The possibility of modelling and simulating entities that perceive their environment, reason upon those perceptions and certain background of knowledge, and then act upon that environment (i.e. as Agents in AI) opens the door to a number of extensions to the established uses of simulation. Economic modelling can, with agents, take into account new and, arguably, more precise characterizations of human beings. This way of modelling economic agents may become an alternative to more traditional mathematical models employed in economics. Those traditional descriptions of human beings normally exclude, for the sake of tractability, fundamental aspects such as qualitative descriptions of the agents' goals and intentions, beliefs and other attributes of human reasoning (e.g. bounded rationality).

However, we do not aim at replacing traditional modelling techniques with a new modelling language, but instead we try to integrate well-known approaches (DEVS modelling and its extensions) together with a new approach that allow us to include, in a simulation model, those reasoning-acting entities: the agents. By appealing to engineering techniques from Artificial Intelligence and Logic Programming, we are capable of producing a light reason engine to simulate each agent observe-reason-act process. With as many agent engines as required for the particular application and with the traditionally conformed simulator for the environment in which those agents are placed (the main simulator), we create a federation (In HLA's sense) which serves as the simulator for the whole multi-agent system.

Universidad de Los Andes
Cesimo - SUMA
Galatea Group
Mérida, Venezuela