机器学习、策略选择与供应链金融合谋行为:基于演化博弈视角的研究

Machine Learning,Strategy Choice and Supply Chain Finance Collusion:An Evolutionary Game Perspective

在机器学习情境下,以供应链金融合谋行为为研究内容,以金融机构(治理方)、中小企业(主谋方)、核心企业或支持性企业(同谋方)为研究对象,分析金融机构科技投入策略与合谋者行为策略选择之间的相互影响。结果表明,在融资申请阶段,当中小企业提供虚假资质获得贷款与提供真实资质获得贷款的收益之差低于其在提供虚假资质进行贷款时所付出的审查逃避费用、给予核心企业或支持性企业附加承诺之和时,供应链金融系统将演化至金融机构选择只应用机器学习、中小企业选择提供真实资质、核心企业或支持性企业选择与中小企业合作的稳定状态;在合同履行阶段,合谋者为逃避审查而付出的成本及其合谋行为被发现并对其定罪后所付出的各类成本远高于其违约收益,此时供应链金融系统将演化至金融机构选择只应用机器学习,中小企业、核心企业或支持性企业选择守约的稳定状态。

In the context of machine learning,takes supply chain finance collusion as the research content and financial institutions(governance party),small and medium-sized enterprises(mastermind party),core enterprises or supporting enterprises(co-conspirator party)as the research object,analyzing the interaction between financial institutions’science and technology investment strategy and conspirators’behavior strategy choice.The results show that,in the financing application stage,the difference between smes obtaining loans by providing false qualifications and obtaining loans by providing real qualifications is lower than the sum of regulatory evasion costs and additional commitments given to core enterprises or supporting enterprises when providing false qualifications.At this point,the supply chain financial system will evolve to a stable state where financial institutions choose to only apply machine learning,smes choose to provide real qualifications,and core enterprises or supporting enterprises choose to cooperate with smes.In the phase of the contract,colluding in evade regulation and to pay for the cost and the conspiracy was found after their conviction is paying for all kinds of costs are much higher than the return on default,the supply chain in the financial system will evolve to financial institutions to choose only the application of machine learning,or supporting small and medium-sized enterprises,the core enterprise choose keep the steady state.