Support for Symmetric/Hermitian Eigenvalue Problems

[jacobdwatters]

The eigenvalues of a symmetric matrix are real. This means that a a specialized algorithm could take advantage of the symmetric structure and do all computations in real arithmetic. Similarly, Hermitian matrices could benefit from a specialized implementation as well.

Tasks:

• - Implement a symmetric and Hermitian Hessenburg decomposition (i.e. tridiagonalization).
• - Implement a symmetric and Hermitian Schur decomposition (utilizing tridiagonalization from symmetric/Hermitian Hessenburg decompositions).
• - Implement eigenvalue/vector computation for symmetric and Hermitian matrices utilizing the associated Schur decompositions.