Positive-definite matrix（正定矩阵）

正定矩阵定义

In linear algebra, a symmetric(对称) n × n real matrix M is said to be positive definite if z^TMz is positive, for any non-zero column vector z of n real numbers; where z^T denotes the transpose of z.

More generally, an n × n complex matrix M is said to be positive definite if z*Mz is real and positive for all non-zero complex vectors z; where z* denotes the conjugate transpose of z. This property implies that M is an Hermitian matrix(埃尔米特矩阵，共轭转置仍是自己).

The negative definite, positive semi-definite, and negative semi-definite matrices are defined in the same way, except that the formula z^TMz or z*Mz is required to be always negative, non-negative, and non-positive, respectively.

特征及性质

Let M be an n × n Hermitian matrix. The following properties are equivalent to M being positive definite:

1. All its eigenvalues are positive.
2. 正定矩阵一定是非奇异的。非奇异矩阵的定义：若n阶矩阵A的行列式不为零，即|A|≠0.
3. It has a unique Cholesky decomposition. The matrix M is positive definite if and only if there exists an unique lower triangular matrix L, with real and strictly positive diagonal elements, such that M = LL*. This factorization is called the Cholesky decomposition of M.
4. The general purely quadratic real function f(z)on n real variables z₁, ..., z_n can always be written as z^TMz where z is the column vector with those variables, and M is a symmetric real matrix. Therefore,the matrix being positive definite means that f has a unique minimum (zero) when z is zero, and is strictly positive for any other z。
5. 判定定理1：对称阵A为正定的充分必要条件是：A的特征值全为正。
   判定定理2：对称阵A为正定的充分必要条件是：A的各阶顺序主子式都为正。
   判定定理3：任意阵A为正定的充分必要条件是：A合同于单位阵。