Hartmut Freitag

A generic measurement proposer

Selecting the next measurement point in order to discriminate among candidate diagnoses is a crucial task for diagnosis system. We present a generic system which allows the fast and effective realization of different measurement selection strategies for model-based diagnosis systems. The minimum entropy technique used in the General Diagnostic Engine (Gde) and a simplification of it presented in [de Kleer 89] are two possible instances of this system. Several other techniques based on this framework are presented, including another simplification of the minimum entropy technique which overcomes deficiencies of [de Kleer 89].

Propagation of Temporally Indexed Values in Multiple Contexts

Modelling multiple views of a dynamic system requires both an efficient technique for dealing with multiple contexts and a mechanism for explicit reasoning about time. Assumption-based truth maintenance systems simultaneously handle multiple, possibly contradictory views, which may be characterized by different sets of assumptions about correctness of components, particular operating modes, etc.. This is mainly achieved by recording inference steps in the form of explicit dependencies and propagating sets of assumptions along these dependencies. Constraint systems explicitly reasoning about time use value-interval pairs, so-called episodes, as parameter values.

An Integrated Non-Monotonic Deduction and Reason Maintenance System

KAPRI⋆⋆ uses rules of the form if … unless … then … . For such a rule to be fired, its monotonie if-antecedents are required to match the current database, while none of its non-monotonic unless-antecedents does. The current state of a KAPRI-system is represented by a dependency network composed of assertions and corresponding justifications for believing in them in terms of belief respectively disbelief in other assertions. This network is maintained by a reason maintenance system, KL-DNMS⋆⋆, that revises the current belief status with respect to every modification due to the firing of a rule or to the addition/retraction of a basic fact.

An Efficient Interpreter for a Rule-Based Non-Monotonic Deduction System

We present an efficient interpreter for CAPRI, a non-monotonic deduction system using rules of the form if...unless...then... .The interpreter is built on top of a reason maintenance system maintaining its current database and includes a fast pattern matcher combining both ideas from RETE and from non-monotonic dependency network update algorithms.