The Agent Communication Simulation Based on the Ego State Model of Transactional Analysis

Inter-agent interaction plays an important role in cyberspace, which has recently been receiving attention as a space for social interaction. These interactions can occur between agents as well as between agents and people. However, for this, agents must be developed to the extent that they do not differ noticeably from human beings; that is, computers must be able to exhibit the same behaviors as humans. To achieve this, we developed a simulation in which human communication is replaced by a card game that corresponds to the ego state model of transactional analysis (TA). The agent communication module that we developed expresses agent-agent transactions based on the ego state model, as adapted to computer agents. The agents, the attacker and defender, continue the simulation with the cards that have messages and personalities. To win this simulation, the agents should find out their opponent’s intentions and show a card that is suitable for their opponent’s sample word or personality.