DPG Phi
Verhandlungen
Verhandlungen
DPG

Berlin 2018 – scientific program

Parts | Days | Selection | Search | Updates | Downloads | Help

SOE: Fachverband Physik sozio-ökonomischer Systeme

SOE 17: Social Systems, Opinion and Group Dynamics II

SOE 17.1: Talk

Wednesday, March 14, 2018, 18:15–18:30, MA 001

How do firms collaborate? A data-driven model — •Giacomo Vaccario1, Mario V. Tomasello2, Claudio J. Tessone3, and Frank Schweitzer11ETH, Zurich, CH — 2E&Y, Zurich, CH — 3UZH, Zurich, CH

How do firms collaborate? To address this question, we propose an agent-based model that replicates two important processes in firm collaborations: i) The selection of the collaborators and ii) The exchange of knowledge. To calibrate our model, we reproduce by computer simulations first the observed collaboration network and secondly the knowledge exchange. For the former, we estimate the collaboration probabilities among firms that best match the empirical network. For the latter, we embed the firms in a multidimensional knowledge space where their positions represent their expertise along technological dimensions. We assume that firms exchange knowledge only while collaborating and approach each other in the knowledge space at a rate mu for the duration of a collaboration tau. We estimate these two parameters by comparing simulated and observed knowledge distances. We find that the average collaboration lasts around two years and that the knowledge transfer occurs at a low rate. In other words, a firm’s position hardly changes during collaboration and is not a consequence of its collaborations. Finally, we introduce a collaboration efficiency measure, that is the distance traveled by the firms in the knowledge space divided by the number of collaborations. We find that the model configuration that best reproduce the empirical data is close to the optimal configuration according to the introduced efficiency measure.

100% | Mobile Layout | Deutsche Version | Contact/Imprint/Privacy
DPG-Physik > DPG-Verhandlungen > 2018 > Berlin