Abstract
The centred simplex gradient (CSG) is a popular gradient approximation technique in derivative-free optimization. Its computation requires a perfectly symmetric set of sample points and it is known to provide an order-2 accuracy with regards to the radius of the sample set. In this talk, we consider the situation where the set of sample points is not perfectly symmetric. By adapting the formula for the CSG to compensate for the misaligned points, we define a new Adapted-CSG. We prove that the Adapted-CSG retains order-2 accuracy and present numerical examples to demonstrate its properties.
Date
Jul 20, 2022
Location
DFOS 2022
University of British Columbia, Kelowna, BC
Yiwen Chen
PhD student in Mathematics
My research interests include derivative-free optimization, numerical optimization, and discrete geometry.