Dynamic trust game model between venture capitalists and entrepreneurs based on reinforcement learning theory

Abstract

It has been extensively demonstrated that the trust between venture capitalists and entrepreneurs plays an important role in improving the success rate of entrepreneurship. However, little research has been performed concerning the dynamic nature of the trust between venture capitalists and entrepreneurs. Therefore, we attempt to reveal the motivations and process of trust between VCs and entrepreneurs from the perspective of dynamics. First, the theoretical model of dynamic trust was proposed. Second, reinforcement learning was introduced to design the multi-stage game model of dynamic trust, and the algorithm was illustrated. Reinforcement learning theory can reveal the reasons for dynamic changes in trust from a psychological perspective. Finally, the action strategies of venture capitalists and entrepreneurs were simulated. We observe that the dynamic characteristics of trust between venture capitalists and entrepreneurs are determined by entrepreneurs’ efforts. Venture capitalists can prevent investment risks and maintain trust stability when fewer funds are invested in the early stage and when more funds are invested in the later stage. Since the dynamic process of trust is also the process of learning about the other, the dynamic process of trust is non-linear, and the relationship is adaptive.