On diagonally structured problems in unconstrained optimization using an inexact super Halley method (Q432797)

For solving the unconstrained minimization problem an iterative method based on the third-order super Halley method is proposed and it is shown how to approximately solve the two systems of linear equations in the inexact super Halley method so that the method retains local and has third-order rate of convergence, thereby making it possible to use iterative methods for solving the linear systems.NEWLINENEWLINE This paper introduces an array of arrays (jagged) data structure for storing the second and third derivative of a multivariate function and suitable termination criteria for the (inner) iterative method to achieve a cubic rate of convergence.NEWLINENEWLINE Further it is shown that the second derivative is diagonally structured, the third derivative also exhibits a diagonal structure which can be utilized in computing with the third-order derivative. It is proved by examples that the exploitation of an a jagged compressed diagonal storage of the Hessian matrices and for the tensor is more efficient than the row or column oriented approach when one uses an iterative method for solving the systems of linear equations since matrix vector products can be implemented very efficiently for diagonally structured matrices.