| Linear parameter estimation is matrix theory and application in the important topic, least squares method is the most commonly used linear parameter estimation method, is a kind of widely used in scientific computing method, many experts and scholars on its entered the intensive research. Indefinite linear least squares problem as least squares problem of promotion, can use to total least-squares problems and geometric similar, oblique mapping problems, etc discussion, is nearly 10 years to propose, and matrix equation haven't condemn least-square problems about the definition, so the study of matrix equation condemn least squares problem will be very important sense, this paper defines the categories from the matrix equations condemn least squares, the problem of the existence of solutions for the concrete expression of the solution.This article totally points three chapters. The first chapter, this paper briefly introduces the indefinite linear least squares problem research background, present situation and indefinite linear least squares problem. The second chapter, defines the matrix equation AX = Bof indefinite least squares(?) , use QR decomposition of matrix, and gives the hyperbolic problems sufficient and necessary conditions of solvability in have the solution, and under the condition of the general expressions are given to understand. Discuss matrix A full rank is listed the situation, considering the(?). And discuss with constraints is indefinite least-square problem. In the third chapter, the expansion is studied for a class of matrix, the matrix equations (?)that key research under the restriction of expansion. |
