Mark 4 of NAG C Subroutine Library Available on RCS

Mark 4 of the Numerical Algorithms Group (NAG) C subroutine library is now available on all RCS machines. The library, which contains a large collection of algorithmic routines written in C, is organized into chapters, with each chapter devoted to a specific branch of numerical or statistical computation.

NAG C Library Features

Some highlights of the library include:

• Minimization, constrained and unconstrained using first derivatives or function values only. Nonlinear least squares. Linear and quadratic programming. Optional argument facility.

• Ordinary Differential Equations, using variable-order and variable-step Adams method with facilities for interpolation and root finding.

• Fourier Transforms, for single or multiple series using real, complex, and Hermitian data.

• Fully-adaptive quadrature, fully adaptive with specialised weight functions over finite and infinite intervals. Multi-dimensional quadrature.

• Linear Algebra, including: Eigensystem analysis of real and Hermitian matrices. Solutions of simultaneous equations. Matrix factorizations.

• Solutions of Systems of non-linear equations, using first derivatives or function values only.

• Zeros of Polynomials, with real or complex coefficients.

• Statistical Functions, common distributions and their inverses, random variate generation.

• Regression Analysis and general linear models, including model updating.

• Time Series Modelling, calculations of acf and pacf, ARIMA model fitting and forecasting, transfer functions, Kalman filtering.

• Curve and Surface Fitting over one and two dimensions (rectangular grid or scattered data) with a choice of fitting strategies. Computation of derivatives and integrals.

• Interpolation over one or two dimensions using cubic splines, or shape-preserving piecewise polynomials.

• Sorting values of arbitrary data types, rank sort, chain sort and linked list, vector search for matched value.

• Special Functions, error functions, Bessel functions, Airy functions, Fresnel functions and elliptic integrals. Kelvin functions, Hyperbolic and inverse hyperbolic functions, Gamma functions.

• Utilities, machine constants, mathematical constants, functions to manipulate complex numbers, BLAS.

• Random Number Generators, generation of sequences of independent pseudo-random numbers from various distributions, and the generation of pseudo-random time series from specified time series models.

• Univariate Estimation, robust estimation, media, media absolute deviation, robust standard deviation.

• Smoothing in Statistics, smoothed data sequence using running medium smoother.

• Operations Research, classical transportation algorithm.

User Interface

The NAG C library functions have been designed to be easy-to-use, while still allowing the user to interact and control even the most sophisticated and advanced of numerical algorithms. This has been achieved by:

• Data abstraction through the use of structures,

• The dynamic allocation of memory for workspace arrays,

• Flexible error handling, and

• The provision of a mechanism for selecting and adjusting optional parameters on the more complex algorithms.

Accessing the Library

To access the library, enter a compilation command similar to the following from a UNIX prompt:

   	cc -I/usr/local/include myprogram.c -L/usr/local/lib -lnagc -lm

Getting More Information

If you have general questions about using the NAG C library, please contact the consultants at the ACS Help Desk locations, or direct electronic mail to consult@rpi.edu.