Helping agents to enter and exploit artificial institutions

In this paper, we are considering agents that act to satisfy their interests in systems composed of institutions. These institutions are unknown to the agents at design time and are endowed with actions (called institutional actions) with functions that bring consequences in the environment that match the interests of the agents. However, existing models of artificial institutions do not make these consequences explicit. Agents need to know the institutions and the consequences of carrying out institutional actions before acting in the systems. Considering this problem, this paper proposes to instrument institutions with a purpose model that makes the environmental consequences of carrying out institutional actions explicit. In addition, the paper also presents some algorithms that agents can use to consider this information in their reasoning and deliberation process. The model is presented, discussed and implemented using the JaCaMo framework, highlighting its advantages for agents when reasoning about the consequences of their actions both in the institution and in the environment.