Paper · conference · under review
Stress testing event-driven agents to find failures that are rare but costly
Kitchen table
Some failures are so rare that a tester almost never tries them. I look at why the tester skips them and how to make it try.
Whiteboard
Adaptive stress testing searches for failures of event-driven agents. When a failure is rare, the proposal distribution starves it. This paper studies that starvation. The paper is under review.
Boardroom
The failures that cost the most are often the rarest. A test that never tries them gives false comfort. This paper is about finding them before they happen. The paper is under review.
Keywords