Algorithm 768: TENSOLVE

Abstract

This article describes a modular solftware package for solving systems of nonlinear equations and nonlinear problems, using a new class of methods called tensor methods. It is intended for small- to medium-sized problems, say with up to 100 equations and unknowns, in cases where it is reasonable to calculate the Jacobian matrix or to approximate it by finite differences at each iteration. The software allows the user to choose between a tensor method and a standard method based on a linear model. The tensor method approximates F ( x ) by a quadratic model, where the second-order term is chosen so that the model is hardly more expensive to form, store, or solve than the standard linear model. Moreover, the software provides two different global strategies: a line search approach and a two-dimensional trust region approach. Test results indicate that, in general, tensor methods are significantly more efficient and robust than standard methods on small- and medium-sized problems in iterations and function evaluations.