Numerical Method For Solving A Class Of Inverse Singular Value Problem

Inverse eigenvalue problem has very important applications in many areas such as structural design,neural network.Inverse eigenvalue problem can be natu-rally extended to Inverse singular value problem which has very important appli-cations in many areas such as circuit theory, orbital mechanics,irrigation theory , pole placement,CT scan and the object's mass distribution. In 1992,M.T.Chu proposed the structural inverse singular value problem and offered two numeri-cal algorithms In recent years, various scholars have proposed different inverse singular problem, the corresponding numerical algorithms are also improving. In this paper,a numerical algorithm is introduced to solve a class of inverse inverse singular value problem.This algorithm is based on Householder transform and rank one updating.We give some basic properties and the computational amount and develop sensitivity analysis of algorithm. The results show that the algo-rithm is very economical and stable.Numerical examples are given to verify the algorithm's practical,economy and stability.In this paper,we use scientific software Matlab 7.5 to testify the rightness of all numerical examples.