Title :
A descent algorithm for L1-deconvolution
Author_Institution :
Bologna Univ., Italy
Abstract :
Presents a descent method for the minimisation of a functional that arises from the least absolute deviation problem. The method is aimed at ill-conditioned problems, i.e. problems where the columns of the matrix are not from linear dependence. In spite of the complications introduced for the sake of numerical stability, the procedure is fast and efficient
Keywords :
matrix algebra; minimisation; numerical methods; L1-deconvolution; descent algorithm; functional; ill-conditioned problems; least absolute deviation problem; matrix; minimisation; numerical stability;
Conference_Titel :
Digital Processing of Signals in Communications, 1991., Sixth International Conference on
Conference_Location :
Loughborough
Print_ISBN :
0-85296-522-2