NAGnews 130 | 23 April 2015

Posted on
29 Apr 2015

In this issue:


NAG Library Mark 25 New Functionality Announced


The latest release of the NAG Library (Mark 25) sees the inclusion of 81 new mathematical and statistical routines, many requested by existing Library users. Mark 25 highlights include:

  • New Matrix Functions
  • Least Angle Regression (LARS), Least Absolute Shrinkage and Selection Operator (LASSO) and Forward Stagewise Regression
  • Nearest Correlation Matrix updates
  • Unscented Kalman Filter
  • Change Point Analysis
  • High Dimensional Quadrature using Sparse Grids
  • Bandwidth Reduction of Sparse Matrix by Reverse Cuthill-McKee Reordering
  • Solutions to the classical Travelling Salesman Problem
  • OpenMP Utilities

The inherent flexibility of routines in the Library enables it to be used across multiple programming languages, environments and operating systems including C and C++, Excel, Java, Microsoft .NET, Python, MATLAB, Visual Basic, Fortran and many more.

If you're an existing NAG Library user and want to use the new functionality in Mark 25, do get in touch with us and we'll guide you through your upgrade, or email to talk about licensing options.


Mark 25: New Library Routine Focus - Mixed Integer Nonlinear Programming


In this issue, and future issues, we will focus on the new functionality in the NAG Library at Mark 25 - Mixed Integer Nonlinear Programming is today's focus.

Whereas solvers in the Optimization Chapter (E04 - Minimizing or Maximizing a Function) of the NAG Library address problems defined by continuous variables, those in the Operations Research Chapter (H - Operations Research) allow variables to be integer. These are known as mixed integer problems.

Examples of mixed integer problems include allocation and planning problems in areas such as finance and engineering, and specific examples include:

  • Portfolio optimization
  • The design of gas distribution networks

The inclusion of variables constrained to be integer at the solution leads to extremely difficult optimization problems and a typical requirement is that integer variables are relaxable. That is, the problem and its derivatives with respect to integer variables can be evaluated at non-integer values. For example, replacing an integrality constraint x 2 {0, 1} by 0 _ x _ 1. This can cause an issue, however, if derivatives with respect to integer variables are not available to the user for a given problem.

The Operations Research Chapter (Chapter H) contains mixed integer solvers for LP and QP problems with linear constraints and simple bounds which require integer variables to be relaxable.

At Mark 25 of the NAG Library a new mixed integer solver (h02da) has been added to the Operations Research Chapter (Chapter H) that does not require integer variables to be relaxable. Instead derivatives with respect to integer variables are approximated internally by numerical differences. Furthermore, h02da is a general nonlinear solver that has no restriction on the form of the objective function other than it is required to be twice continuously differentiable. Similarly there is no restriction on problem constraints which, unlike other solvers in Chapter H, can be nonlinear.


Optimization Algorithm Speeds Up Configuration Testing of Powerful MRI Magnet and Helps Secure Development


The most powerful MRI magnet in the world has been developed by Irfu, Institute de Recherche sur les lois fondamentales de l'univers (Institute of Research into the Fundamental Laws of the Universe), a CEA Institute in Saclay, France. Magnetic resonance imaging (MRI) is a diagnostic and research tool used in the clinical research and diagnosis. The whole body "Iseult" magnet is the core component of a magnetic resonance imaging (MRI) scanner that is expected to set new standards for cerebral imaging.

The Iseult magnet is an 11.7 teslas superconducting actively shielded magnet of about 120t with a warm aperture of 900mm in diameter and 5000mm in length. Once built, the Iseult will be installed in the NeuroSpin Imaging Centre at Saclay and dedicated to brain neuroimaging. It will be the most powerful MRI magnet in the world.

Read the full story here.


Nearest Correlation Matrix Solutions Webinar - 21 May 2015


This 30 minute webinar 'Nearest Correlation Matrix Solutions' by Craig Lucas, will give an understanding of how to deal with the issues in forming a nearest correlation matrix from real data.

Attendees will discover:

  • How issues with data can lead to approximate correlation matrices,
  • Some theoretic approaches and,
  • How to use a set of alternative specialized routines to compute true correlation matrices, while fixing some of the original entries.

Register for the webinar here. Learn about the Nearest Correlation Matrix functionality in the NAG Library.


Best of the Blog


Optimization, NAG and Me - 5 Years and Counting

I was always a blend of a computer scientist and a mathematician. Computers and programming were my hobby but it did not feel quite right to choose either as the main subject at university so I picked mathematics. Thereafter began my hunt to get as close to computers as possible. Numerical analysis and mathematical programming/optimization were my lucky answer. Read more.

Introducing the team: Mick Pont, NAG Principal Technical Consultant

I'm a Principal Technical Consultant, and Deputy Manager of the Development Division. I'm involved in the development and peer review of new NAG Library software and documentation, and in the scheduling of software production in line with company targets. Read more.


Training Courses & Events


NAG will be at the following exhibitions and conferences over the next few months.

NAG are once again sponsoring the following six-day course later this year:


We're Hiring! Developer in Mathematical Optimization


We are looking for a developer in Mathematical Optimization to join NAG's Optimization development team in either Oxford or Manchester, UK. For more information on this exciting role visit the 'Careers at NAG' area of the NAG website.

Jan Fiala, NAG Numerical Software Developer has written a blog post about his time working at NAG in the Optimization Team.


NAGnews - Past Issues


We provide an online archive of past issues of NAGnews. For editions prior to 2010, please contact us.