University of Illinois Urbana-Champaign

Application of black-box optimization on system verification

Leung, Peter

This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.

Permalink

https://hdl.handle.net/2142/107260

Description

Title
Application of black-box optimization on system verification
Author(s)
Leung, Peter
Contributor(s)
Mitra, Sayan
Issue Date
2020-05
Keyword(s)
  • Verification
  • Optimistic Optimization
  • Black-box
  • DryVR
Abstract
System verification is a very generalizable problem about ensuring how a system behaves. Reliable system verification techniques are essential as autonomous systems are becoming more prevalent. This can be seen with increasing numbers of self-driving cars and drone deliveries. Our research approaches this problem by determining if a system is safe. In other words: Given the possible states from different initial conditions, do all these states avoid a predefined unsafe set? We have set up a method to utilize optimization techniques to help verify systems. Our research focuses on black-box verification, where we can only take zero-order evaluations of the function in question. There is already much research into problems where we have full knowledge of the function to be optimized, or no knowledge at all. But limited work has been done in the area in between. We hope to leverage some knowledge of the “smoothness” of the function to help us verify systems. This thesis explains how we set up these systems and utilized previous research on optimistic optimization. The thesis proposes a new optimization algorithm. Lastly results are compared to suggest what systems and conditions would be best suited to each algorithm, or whether an approach entirely different from ours would be more appropriate.
Type of Resource
text
Language
en
Permalink
http://hdl.handle.net/2142/107260

Owning Collections

Senior Theses - Electrical and Computer Engineering PRIMARY
The best of ECE undergraduate research