Practionist.org

:. Features

We have included several features into the PRACTIONIST Runtime & Framework in order to support the development of agents with built-in BDI capabilities, with as less effort as possible. Some of them are:

a prolog-like language to represent and express agents’ beliefs. Such a language relies on a belief logic, which guarantees that at any moment the set of beliefs are compliant with the doxastic logic (KD45);
ability to dynamically build plans in order to pursue goals. These planning capabilities are based on a backward search algorithm in the states space;
the deliberation process by which some desires may be promoted into intentions under some conditions. In other words, when an agent wishes to pursue a goal, this desire can become an intention only if the agent really believes that such a goal is both possible and not inconsistent with active goals;
the means-ends reasoning, which lets the agent figure out some both practical and applicable plans that are able to manage an intended goal, an incoming perception, or a belief update;
the ability to concurrently execute plans, which encompass the actual acts of agents, such as updating and querying beliefs, doing actions, desiring or intenting to pursue goals, waiting for incoming perceptions or successful or failing goals, and so forth;
the opportunity to define and use several perceptors and the corresponding ability to select the proper one that is able to handle an incoming perception;
the opportunity to define and use several effectors and the corresponding ability to select the appropriate effector that is able to execute an action.

:. Other tools

PRACTIONIST is built on top of JADE. Thus, several development tools available in JADE (e.g. Sniffer agent and Dummy agent) and other thirdy party tools for JADE (e.g. the beangenerator for the ontology designed with Protégé) can also be used with PRACTIONIST. The framework also provides developers with the PRACTIONIST Agent Introspection Tool (PAIT), a visual integrated monitoring and debugging tool, which supports the analysis of the agent's state during its execution. In particular, the PAIT can be suitable to display, test and debug the agents' relevant entities and execution flow. Each of these components can be observed at run-time through a set of specific tabs (see figure); the content of each tab can be also displayed in an independent window.

(click on the picture to enlarge)