Analysis of some discrete-automaton models of collective behavior

We analyze the phenomenon of collective behavior using a discrete-automaton model. This model is similar to those used in modern computational biology to explain dynamical processes in gene networks. Within the proposed framework we study a multiagent system, which contains not only conformists who make their decisions depending on other agents’ opinions but also agents with fixed decisions (the, so called, «agitators» and «loyalists»). We solve the problems of of small number of agitators (loyalists) placement, which forces the system to transit to a state where more than a half of agents is active (inactive). Our computational method based on SAT-approach made it possible to solve corresponding problems for systems with a hundred vertices with PC.