Threshold agent networks: an approach to modelling and simulation

Threshold agent networks (TANs), a discretized version of threshold or neural networks, are proposed as alternative platforms to sequential dynamical systems for modelling computer simulations. It is argued that each model has its own advantages and disadvantages compared to the other and the choice on each occasion should depend on the particular characteristics of the application at hand. Some results on the expressive power and the limitations of TANs are presented. Finally, equivalence classes of TANs that are introduced based on characteristics of their state spaces are studied in detail and upper bounds are given on their cardinalities.