Nonlinear Krylov-Secant Solvers Héctor Klíe, and Mary F. Wheeler The Center for Subsurface Modeling (CSM) The Institute for Computational Engineering and Sciences (ICES) The University of Texas at Austin, Austin, TX, 78712 Contents 1 Introduction 2 2 Newton-Krylov Framework 3 3 The Family of Secant Solvers 5 3.1 Broyden’s Method ........................................... 7 3.2 The Nonlinear Eirola-Nevanlinna (NEN) Method .......................... 7 4 Krylov-Secant Updates 9 4.1 Secant Updates constrained to the Krylov subspace ......................... 9 4.2 Krylov Secant vs. Standard Secant Update ............................. 12 4.3 The Nonlinear Krylov-Eirola-Nevanlinna Algorithm ......................... 14 4.4 A high-order Newton–Krylov algorithm ............................... 15 5 Implementation issues 15 5.1 Preconditioning ............................................ 15 5.2 Globalization Strategy and Forcing Terms .............................. 16 6 Numerical Experiments 17 7 Conclusions and Further Work 19 Abstract This report describes a new family of Newton-Krylov methods for solving nonlinear systems of equations arising from the solution of Richards’ equation and in fully implicit formulations in air-water sys- tems. The basic approach is to perform secant (Broyden) updates restricted to the Krylov subspace generated by the GMRES iterative solver. This approach is introduced as Krylov-secant methods. One of the most attractive features of these methods is their performance of sequence of rank-one updates without explicitly recalling the com- putation or action of the Jacobian matrix. Implications of these up- dates in line-search globalization strategies, computation of dynamic * Keywords and phrases: Krylov-secant, Broyden’s method, Newton-Krylov, GMRES, Arnoldi factorization, rank-one updates; Author’s e-mail: {klie,mfw}@ices.utexas.edu; UT, Austin. 1