of data users to acquire data, the value of open data, the reward (penalty) from the regulators, and the data mining capability of data users to group strategies and uncover some regular patterns. Some meaningful results are also obtained through simulations, which can guide stakeholders to make better decisions in the future.

Figures (5)

• Figure 1: The game relationship among data providers, users, and regulators. Each of them has two different strategies to choose from, and the combination of strategies among them will have an impact on their payoffs.
• Figure 2: The Evolutionary Trajectories of Three Players. The $x$-axis and the $y$-axis are set as the time and the cooperation ratio separately. (a) The parameters are set as $C_1 = 8, C_2 = 10, C_3 = 6, \alpha = 0.7, V_1 = 5, V_2 = 3, R_1 = 2, L_1 = 2, L_2 = 2, P = 8, R_2 = 5$, and $F = 3$ to follow the conditions for (0, 0, 1) to become ESS. (b) The parameters are set as $C_1 = 15, C_2 = 10, C_3 = 6, \alpha = 0.8, V_1 = 10, V_2 = 8, R_1 = 4, L_1 = 2, L_2 = 5, P = 10, R_2 = 8$, and $F = 3$ to satisfy the condition for (0, 1, 1) to become ESS. (c) The parameters are set as $C_1 = 4, C_2 = 13, C_3 = 6, \alpha = 0.2, V_1 = 8, V_2 = 5, R_1 = 3, L_1 = 5, L_2 = 1, P = 7, R_2 = 6$, and $F = 3$ to obey the condition for (1, 0, 1) to become ESS. (d) The parameters are set as $C_1 = 8, C_2 = 6, C_3 = 7, \alpha = 0.7, V_1 = 10, V_2 = 8, R_1 = 4, L_1 = 4, L_2 = 3, P = 9, R_2 = 5$, and $F = 4$ to meet the two different conditions for (1, 1, 1) to become ESS. It can be seen that the results of the numerical simulations are in good agreement with our previously derived theory.
• Figure 3: Effects of $\alpha$, $C_2$, $V_1$, and $F$ on the fraction of the data users choosing to acquire data. Here, we fix $C_1=6$, $C_3=5$, $V_2=1$, $R_1=2$, $L_1=5$, $L_2=7$, $P=3$, $R_2=4$. Regarding the concerned $C_2$, $V_1$, and $F$, we have (a) $C_2= [3.4, 3.6, 3.8, 4.0]$, $V_1=2$, $F=1$. (b) $C_2=4$, $V_1=[1.5, 2.5, 3.5, 4.5]$, and $F=1$. (c) $C_2=4$, $V_1=3$, and $F=[0.3, 0.7, 1.1, 1.5]$. The $x$ and $y$ axes are set as the data mining capability $\alpha$ and the fraction of the data users adopting to acquire data, respectively. It can be observed that a larger value of data mining capability $\alpha$ results in a greater cooperation level of the data user.
• Figure 4: Heat maps of the fraction of the data user performing to acquire data considering $V_1$ and $C_2$. Here, we set $V_1$ and $C_2$ in $[1,5]$ and $[3,7]$ respectively, and fix $C_1=6$, $C_3=5$, $V_2=1$, $R_1=2$, $L_1=5$, $L_2=7$, $P=3$, $R_2=4$, and $F=1$. Besides, we have (a) $\alpha=0.30$, (b) $\alpha=0.60$, and (c) $\alpha=0.90$. The color bar is in the range $[0, 1]$ as shown in each subgraph. For the larger data mining capability $\alpha$, the ratio of the data user choosing to acquire data can be enhanced by decreasing the cost of the data users acquiring data $C_2$ or increasing the value of high quality data $V_1$.
• Figure 5: 3D traces of $x$, $y$, and $z$ in the evolution processes. Here, the fixed parameters are $C_1=6$, $C_3=5$, $V_2=1$, $R_1=2$, $L_1=5$, $L_2=7$, $P=3$, and $R_2=4$. $\alpha\in[0.30, 0.60, 0.90]$ is set for the cross simulation. In (a), (b), and (c), we set $C_2=[3.4, 3.6, 3.8, 4.0]$, $V_1=3$, and $F=1$. In (d), (e), and (f), we set $V_1=[1.5, 2.5, 3.5, 4.5]$, $C_2=4$, and $F=1$. In (g), (h), and (i), we set $F=[1, 2, 3, 4]$, $C_2=4$, and $V_1=3$. Each curve stands for a changing trajectory of $x$, $y$, and $z$, of which the initial values are all $0.5$. The 3D traces exhibit that an increase in the data mining capability $\alpha$ will provide more relaxed conditions for the whole group to adopt cooperative strategies.

Theorems & Definitions (4)

• Theorem 1
• proof
• Theorem 2
• proof