Resumo:
The proposal of this work is to implement a method for exact solution of systems linear, non rational, whose coefficients are integers, using a packet of multipoint precision floating point arithmetic. The work presents a study of the direct methods of LU factorization, elimination of Gauss and Jordan elimination and then a detailed study of the method proposed by Fox for solution of linear systems, without rounding errors. Finally, some results obtained by direct proposed by Fox.
