Decompostition of arithmetic expressions to improve the behavior of interval iteration for nonlinear systems (Q1180347)

In order to verify and to enclose solutions of nonlinear systems within given boxes, the author combines two algorithms based on interval arithmetic. The first expands the given nonlinear system \(F(x)=0\) to a larger one using the code list of \(F\). Solving these new equations with respect to their variables repeatedly according to a particular scheme (introducing intervals and using a stack and intersections) yields a starting interval vector for the second algorithm which is a preconditioned interval Gauss-Seidel method presented in an earlier paper of the author. Theoretical results complete and suppport the computational method, experimental results illustrate its efficiency. A comparison with the well-known Hansen-Greenberg algorithm and some items for future work conclude the paper.