Conjugate gradient methods for continuation problems II

We study linear conjugate gradient (CG) methods for large sparse continuation problems. First we show how the CG methods can be incorporated as linear solvers in the context of a continuation method. Next we indicate how the preconditioned generalized minimal residual algorithm may be used to test for simple and multiple bifurcations without solving the bordered linear systems. Sample numerical results are reported. Some concluding remarks concerning the performance of three-term CG methods are given.