Estimation of algebraic solution by limiting the solution set of an interval linear system (Q2391679)

The paper deals with the solution of an interval linear system (ILS) whose coefficient matrix is real-valued and right-hand side vector is interval-valued. An interval-number vector \([X_A]\) satisfying the ILS is called the algebraic solution. The interval Gaussian elimination procedure (GEP) is used for solving the ILS. The interval GEP is obtained from the ordinary GEP replacing the real numbers by the related intervals and the real operations by the respective interval operations. The solution \([X_G]\) obtained by the interval GEP is not, in general, the algebraic solution of the ILS, but \([X_A]\subseteq [X_G]\), if \([X_A]\) exists. In the paper, an algorithm for computing \([X_A]\) is proposed. It is based on three steps: (1)~computing \([X_G]\) using the interval GEP; (2)~computing limiting factors using the usual GEP; (3)~if limiting factors belong to positive intervals determined by \([X_G]\), then \([X_A]\) can be computed from \([X_G]\); if not, \([X_A]\) does not exist. The problem is illustrated by simple examples. It is shown, among others, that \([X_A]\) does not exist even for the non-singular triangular matrix. The proposed applications of the method are fuzzy linear systems. The paper is well-written and simply explains all facts.