A Causal Framework for Intelligent Agents and Multiagent Systems

Automation of organizational processes that make intensive use of expert knowledge has on the intelligent agent paradigm an appropriate metaphor. The idea would be introducing autonomous agents carrying out tasks on behalf of human users in organizational processes. Knowledge transference from experts to autonomous agents has been treated on different ways taking advantage of more than 20 years of experience in knowledge representation in Artificial Intelligence. Data mining and case-based reasoning use historical information for identifying patterns that guide agent’s decision. Inference rules are used for explicitly capturing expert knowledge. Agent’s capacity for adapting to changes rely on the employed formalism. In many cases, the expert is left aside in order to provide an automatic solution. On the other hand, Autonomic Computing is providing interesting approaches for dealing with the dynamic adaptation of components to organizational and environmental changes. This theory proposes incorporating self-configuration, self-optimization, self-healing and self-protecting capabilities on information systems. Solutions on this field are usually partial and seem to lack a coherent framework for representing the notion of integral autonomy. In order to propose a conceptual framework that provide coherence on the modeling of autonomic systems we are introducing some notions of Classical Philosophy that have not been taken full advantage of. This framework intends to facilitate knowledge transference from human experts to autonomous agents and, on the opposite direction, the participation of human users on the adaptation to environmental changes.