Abstract:In the process of proving a sharpened form of Gårding's inequality, Fefferman & Phong demonstrated that every non-negative function $f\in C^{3,1}(\mathbb{R}^n)$ can be written as a finite sum of squares of functions in $C^{1,1}(\mathbb{R}^n)$. In this thesis, we generalize their decomposition result to show that if $f\in C^{k,\alpha}(\mathbb{R}^n)$ is non-negative for $0\leq k\leq 3$ and $0<\alpha\leq 1$, then $f$ can be written as a finite sum of squares of functions that are each `half' as regular as $f$, in the sense that they belong to the Hölder space \[ C^\frac{k+\alpha}{2}(\mathbb{R}^n)=\begin{cases}\hfil C^{\frac{k}{2},\frac{\alpha}{2}}(\mathbb{R}^n) & k\textrm{ even}, \\ C^{\frac{k-1}{2},\frac{1+\alpha}{2}}(\mathbb{R}^n) & k\textrm{ odd}. \end{cases} \] We also investigate sufficient conditions for such regularity preserving decompositions to exist when $k\geq 4$, and we construct examples of functions which cannot be decomposed into finite sums of half-regular squares. The aforementioned result of Fefferman & Phong, as well as its subsequent refinements by Tataru, Sawyer & Korobenko, and many other authors, have been repeatedly used to investigate the properties of certain differential operators. We discuss similar applications of our generalized decomposition result to several problems in partial differential equations. In addition, we develop techniques for constructing non-negative polynomials which are not sums of squares of polynomials, and we prove related results which could not be found in a review of the literature.

On the degree of varieties of sum of squares

Sum of Squares Decompositions in Hölder Spaces

Generalized identifiability of sums of squares

High-order sum-of-squares structured tensors: theory and applications

Sums of squares representations on singular loci

Strictly positive polynomials in the boundary of the SOS cone

Weighted Sum-of-Squares Lower Bounds for Univariate Polynomials Imply

Sparse sum-of-squares (SOS) optimization: A bridge between DSOS/SDSOS and SOS optimization for sparse polynomials

The structure of Gram matrices of sum of squares polynomials with restricted harmonic support

Decomposition and Completion of Sum-of-Squares Matrices

Sum-of-squares Chordal Decomposition of Polynomial Matrix Inequalities

Kloosterman sums on orthogonal groups

On Difference-of-SOS and Difference-of-Convex-SOS Decompositions for Polynomials

Improving Efficiency and Scalability of Sum of Squares Optimization: Recent Advances and Limitations

On the Rank of Hermitian Polynomials and the SOS Conjecture

Sum-of-squares certificates for Vizing's conjecture via determining Gröbner bases

S-SOS: Stochastic Sum-Of-Squares for Parametric Polynomial Optimization

Regularity Preserving Sum of Squares Decompositions

Isolations of the sum of two squares from its proper subforms

Theory and applications of the Sum-Of-Squares technique