PURPOSE: To increase the converging speed when an answer is obtained for the matrix calculation of a repeating solution by changing subsequently the value of a parameter which accelerates the convergence.

CONSTITUTION: When a multi-dimensional simultaneous linear equation is solved by a repeating solution, the value α of a proper acceleration parameter is inputted as the initial value and the correction value α1.δx is stored after a matrix is solved. The new correction value α2.δx. If the former correction value is smaller than the latter one, the value α is increased to increase the acceleration. While the value α is suddenly decreased with no acceleration applied if the former correction value is smaller than the latter one. When the decreasing procedure of the value α is through, the old correction value is changed to the new one and stored. Thus the converging speed of the solution can be increased.