Abstract

The need for a computerized diagnostic strategy choice aiding tool is identified, and a methodology for its implementation is proposed. In troubleshooting practice, a diagnostic strategy may be described by the test sequencing optimization objective and the algorithm to produce the optimal test sequence. It is possible to pose a constrained optimization problem in the space of the diagnostic strategies. Solving such optimization problems requires a methodology for representing and manipulating the strategies in conjunction with the given problem in the troubleshooting domain. It is emphasized that an expert diagnostic system should be able to choose and implement a diagnostic strategy depending on the current situation. To facilitate such an ability, it is necessary to create a methodology for representing different diagnostic strategies, to provide tools for evaluating the current situation (e.g. specifics of the failure rates, costs, system topology, etc.), and to provide a mechanism for matching and implementing a diagnostic strategy in the given situation. Such an evaluation of the system under diagnosis is called metadiagnosis. A semantic control approach for a computerized strategy choice aiding tool architecture is proposed, and an example of its application is provided.