A study on coevolutionary dynamics of knowledge diffusion and social network structure
Introduction
The smooth transfer of knowledge in human groups is increasingly important to our society. Accordingly, in the last decades, knowledge transfer or diffusion has been extensively studied in various disciplines, especially in social psychology and in economics and management science. In social psychology, one research thread that is closely related to knowledge diffusion is on social cognition and social learning (Fiske & Taylor, 1991), which study communications and information processing in human groups by investigating the underlying cognitive processes. In economics and management, the studies on knowledge diffusion can trace back to the studies on the diffusion of innovations and technologies since the 1950s (Coleman et al., 1957, Rogers, 2003). However, it is with the flourishing of knowledge management since the 1990s, the explicit attention on knowledge transfer or diffusion has been widespread (Argote and Ingram, 2000, Ernst and Kim, 2002).
The studies on knowledge diffusion cover a wide range of research topics, e.g., the factors that affect knowledge transfer in organizations (Darr and Kurtzberg, 2000, Szulanski, 2000), the effects of knowledge transfer on organizational performance (Argote and Ingram, 2000, Cavusgil et al., 2003) and methods and tools to foster knowledge transfer within organizations (Bartol and Srivastava, 2002, Gurteen, 1999). More recently, with the rapid growth of network science (Barabasi, 2002), the studies on knowledge diffusion in social networks have attracted much research attention (Herie and Martin, 2002, Phelps et al., 2012).
Regarding knowledge diffusion in social networks, two research issues come to the fore, i.e. how knowledge is diffused in a social network and how the social network itself, in which knowledge is diffused, evolves.
For the former issue, enormous researches have been conducted to explore knowledge diffusion in the inter-organizational networks (Cowan and Jonard, 2004, Kim and Park, 2009, Paruchuri, 2010), in the intra-organizational networks (Boone and Ganeshan, 2008, Hansen, 1999, Mu et al., 2010, Reagans and McEvily, 2003), and in the multi-level networks that combine the internal and external links of an organization (Walter, Lechner, & Franz, 2007). In these researches, one key issue is to estimate the effects of the network properties on knowledge transfer. Typical research topics include the effect of the tie strength in the network on the knowledge transfer (Hansen, 1999), the effects of cohesion and range on knowledge transfer (Reagans & McEvily, 2003). The network structure or topology is one key property that has great impact on knowledge diffusion. Thus, knowledge diffusion on different types of network structures has been extensively studied, e.g. on grid-based or regular networks (Deroian, 2002, Ellison and Fudenberg, 1995), on scale-free networks (Amblard and Deffuant, 2004, Delre et al., 2004, Lin and Li, 2010, Stauffer and Sahimi, 2005, Tang et al., 2006, Tang et al., 2010), and on small-world networks (Cowan and Jonard, 2004, Delre et al., 2007, Eslami et al., 2013, Kim and Park, 2009). More recently, knowledge generation and transfer on collaborative hypernetworks have also been taken into account (Liu et al., 2014, Yang et al., 2015).
The latter issue is associated with the research efforts on the topological structures and dynamics of networks, which have attracted great attention since Watts and Strogatz (1998)’s work on the small-world networks and Barabasi and Albert (1999)’s work on the scale-free networks. Inspired by these endeavors, various researches have also studied the structure and evolution of the knowledge-intensive social network, especially the collaboration networks of different types. The scientific collaboration networks have been extensively studied; and typical examples include (Barabasi, 2002, Demirkan et al., 2013, Guimera et al., 2005, Lee et al., 2010, Tang et al., 2013). The technological innovation networks are another type of well-discussed networks. There have been plenty of researches on the intra-firm, inter-firm, national-wide, and international innovation networks, e.g. (Cowan et al., 2007, De Prato and Nepelski, 2014, Hermans et al., 2013, Lee, 2010, Lovejoy and Sinha, 2010, Ter Wal, 2014). In these endeavors, the behavioral and social factors that affect the network structure are discussed and the mechanisms that underlie the evolution of the network are explored.
The prior efforts on the dynamics of knowledge diffusion on social networks and those on the dynamics of the social networks are both fruitful. However, one limitation of the aforementioned efforts lies in that the two dynamics are studied separately. To address this, there is a need to study the coevolution of network and knowledge. Such coevolutionary dynamics, which combines the topological evolution of the network with the knowledge dynamics in the network nodes, is a recent focal research direction in the field of complex networks (Gross & Blasius, 2008). Various research efforts have been reported, especially on the coevolution of network and opinions, e.g. (Holme and Newman, 2006, Gil and Zanette, 2006, Luo et al., 2014, Su et al., 2014, Vazquez, 2013). The studies on the coevolution of network and knowledge are also reported, although they are relatively fewer. For example, based on the transactive memory theory (Wegner, 1995), Palazzolo et al. give a thorough analysis on the coevolution of the social and knowledge networks through agent-based modeling (Palazzolo, Serb, She, Su, & Contractor, 2006). Roth and Cointet empirically examine the coevolution of social relation network and the social-semantic network (Roth & Contet, 2010). In a similar spirit, Wang and Groth present a framework to measure the dynamic bi-directional influence between communication content and social networks (Wang & Groth, 2010). Iñiguez et al. study the spreading of scientific concepts in an adaptive network and find that scientifically-sound concepts are difficult to spread as opposing individuals tend to form close communities that prevent opinion consensus (Iñiguez, Taǵ’ueña Martĺnez, Kaski, & Barrio, 2012).
The previous contributions on the coevolutionary adaptive networks show a promising research direction. Especially, further attentions on the coevolution of network and knowledge are deserved, as the current researches are largely insufficient due to the inherent complexity of this subject of inquiry. As an attempt complementary to the prior contributions, we in this work set the research background to the community-based human organizations in the real- and/or cyber-space, such as communities of practice (Wenger & Snyder, 2000), knowledge-building communities (Scardamalia & Bereiter, 1994), and many types of scientific and technological communities. For simplicity, we uniformly call these different communities as “knowledge communities”. Knowledge communities have been playing an important role in todays knowledge creation and diffusion (Lee and Cole, 2003, Hussler and Rondé, 2007, Luo et al., 2013). The communication networks underlying these communities are intrinsically dynamic; and this dynamics of network is inherently interwoven with the dynamics of knowledge transfer in the network, owing to the bidirectional influences between network and knowledge. Thus, we in this work propose an agent-based model to explore such co-evolutionary dynamics of network and knowledge.
The key assumption underlying our model is that the “knowledge distance” is a critical factor that impacts on both the performance of knowledge transfer and the stability of communicative link between two agents. In professional communications, knowledge transfer between two highly-heterogeneous agents may often be ineffective due to big knowledge gap; meanwhile, the transfer of knowledge may neither be favorable between two highly-homogenous agents because high knowledge-similarity decreases the benefit from knowledge-exchange.Thus, as pointed out by some scholars in cognitive-psychology and pedagogy, knowledge transfer is most effective when the “knowledge distance” is neither too large nor too small (Piaget, 1976, Scholl, 1996, Van Der Vegt and Bunderson, 2005, Witte and Davis, 1996). On the basis of this assumption, the agents are designed to seek for better chances of effective knowledge exchange; and the links that connect to highly-heterogeneous or highly-homogeneous neighbors become unstable. With the proposed model to reflect a simplified scenario of the knowledge-transfer network in real-world communities, we aim to explore possible dynamic patterns of network evolution and knowledge transfer driven by knowledge distance. To the best of our knowledge, this issue has not been well studied in literature, although there have been outstanding works on the impacts of knowledge heterogeneity or diversity on the performance of problem-solving and knowledge production, e.g. (Guimera et al., 2005, Page, 2008).
The remainder of the paper is organized as follows. In Section 2, the agent-based simulation model is proposed. Subsequently, in Section 3, we analyze the simulation result. In Section 4, the key findings are summarized and discussed; and lastly the whole paper is concluded in Section 5.
Section snippets
General description of the model
As stated in Introduction, this paper proposes an agent-based simulation model to study the knowledge-distance-driven coevolutionary dynamics of knowledge networks. Before formulating the model, we give a further explanation on its enabling ideas.
We consider the proposed model as a special case of the “coevolutionary adaptive networks” researches. Thus, the proposed model follows the overall structure of the interaction between human dynamics and network evolution (Gross & Blasius, 2008), as
Setting the simulation conditions
Using the previously-described model, we conduct computational simulations to investigate the coevolutionary dynamics of network and knowledge. Here we set the basic conditions for executing the simulations and the basic measurements evaluate the simulation results.
We consider a population of N agents with M edges in the network. In addition, edges are undirected and duplicate links are not allowed. Each agent has a 5-category knowledge vector (i.e. l = 5), initialized by setting
Summarization of simulative results
Although the model presented in this paper is primitive and doubtlessly oversimplified to cater for the complexity of knowledge diffusion in the real-world social contexts, this work may be helpful for improving our understanding on the dual dynamics of network and knowledge. The major results from our simulations can be summarized as follows:
(1) As the parameter d actually reflects the “gap” that hampers the transfer of knowledge between individuals, it is therefore straightforward that the
Conclusion
The results obtained in this work are helpful for enrich our knowledge on the coevolutionary nature of the structural change of the network and the knowledge-diffusion activities on the networks. In particular, they may shed some light on the underlying social forces and mechanisms for the small-world dynamics of the knowledge-intensive social networks.
In Watts and Strogatz (1998)’s classic work, they have shown a typical structure of the small-world network, in which the network is comprised
Acknowledgements
This work is partly supported by National Natural Science Foundation of China under Grant Nos. 71401024, 71371040, respectively. The authors are grateful for constructive comments from the anonymous reviewers, which are of great help to improve the paper.
References (92)
- et al.
The role of network topology on extremism propagation with the relative agreement opinion dynamics
Physica A
(2004) - et al.
Knowledge transfer: A basis for competitive advantage in firms
Organizational Behavior and Human Decision Processes
(2000) - et al.
Knowledge flows and the geography of networks: A strategic model of small world formation
Journal of Economic Behavior & Organization
(2009) - et al.
The impact of small world on innovation: An empirical study of 16 countries
Journal of Informetrics
(2010) - et al.
Network structure and the diffusion of knowledge
Journal of Economic Dynamics and Control
(2004) - et al.
Investigation of partner similarity dimensions on knowledge transfer
Organizational Behavior and Human Decision Processes
(2000) Formation of social networks and diffusion of innovations
Research Policy
(2002)- et al.
Global production networks, knowledge diffusion, and local capability formation
Research Policy
(2002) - et al.
Coevolution of agents and networks: Opinion spreading and community disconnection
Physics Letters A
(2006) - et al.
Niches and networks: Explaining network evolution through niche formation processes
Research Policy
(2013)
The impact of cognitive communities on the diffusion of academic knowledge: Evidence from the networks of inventors of a french university
Research Policy
Communities, knowledge creation, and information diffusion
Journal of Informetrics
Scale-free network provides an optimal pattern for knowledge transfer
Physica A
Technological overlap and inter-firm cooperation: Implications for the resource-based view of the firm
Research Policy
Absorptive and disseminative capacity: Knowledge transfer in intra-organization networks
Expert Systems with Applications
Renormalization group analysis of the small-world network model
Physics Letters A
Social and semantic coevolution in knowledge networks
Social Networks
The process of knowledge transfer: A diachronic analysis of stickiness
Organizational Behavior and Human Decision Processes
Disseminative capacity, organizational structure and knowledge transfer
Expert Systems with Applications
Estimating the effect of organizational structure on knowledge transfer: A neural network approach
Expert Systems with Applications
Knowledge transfer between and within alliance partners: Private versus collective benefits of social capital
Journal of Business Research
Knowledge diffusion in the collaboration hypernetwork
Physica A: Statistical Mechanics and its Applications
Technological acquisitions and the innovation performance of acquiring firms: A longitudinal study
Strategic Management Journal
Linked: How everything is connected to everything else and what it means for business, science, and everyday life
Emergence of scaling in random networks
Science
Encouraging knowledge sharing: the role of organizational reward systems
Journal of Leadership and Organization Studies
Network-independent partner selection and the evolution of innovation networks
Management Science
Where do small worlds come from? Industrial and corporate change
Knowledge acquisition and transfer among engineers: Effects of network structure
Managerial and Decision Economics
Structural holes and good ideas
American Journal of Sociology
Tacit knowledge transfer and firm innovation capability
Journal of Business and Industrial Marketing
The diffusion of an innovation among physicians
Sociometry
Bilateral collaboration and the emergence of innovation networks
Management Science
Mixing beliefs among interacting agents
Advances in Complex Systems
Diffusion dynamics in small-world networks with heterogeneous consumers
Computational and Mathematical Organization Theory
Exploring the role of network characteristics, knowledge quality, and inertia on the evolution of scientific networks
Journal of Management
Global technological collaboration network: Network analysis of international co-inventions
The Journal of Technology Transfer
Evolution of networks: From biological nets to the internet and www
Word-of-mouth communication and social learning
Quarterly Journal of Economics
Effect of collaboration network structure on knowledge creation and technological performance: The case of biotechnology in canada
Scientometrics
Social cognition
Social cohesion
Annual Review on Sociology
Coevolutionary dynamics of opinions and networks: From diversity to uniformity
Physical Review E
Adaptive coevolutionary networks: A review
Journal of the Royal Society Interface
Cited by (69)
Finite-time optimal feedback control mechanism for knowledge transmission in complex networks via model predictive control
2022, Chaos, Solitons and FractalsMulti-level simulation of the physical, cognitive and social
2022, Computers, Environment and Urban SystemsCitation Excerpt :One or more diffusion and network models need to be chosen for a diffusion application. Different diffusion models may be relevant, depending on the influence of interest, e.g., opinion (Deffuant et al., 2002), rumour (Kimura et al., 2009), knowledge (Luo et al., 2015). In our applications, we chose the LT model for the spread of influence to evacuate (similar to (Hasan & Ukkusuri, 2011)) and the IC model to represent an information diffusion process (similar to (Qi, Liang, Wang, & Cheng, 2018)).
Formation mechanism for integrated innovation network among strategic emerging industries: Analytical and simulation approaches
2021, Computers and Industrial EngineeringEnhancing the competitiveness of multi-agent cooperation for green manufacturing in China: An empirical study of the measure of green technology innovation capabilities and their influencing factors
2020, Sustainable Production and ConsumptionCitation Excerpt :Vick et al. (2015) pointed out that network feedback is conducive to achieving external advantages (Vick et al., 2015). Luo et al. (2015) suggested that multi-agent cooperation driving depends on network structure (Luo et al., 2015). Focusing on Chinese cooperative innovation projects, Wu et al. (2017) found that specific investments indirectly influenced cooperation innovation performance through relational trust (Wu et al., 2017).
Exploring the dynamic growth mechanism of social networks using evolutionary hypergraph
2020, Physica A: Statistical Mechanics and its ApplicationsKnowledge growth in university-industry innovation networks – Results from a simulation study
2020, Technological Forecasting and Social Change