Model-Based Reasoning

- People

1. Automated support for the design of qualitative models

2. Decentralized model-based and diagnosis with qualitative models

The first activity has been mainly carried out within the AUTAS EU project, whose goal was to automate the production of Failure Mode Effects Analysis (FMEA) on Aircraft Systems. The AUTAS toolset, a prototype developed in the project, makes use of Model-Based Reasoning techniques together with Qualitative Reasoning in order to derive automatically the set of effects that occur as a consequence of a failure (or combination of failures). The AUTAS project has shown that, although Qualitative Models make it possible to describe in an abstract and simpler way several fault situations, building a correct Qualitative Model requires highh degree of expertise on Qualitative Physics and logic, which company engineers often do not have. For this reason the group has studied the problem of intelligent automated support to the design of Qualitative Models.

The second activity has started as a continution of the group previous work on Model-Based Diagnosis, particularly in connection with Contraint Propagation (as the main technique used for solving diagnostic problems) and Qualitative Modelling (as the set of languages used to describe diagnostic problems). The group has proposed a decentralized diagnostic framework that makes use of these techniques, since the works on decentralized diagnosis in the literature were all focussing on different diagnostic techniques (mainly Discrete Events Systems and Finite State Machines). The framework has been applied within the WS-DIAMOND EU project, where it used for the decentralized diagnosis of complex Web Services. This work has been carried out in cooperation with the research group on Artificial Intelligence and Human-Computer Interaction, which Prof. Console is also member of.

- Publications