Developing Intentional Systems with the PRACTIONIST Framework

Agent-based systems have become a very attractive approach for dealing with the complexity of modern software applications and have proved to be useful and successful in some industrial domains. However, engineering such systems is still a challenge due to the lack of effective tools and actual implementations of very interesting and fascinating theories and models. In this area the so-called intentional stance of systems can be very helpful to efficiently predict, explain, and define the behaviour of complex systems, without having to understand how they actually work, but explaining them in terms of some mental qualities or attitudes, rather than their physical or design stance. In this paper we present the PRACTIONIST framework, that supports the development of PRACTIcal reasONIng sySTems according to the BDI model of agency, which uses some mental attitudes such as beliefs, desires, and intentions to describe and specify the behaviour of system components. We adopt a goal-oriented approach and a clear separation between the deliberation phase and the means-ends reasoning, and consequently between the states of affairs to pursue and the way to do it. Moreover, PRACTIONIST allows developers to implement agents that are able to reason about their beliefs and the other agents' beliefs, expressed by modal logic formulas.