The coevolution of endogenous knowledge networks and knowledge creation

Abstract

This paper presents a simulation model of the coevolution of an endogenous network of agents and the knowledge they create. The model is able to reconcile the diferent scenarios of knowledge creation in networks of previous empirical studies: positive, none, or negative relation between collaboration and performance. Thus, it can be used to describ e a stylized process of knowledge creation through collaborations. Moreover, the model highlights some key aspects of the process which are central when trying to promote knowledge creation. First of all, in some cases collaboration can be harmful for knowledge creation. Also, that collaboration may not boost knowledge creation, but rather the better performing agents attract more collaborators. Finally, when collaborators are chosen based on previous history rather than attractiveness, the performance of agents is reduced. These key aspects can be of high importance for the design of policies aimed to improve knowledge creation processes.