Spring 2019

An Interior-Point Method for Convex Optimization over Non-symmetric Cones

Wednesday, May 1, 2019 3:30 pm4:15 pm PDT

Add to Calendar


Dávid Papp (North Carolina State University)

We present a highly general interior point method for non-symmetric conic optimization and its Matlab implementation. The algorithm can solve optimization problems over a cone as long as the user can supply the gradient and Hessian of a logarithmically homogeneous self-concordant barrier for either the cone or its dual. We outline three applications: hyperbolic programming, sum-of-squares optimization without semidefinite programming, and bounding polynomials using nonnegative circuit polynomials.