Test Strategy optimization requires the following questions be asked:

• What are the goals for which test strategies are being optimized?
• How can the degree of optimization be assessed?
• What can be done to optimize a test strategy?
• How does diagnostic tools help perform optimization?

Test Strategy optimization is achieved by focusing on the overall “diagnosability” of a system or device during development. In the most limited sense, test strategy optimization refers to ensuring that tests are performed in an order that best serves development and maintenance goals. In a broader sense, optimization may encompass not only the ordering of tests, but also the enhancement of the hardware’s inherent capacity to facilitate and support effective diagnostics.

Optimization can serve a number of goals, including:

• Reducing Cost of Ownership
• Improved Availability / Readiness
• Reducing Support Environment Requirements
• Greater Safety

Optimizations with multiple goals can be at cross-purposes. For example, improving Availability may result in increased Life Cycle Cost. Finding solutions that satisfy multiple goals can be achieved incrementally by iteratively improving diagnostic capability. For each iteration, the most out-of-range goals are used to provide limits to drive new cases, thereby converging on a solution.

The degree to which case studies meet goals is often determined by examining familiar testability, reliability and maintainability measures, such as:

• Fault Detection / Fault Isolation Levels (FD/FI)
• System Mean Time Between Failures
• Mean Cost / Time to Repair
• Likelihood of Critical Events
• Inherent Availability

While diagnostic optimization may lead to more fundamental design changes (test point placement, partitioning, etc.), development cost can sometimes be reduced by first focusing on changes to testing, such as:

• Refining Test Selection / Ordering Criteria
• Additional Test Procedures (automated or otherwise)
• Changing Test Implementation (BIT, ATE, Probe, etc)

It’s important to recognize when changes to testing result in diminishing returns, an indication that design changes should be considered, or that the proposed changes to testing are ineffective. More sophisticated testing can involve greater development costs, often overriding any benefits which lead to its consideration. Similar changes made earlier in the design’s development, however, can achieve much of the same optimization at a greatly reduced cost. Optimization is about finding the balance between seemingly competing criteria.