Argumentation based modeling of decision aiding for autonomous agents

Decision aiding can be abstractly described as the process of assisting a user/client/decision maker by recommending possible courses of his action. This process has to be able to cope with incomplete and/or inconsistent information for the following reasons. First, the recommendations provided during this process depend heavily on the environment the decision is made. Since complete knowledge of this environment is almost impossible, decision aiding has to be carried out under incomplete information. Second, the decision aiding process also depends on the preferences of its user. However, such subjective information is affected by uncertainty, possible inconsistencies and is dynamically revised due to the time dimension of decision aiding. A complete description of a model of the user is also almost impossible, therefore a decision aiding process must also account for this source of incompleteness. This work presents a model of decision aiding that is amenable to automation and shows how it can be embedded in autonomous agents and thus give them the capability to provide decision aiding to a human user or to completely substitute him for some decision task. The whole process is modelled in a suitable argumentation framework that treats decision aiding as an iterative defeasible reasoning process.