Numerical methods -- such as optimization methods, methods for solving systems of equations, etc. -- usually produce approximate solutions. Often, there are no guaranteed bounds on the accuracy of these approximate solutions -- or there are bounds but these bounds are too wide to be practically useful. In such situations, it is desirable to have verified numerical computing, i.e., computing that produces results with verified (provable) accuracy.
When a (final or intermediate) approximate result x of the computation comes with a verified bound D, it means that the actual (unknown) value of the estimated quantity belongs to the interval [x-D,x+D]. In view of this fact, verified numerical computing is also known as interval analysis.
Interval analysis is not only about the bounds. In addition to bounds, we can also have, e.g., partial information about the probabilities of different values from the corresponding intervals or expert information on some of the quantities.
The main purpose of this mini-symposium is to cover the numerical and algorithmic aspects of scientific computing, with a strong emphasis on verification leading to guaranteed properties of computed results as well as on arithmetic, programming, and algorithmic tools that provide and/or enhance this verification.
The intent is to present a state-of-the-art overview on the challenging and dynamic field of verified computing techniques and interval analysis for researchers, experts, and scientists who apply these techniques. We seek contributions that provide competent and concise information on recent hardware and software standards, language support for interval analysis techniques, algorithms with result verification, and applications in various fields.
Paper Submission and Publication The rules of the PPAM conference apply. In particular:
Minisymposium Organizers (in alphabetic order):
Back to the Forthcoming Conferences part of the Interval Computations website