The effect of network structure on innovation initiation process: an evolutionary dynamics approach

TL;DR

This study uses an agent-based evolutionary dynamics model to examine how different network structures influence the initiation and success of innovations in firms, revealing that local networks favor radical innovations while non-local networks support incremental ones.

Contribution

It introduces a novel agent-based model to analyze the impact of various network structures on innovation processes, highlighting the differential effects on radical and incremental innovations.

Findings

Local networks outperform in radical innovation scenarios.

Non-local networks are more effective for incremental innovations.

Network structure influences innovation success depending on innovation frequency.

Abstract

In this paper we have proposed a basic agent-based model based on evolutionary dynamics for investigating innovation initiation process. In our model we suppose each agent will represent a firm which is interacting with other firms through a given network structure. We consider a two-hit process for presenting a potentially successful innovation in this model and therefore at each time step each firm can be in on of three different stages which are respectively, Ordinary, Innovative, and Successful. We design different experiments in order to investigate how different interaction networks may affect the process of presenting a successful innovation to the market. In this experiments, we use five different network structures, i.e. Erd\H{o}s and R\'enyi, Ring Lattice, Small World, Scale-Free and Distance-Based networks. According to the results of the simulations, for less frequent innovations like radical innovation, local structures are showing a better performance comparing to Scale-Free and Erd\H{o}s and R\'enyi networks. Although as we move toward more frequent innovations, like incremental innovations, difference between network structures becomes less and non-local structures show relatively better performance.