The fundamental idea underlying agent-oriented programming (AOP)  is that agents are modeled in terms of their “mental states” (consisting in this work of beliefs, plans, capabilities, and intentions), both by other agents and by their designers and programmers. An agent program gives the agent’s initial mental state and rules describing a transition function: given an agent’s current state and input, the rules specify its new state and output. An agent’s communicative acts, then, are actions that affect the mental states of the agents involved, just as physical actions affect the agent’s physical environment. We present and discuss a new agent-oriented programming language, PLACA, a descendant of AGENTIO . Unlike AGENTO, PLACA capitalizes on agents’ planning abilities. Assuming that all agents have at least elementary planning abilities, PLACA agents can make high-level requests of each other without worrying about how these requests will be carried out. We describe PLACA, show a short example program, briefly describe a PLACA interpreter, and discuss the mental states of agents.
Thomas, S. R. (1995). The PLACA agent programming language. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 890, pp. 355–370). Springer Verlag. https://doi.org/10.1007/3-540-58855-8_23