How to model Diagnostic & Troubleshooting Applications?

Diagnostics is the process of identifying the "root cause" of a problem based on data about the symptoms plus any additional data about the conditions under which these symptoms occur. Troubleshooting is the process of resolving a problem by addressing the root cause or, failing this, to manage the symptoms to minimise their their impact.

The standard approach to model such applications is to apply Diagnostic rules followed by the Troubleshooting Rules. However in some cases the diagnostic process can involve a process of 'trial and error' by applying a series of potential solutions until one is found that resolves the problem.

• Use the known symptoms plus any other available information to narrow down the list of possible root causes. If this requires questions to be asked or or tests to be carried out, then ensure that the order of these should be optimised to minimise the number and/or the cost of applying these tests.
• If more than one potential root cause is identified, rank the list of possible root causes by how likely they are to occur (e.g. from experience).
• Where feasible, apply each potential "fix", in turn, until the problem is resolved. The order of applying the potential fixes should be based on their historical success rate but should also take into account the cost/risk of applying each "fix". For example giving a patient medication to see if it is resolves their condition before attempting major surgery, even if the surgery is more likely to prove effective.