A two-layer team-assembly model for invention networks

Since companies are exposed to rigid competition, they are seeking how best to improve the capabilities of innovations. One strategy is to capitalize on other companies’ knowledge in order to speed up their own innovations. However, acquiring other companies’ knowledge is not easy since they have different cultures and tacit knowledge. To tackle this problem, we need to consider interaction between companies and inventors because the actual providers of innovations are inventors and there is definitely a synergy between inventors and companies. This paper clarifies the dynamics between two networks whose nodes represent inventors and companies. By using patent data from the United States and Japan, the author extracted collaborations between inventors and companies and created two different projected networks where one of the nodes represented inventors and the other represented companies. The author discusses the impact of patents on networks and found that this impact was significantly affected by both inter-inventors’ and inter-companies’ collaborations. This means that it is important to consider layered networks. Finally, the author created a model to replicate a two-layered network. The proposed model could replicate the observed network well in terms of degree distributions. The proposed model could give us insights into the inter-company activities of innovations. (1) Inventors with many connections to other inventors have a greater possiblity of obtaining other connections in the future. (2) Companies that have the densely connected inventors have a greater possiblity of obtaining connections to other companies. (3) As most connections are inside companies, they have to retain many inventors to acquire inter-company connetions.