Experimental economic research of communications networks
Ural Federal University, Russia, 620002, Ekaterinburg, Mira street, 19, phone: +7 (343) 375-44-741 pp. (accepted)
Relationship between market participants may be presented in the form of communications networks. However, the network aspects of economic activity has only recently attracted the attention of researchers and therefore not reflected in economic theory. The aim of this work is to study communications networks in localized systems by economies experimental method.
Communications networks have been received as a result of two types of experiments: in the short-term (1 hour) business games and the long-term (2 months) real activity of the participants in the experimental part of the economic system.
The study of the data is carried out using the methodology of social network analysis (SNA), which can be successfully used for the analysis of economic networks. Graphs of links are based on the matrix of agents communications networks. Since the communications networks are directed, to study were selected the following main parameters: the number of nodes and edges, density, diameter, clustering coefficient, transitivity and the input and output indices for Freeman centralization.
All obtained values were compared with the corresponding parameters of the Bernoulli random graphs with similar density, and their absolute and relative deviations were calculated. For a qualitative comparison of enterprise networks of two studied types, their absolute values of the network parameters and the relative deviation from the parameters of the Bernoulli random graphs were averaged. Matches or deviation from random graphs allow us to determine the properties of studied graphs.
As a result, there was revealed contrast studied graphs from graphs Bernoulli more than 22%, as well as the conclusions of the difference between the structures of communication networks of the two species studied were found: network of game is closer in performance to the network formed by random graphs Bernoulli than network by long-term real activity of students . This is due to the different nature of these networks. The findings determine the direction of future more detailed research and can provide enough material for an understanding of the phenomenon of autopoiesis (self-reproduction) in closed circuits of economic exchange.
Present study was carried out under financial support of the Russian Fund of Fundamental Research grant № 15-06-04863 “Mathematical models of local payment system lifecycles”.