Backcalculation

Engineering design can be formulated as the solution to an equation or inequality involving uncertain numbers. Given variation and unpredictability, how can the values of variables be manipulated to insure that particular constraints are observed, or exceeded in statistically controlled ways?

Scott's slides about untangling equations involving uncertainty (backcalculation v. deconvolution)

NASA final report from AB's SBIR award

Yan Wang's slides and recorded lecture at https://riskinstitute.uk/riskinstituteonline/

Ward's MIT dissertation about designing fits off the shelf

Robo-engineer

Scott:

Thank you for the opportunity to present the work at your institute. I also very much enjoyed the discussion with the students. We need more young and brave minds like them in the imprecise probability community. Thought the ShutdownSTEM is a U.S. activity. The world is getting really small now. Hope would be able to catch up after this virus episode is done, maybe at next SRA conference, and continue the conversation. If not sooner, have a great summer!

Best regards,

Yan

Dear Yan:

Thanks a lot for your talk this afternoon. My informal survey of students after the event was that it was “great”. It really helped a lot, and was amazingly timely.

Sorry we didn’t get a chance to chat after the talk. I was hoping to catch up. As you know, I am convinced that your approach will be the future of uncertainty in engineering, and perhaps the future of engineering itself. I think that Ward’s scheme can be operationalised that way. I don’t really understand how the generalisation to p-boxes works though.

By the way, I was embarrassed to learn this evening that there was a call to pause STEM research for today in support of Black Lives Matter. Like most parties, I didn’t even know about it until it was over. Had I realised I would have at least offered you the option to reschedule!

Best regards,