Quadratic models Q_k(.) of the objective function F(.) are used by many successful iterative algorithms for minimization, where k is the iteration number. Given the vector of variables x_k, a new vector x_{k+1} may be calculated that satisfies Q_k(x_{k+1}) < Q_k(x_k), in the hope that it provides the reduction F(x_{k+1}) < F(x_k). Trust region methods include a bound of the form ||x_{k+1}-x_k|| .leq. Delta_k. Also we allow general linear constraints on the variables that have to hold at x_k and at x_{k+1}. We consider the construction of x_{k+1}, using only of magnitude n squared operations on a typical iteration when n is large, where n is the number of variables. The linear constraints are treated by active sets, which may be updated during an iteration, and which decrease the number of degrees of freedom in the variables temporarily, by restricting x to an affine subset. Conjugate gradient and Krylov subspace methods are addressed for adjusting the reduced variables, but the resultant steps are expressed in terms of the original variables. Termination conditions are given that are intended to combine suitable reductions in Q_k(.) with a sufficiently small number of steps. The reason for our work is that x_{k+1} is required in the LINCOA software of the author, which is designed for linearly constrained optimization without derivatives when there are hundreds of variables. Our studies go beyond the final version of LINCOA, however, which employs conjugate gradients with termination at the trust region boundary. In particular, we find that, if an active set is updated at a point that is not the trust region centre, then the Krylov method may terminate too early due to a degeneracy. An extension to the conjugate gradient method for searching round the trust region boundary receives attention too, but it was also rejected from LINCOA because of poor numerical results. The given techniques of LINCOA seem to be adequate in practice.
Citation
Report DAMTP 2014/NA02, Centre for Mathematical Studies, University of Cambridge, UK, August, 2014.
Article
View On fast trust region methods for quadratic models with linear constraints