基于Agent仿真的个体有限理性与众包绩效的关系研究

如何筛选有限理性参与者是企业实施众包的关键问题。通过引入个体有限理性,扩展了经典的NK模型,构建了模拟竞赛式众包问题解决过程的多主体仿真模型,研究了个体有限理性水平、个体有限理性系统化程度以及个体有限理性水平标准差对众包绩效的影响。仿真结果显示,个体有限理性水平对提高众包绩效有显著的正向影响,尤其是在任务复杂性较高的情况下;个体有限理性系统化程度和个体有限理性水平标准差对众包绩效也有正向影响,但依赖于个体有限理性水平。当企业有一个复杂性较高的任务时,如果追求所有方案的整体改善,企业需要招募有限理性水平较高的个体,并组成有限理性系统化程度较高且有限理性水平标准差较大的群体;如果追求少数优质方案,企业需要招募有限理性水平较高的个体,并组成有限理性系统化程度较低且有限理性水平标准差较大的群体。

Individual Bounded Rationality and Crowdsourcing Performance: An Agent-based Simulation

The advancement of internet makes firms able to use collective intelligence to solve complex problems with the help of crowdsourcing by combining knowledge of outside participants. However, compared to inside employees, the bounded rationality level of outside participants is relatively low, which becomes a major concern of the firms. Therefore, how to recruit participants is a key issue for firms to implement crowdsourcing. To solve the problem, by introducing individual bounded rationality, an agent-based model that simulates problem solving process of tournament-based crowdsourcing is constructed by extending the NK model to explore the effects of the level of bounded rationality, the systematization of bounded rationality and the standard deviation of bounded rationality level on crowdsourcing performance. The results of simulation experiments suggest that the level of bounded rationality has a significant positive effect on crowdsourcing performance, especially in cases where the task complexity is high. The systematization of bounded rationality and the standard deviation of bounded rationality level also have a positive effect on the crowdsourcing performance, but they depend on the level of bounded rationality. Therefore, when there is a higher complexity task, and if the firm pursues the overall improvement of all solutions, the best strategy is to recruit individuals with higher bounded rationality level and form a group with higher systematization of bounded rationality and higher standard deviation of bounded rationality level; but if the firm pursues a small number of high-quality solutions, the best strategy is to recruit individuals with higher bounded rationality level and form a group with lower systematization of bounded rationality and higher standard deviation of bounded rationality level.