6.4 Scenario Conclusions

This scenario has illustrated our proposed PCM method. It shows the method proposal (guided by the static goals), the knowledge-verification (of constraints and assumptions), the simulation-verification (by computing diagnoses and testing the dynamic goals), and the critique and modify (for recovering from violations of dynamic goals). We have show three possibilities for recovering from the inability to handle normal observed behaviour:

• by adapting the Obs-mapping as a way to get more explanations;
• by adapting the Vocabulary, to better tune the Cover, Obsmap and Vocabulary components to each other;
• by choosing an Obs-mapping that depends on the current observations.

Furthermore we have shown three possibilities of recovering from ``too many solutions'':

• by strengthening the Selection component, because the used Selection is a to weak filter.
• by adding more data, because more data will exclude diagnoses;

The scenario and the alternative paths are summarised in Figure 9. The choice on the brancing nodes of this search space (ie. which repair action has to be taken) is difficult, and needs further study.

  
Figure: The steps of the PCM method in the scenario and the alternative steps. The Method-i and the Diagnoses-i are explained in table 1.

  
Table: Explanation of the terms in figure 9