There are many diagnostic reasoning approaches, each having their own distinct advantages and disadvantages. eXpress allows the Analyst to develop diagnostics based on some combination of the following two approaches:
Functional Dependency-Based Reasoning
Failure Mode Dependency-Based Reasoning
Because diagnostics can be developed using these two approaches during relatively early phases of product design and development, these approaches are ideal for analyzing the diagnostic capacity of a design. Actual run-time diagnostics, however, may be implemented using an approach that does not involve dependency-based reasoning. These approaches include:
Model-Based Reasoning
Rule-Based Expert Systems
Case-Based Reasoning
AI-Based Reasoning (Neural Net, Bayesian, etc.)
Unfortunately, diagnostics that utilize these approaches may not be able to be developed until relatively late in product development, at a point in time when feedback about how well the design supports diagnostics may no longer be useful. Frequently, design-time analysis will be performed using proto-diagnostics created using a dependency-based approach, even though the actual run-time diagnostics will employ a different type of reasoning. Rather than optimizing the design to support a specific diagnostic strategy, different case studies can be employed to ensure that the design is optimized to support diagnosis in general.
The following chart shows which diagnostic approaches can be developed during different phases of product development and deployment.
