Abstract: In this talk, we will present interactions between conic programming, a branch of optimization, and real algebraic geometry, the algebraic and geometric study of real polynomial systems. This has led to a number of beautiful mathematical theorems, both improving our understanding of the geometric picture as well as sharpening our tools for applications.
On the optimization side, we will take a look at hyperbolic programs, an instance of conic programs first studied in the 1990s that comprises linear programs as well as semidefinite programs. On the algebraic side, it is based on the theory of hyperbolic respectively real stable polynomials, which show up in many parts of mathematics, ranging from control theory to combinatorics. In many ways, hyperbolic polynomials behave like determinants of families of symmetric matrices. We will compare these two paradigms, one based in matrix calculus, the other purely in algebraic geometry, through a series of examples, pictures, theorems, and conjectures.