A Subspace Minimization Method for Constrained Optimization
We will discuss how certain properties of quasi-Newton methods have been exploited to derive an efficient algorithm for unconstrained optimization, which works by restricting search directions to a subspace at each iteration. Then we will present a new algorithm, RH-B, which applies these principles to problems with bound constraints. This will include a discussion about issues with the current implementation, suggestions for future versions and numerical results.
Tuesday, February 24, 2009
11:00AM AP&M 2402
Center for Computational Mathematics9500 Gilman Dr. #0112La Jolla, CA 92093-0112Tel: (858)534-9056