多主体博弈视角下供应链网络幂律分布研究 ——以中国新能源汽车零部件行业为例

Research on Power Law Distribution of Supply Chain Network from the Perspective of Multi-agent Game ——Taking China's new energy vehicle parts industry as an example

摘要:

为了探究供应链网络中幂律分布演化趋势的规律，针对演化博弈理论在幂律分布的突变现象和平稳分布方面研究的不足，在选择-漂移动力学模型的基础上对突变效用矩阵和节点策略选择概率的变化进行改进分析。通过得到的效用矩阵优化值和行为策略转变概率，结合中国知网提供的新能源汽车零部件数据进行实验仿真，以2018年至2021年数据集为主要研究目标，分析在改进选择-漂移动力学模型的作用下供应链网络的幂律分布特征的变化情况。结果表明：不同的行为策略占比会影响网络幂律分布的演化趋势；存在一个最佳调节参数使得网络在演化过程中避免幂律分布的相变现象发生，并使网络的分布更加平稳。

Abstract:

In order to explore the evolution trend of power-law distribution in supply chain networks and address the shortcomings of evolutionary game theory in studying the phenomena of power-law distribution mutations and stable distributions, an improved analysis was conducted on the changes in mutation utility matrix and node strategy selection probability based on the selection drift dynamics model. By using the optimized values of the utility matrix and the probability of behavior strategy transition obtained, combined with the new energy vehicle component data provided by China National Knowledge Infrastructure, experimental simulations were conducted. The main research objective was to analyze the changes in the power-law distribution characteristics of the supply chain network under the influence of an improved selection drift dynamics model, with the dataset from 2018 to 2021. The results indicate that the proportion of different behavioral strategies will affect the evolutionary trend of the power-law distribution in the network; There exists an optimal tuning parameter that prevents the occurrence of power-law distribution phase transitions during the evolution of the network, and makes the distribution of the network more stable.