This preview shows page 1. Sign up to view the full content.
Unformatted text preview: tomer for a bit longer. Example 3: weekly Timing Prediction improvement at our client Philips Assembléon (2001), using
evolutionary methods. The result was delivered directly to customers. The improvement is in ability to
correctly predict the real factory run time for chip mounting processes. Previously there had been an over
18-month ‘drought’ of improvement to customers, while they worked on much larger delivery cycles. “The
results were received with great applause.” 4. Use metrics to describe metrics, for credibility, and uncertainty.
A metric has its own numeric attributes.
Mertic qualities are for example:
accuracy, credibility, relevance, impact
and costs, like learning cost, test setup cost, test process costs, test analysis costs
We can use metrics to describe and understand our primary metrics.
And to better select both scales of measure, and corresponding measurement processes.
In the Impact Estimation method we can describe each estimate of impact using a set of data:
• the real estimate on the defined scale (...
View Full Document
- Spring '13
- Systems Engineering