sqrtM2

sqrtM2 is a notation used to denote the square root of a quantity called M2, which can be either a scalar or a matrix depending on the context. In the scalar case, sqrtM2 refers to the nonnegative real square root of M2, defined for M2 ≥ 0 as the number whose square equals M2.

In linear algebra, sqrtM2 commonly refers to a matrix square root: a matrix S such that S^2 = M2. When M2 is Hermitian and positive semidefinite, there exists a unique principal square root, which is Hermitian and positive semidefinite. This principal square root is often denoted by sqrt(M2) or by sqrtM2 in informal notation. For general matrices, a square root may not exist, and if it does, there can be multiple distinct square roots.

Existence and uniqueness depend on the properties of M2. If M2 is real symmetric and positive semidefinite, a real, unique principal square root exists. If M2 has negative or complex eigenvalues, a real square root may not exist, and complex square roots may be defined. Even when a square root exists, it is not guaranteed to be unique unless a specific branch (such as the principal root) is fixed.

Computational approaches include eigenvalue decomposition M2 = Q Λ Q^{-1}, yielding sqrtM2 = Q sqrt(Λ) Q^{-1} when Λ has nonnegative eigenvalues, and iterative methods like Denman–Beavers or Newton’s method for numerical stability. Related concepts include the matrix square root, the principal square root, and the matrix absolute value defined by sqrt(M2) for certain interpretations of M2.

---