Numerical methods for fuzzy system of linear equations (Q1883554)

The paper deals with systems of linear equations \(Ax = b\) with fuzzy right-hand side (\(b\) and \(x\) are vectors of fuzzy numbers). It is a continuation of papers by \textit{M. Friedman} [Fuzzy Sets Syst. 96, 201--209 (1998; Zbl 0929.15004)] and \textit{M. Friedman}, \textit{A. Kandel} and \textit{M. Ma} [Fuzzy Sets Syst. 109, 55--58 (2000; Zbl 0945.15002)], where the associated matrix \(S\) was introduced, \[ S = \left[\begin{matrix} P & Q \\ Q & P \end{matrix}\right] \] with \(p_{i,j} =\max(0,a_{i,j})\), \(q_{i,j} = -\min(0,a_{i,j})\), \(i,j =1,\dots,n\), where the system \(Sy=c\) is crisp (real matrix and vectors). If the matrix \(A\) is diagonally dominant, then \(S\) has also this property (Theorem 3.2). Therefore, approximate solutions can be obtained by Jacobi iterations or Gauss-Seidel iterations.