VARIATIONS OF CRAMER'S RULE IN WZ FACTORIZATION

Abstract

 This article compares four established variations of Cramer's rule to solve the linear systems in WZ factorization. The methods are implemented on MATLAB and the results show the matrix norms of the methods are better than classical Cramer's rule. The application of the variations of Cramer's rule in WZ factorization can now be compared on mesh multiprocessors, since the methods minimize round-off error.