There exist numerical methods, called eigensolvers, which can compute eigenvalues (and eigenvectors) even for very large matrices, with hundreds of rows/columns, or larger. How could this be? The ans...There exist numerical methods, called eigensolvers, which can compute eigenvalues (and eigenvectors) even for very large matrices, with hundreds of rows/columns, or larger. How could this be? The answer is that numerical eigensolvers are approximate, not exact. But even though their results are not exact, they are very precise—they can approach the exact eigenvalues to within the fundamental precision limits of floating-point arithmetic.
