NAG Library News, Mark 28.3 (March 2022)

This Mark begins a new numbering scheme for the NAG Library. The point release number, $3$ in this case, now denotes the month of release (March). Further releases in the same year will have a point release matching the number of the month of release (e.g., a release in May of this year would be number $28.5$).

At Mark 28.3, the NAG Library consists of four main interface sets: the NAG FL Interface, the NAG CL Interface, NAG CPP Interface and the NAG AD Library Interfaces. The NAG FL Interface now consists of 1945 user-callable routines while the NAG CL Interface now consists of 1659. All of these routines are fully documented and 4 are new since Mark 27.3. See Section 4 for the full list of new routines.

The following summarises additional content and changes to the NAG Library.

Chapter E04 (Minimizing or Maximizing a Function) has two additions to the NAG optimization modelling suite.

A modern active-set sequential quadratic programming (SQP) solver for sparse nonlinear programming (NLP). Among the main updates are improvements to the internal QP solver and reworked elastic mode to increase reliability. Overall, the new solver offers a performance increase over 15% when compared to e04ugf/​e04uga and e04vhf and several problems which were unsolvable are now successfully addressed. In addition, the integration to the NAG optimization modelling suite offers clarity and consistency of the interface of the solvers within the suite, making it trivial to switch among compatible solvers and benefit from the intuitive input.

A new general nonlinear data-fitting solver. Various loss functions from ordinary least squares to robust regression are supported. Regularizations from LASSO and ridge regression can be easily added to the selected loss function. This solver handles bound constraints, linear constraints, quadratic constraints, and general nonlinear constraints. It is an ideal starting point for nonlinear regression and calibration.

Chapter E05 (Global Optimization of a Function) has a new interface to the Multi-level Coordinate Search (MCS) algorithm to integrate it to the NAG optimization modelling suite. This new routine gives access to all of the suite facilities and uses the same interface as all the compatible solvers, simplifying experimentation significantly. The e05kbf interface is intended as a replacement for e05jbf.

Chapter G05 has a new routine to do weighted sampling with replacement, with a choice of resampling method and can return multiple samples.

The News document for the previous release, Mark 27.3, is available on the NAG website.

Some routines in the Library may be classified as experimental. These routines will be flagged by a note at the top of the documentation.

A routine classified as experimental has the following properties:

1. (a)The routine interface may have to change at a later release to accommodate additional features or improve its usability. The routine will have been designed in such a way as to minimize the number of such changes.
2. (b)It has gone through the same stringent testing and review processes as a normal Library routine.

A routine classified as experimental may, at some point, be reclassified as no longer being experimental, at which point the interface will become fixed as per a normal Library routine. It is not expected that an experimental routine would be removed from the Library without a replacement routine being added.

At Mark 28.3 we no longer distribute separate installers for the NAG Library Manual with each implementation. Please refer to the NAG Library documentation page on the NAG website, which has links to the manual for all recent releases of NAG Library.

The lists in this section refer and link to the NAG FL Interface. In these lists there is a one-to-one correspondence between the NAG FL Interface and the NAG CL Interface; and, there is a direct hyperlink from the NAG FL Interface routine document to its corresponding NAG CL Interface routine document.

The new NAG AD Library routine documents currently available are listed in the AD Library Contents and there is a direct hyperlink from the NAG FL Interface routine document to its corresponding NAG AD Library routine document, and vice versa.

Since Mark 27.1, all NAG AD Library numerical routines have variants for computing adjoints, tangent linear derivatives or the purely primal computation using consistent interfaces.

Since Mark 27.1, C++ interfaces for a subset of NAG Library are also provided in the NAG CPP Interface. The currently available interfaces are listed in the CPP Interface Contents.

The following routines have been significantly updated or enhanced at Mark 28.3 and details are available in each routine document.

For details of all known issues which have been reported for the NAG Library please refer to the Known Issues lists for the Library FL Interface and CL Interface.

Since Mark 27.3, the following routines have been withdrawn or scheduled for withdrawal from the NAG Library, or have been deprecated. For withdrawn routines, warning of their withdrawal has been included in the NAG Library Manual since Mark 26, together with advice on which routines to use instead.

Please note that unless a Mark of Withdrawal is shown below, the routine will continue to form part of the Library but you are advised to stop using the deprecated routines and to use the recommended replacement routines instead.

Please see the chapter introduction documents and the ‘FL Interface Replacement Calls Advice’ (or the CL version ‘CL Interface Replacement Calls Advice’) for more detailed guidance, including advice on how to change a call to the deprecated routine into a call to its recommended replacement, or use the links provided in the following table.