Multiagent modelling for telecommunication market structure evolution

In our paper we investigate how network effects influence the dynamics of oligopoly on mature telecommunication markets. The model covers four categories of agents: telecommunication operators, customers joined into housholds and market regulator. The market state aware firms use reinforcement learning to set prices, discriminating between inter and intra network calls. The customers minimize the cost of fulfilling theirs communication needs on the household level, which are driven by a network of interactions between them. The market regulator sets the number of companies on the market and interconnection charges level. We find that the networked nature of the demand side has significant influence on long run price levels. Market without networking (no connections other than those between customers belonging to the same household) is found to reach the Bertrand equilibrium. However, in the real world case, where networking effects are present, the price level is well above marginal costs. Consequently, we describe the influence of customer interaction network topology and market regulator decisions on the long run price level.