Rambling Abductive Agents in ALIAS

This paper presents the Abductive LogIc Agent System (ALIAS, for short), a multi-agent architecture based on cooperating intelligent logic agents. Agents in ALIAS are grouped into bunches; each bunch represents an agora where agents together discuss about common arguments. ALIAS agents can ramble from bunch to bunch at runtime, thus collecting information (i.e. agent's experience) derived from the interactions with other agents in diierent bunches. ALIAS agents are equipped with hypothetical reasoning capabilities, performed by means of abduction: if the knowledge available to an agent is insuucient to solve a query, the agent could abduce new hypotheses. To this purpose, each agent is characterized by a local knowledge base represented by an abductive logic program. Agents in the same bunch might diier in their knowledge bases, but must agree on assumed hypotheses. Abduced hypotheses within a bunch build up a common knowledge base that is dynamically created and managed by means of a tuple space .