generalized_inverse外文电子书籍.pdfVIP

Chapter 4 Generalized Inverses Definition 4.1: A generalized inverse of an m × n matrix A is any n × m matrix G such that AGA = A. The notation A− is often used to indicate a generalized inverse of A. Result 4.2: If A is an n × n non-singular matrix, the unique generalized inverse of A is A−1 . Proof: Since AA−1 = In , we have AA−1A = A. Clearly, A−1 is a g-inverse of A. Suppose B (n × n) is a g-inverse of A. Then ABA = A ⇔ (A−1A)BAA−1 = A−1AA−1 i.e., B = A−1 , (since A−1A = AA−1 = In ). So g-inverse of a non-singular matrix is unique and it is the regular inverse.    1 0     1 0      1 0 0 1 0 0         Example 4.3: Let A = . Let G1 = 0 1 . Then AG1 = 0 1 = 0 1 0  0 0  0 1 0  0 0   1 0    = I ⇒ AG A = I A = A. Thus,k G is a g-inverse of A. Let G = 2 1 2 1 2 0 1  1 0     1 0        1 0 0 1 0         0 1 . Then AG2 = 0 1 = . Thus AG A