308 Int. J. Operational Research, Vol. 15, No. 3, 2012 Copyright © 2012 Inderscience Enterprises Ltd. A constructive bandwidth reduction algorithm P. Arathi, L. Jones Tarcius Doss* and K. Kanakadurga The Department of Mathematics, CEG, Anna University, Chennai 600025, Tamilnadu, India E-mail: arathi.ps@gmail.com E-mail: jones@annauniv.edu E-mail: durgakk@annauniv.edu *Corresponding author Abstract: In this paper, a viable bandwidth minimisation algorithm based on graphs, for reducing the bandwidth of sparse symmetric matrices, is presented. The algorithm is tested for various sparse symmetric matrices arising from standard structured and random graphs. Bandwidth results for these matrices are also obtained using the existing algorithms and compared. The proposed algorithm is easy to implement and the bandwidth reductions obtained with the proposed algorithm are good when compared with the existing algorithms. Keywords: bandwidth; sparse symmetric matrix. Reference to this paper should be made as follows: Arathi, P., Doss, L.J.T. and Kanakadurga, K. (2012) ‘A constructive bandwidth reduction algorithm’, Int. J. Operational Research, Vol. 15, No. 3, pp.308–320. Biographical notes: P. Arathi is a Research Scholar, working under Dr Jones, in the Department of Mathematics. She did her MSc Mathematics in Anna University in the year 2004–2006. L. Jones Tarcius Doss is an Associate Professor working in the Department of Mathematics at Anna University, Chennai, India. His areas of interest include numerical mathematics of differential equations and algorithms. K. Kanakadurga is formerly affiliated with the Department of Mathematics, Anna University, Chennai, Tamilnadu, India as a Lecturer. Her areas of interest include computational fluid dynamics and algorithms. 1 Introduction The determination of numerical solutions in science and engineering results in large sparse linear algebraic equations of the form Ax = b, where A is the sparse symmetric. For many decades, solving these equations at a low cost of time and storage has been a challenging task. The problem of reordering a sparse symmetric matrix to reduce the bandwidth has played a central role in the solution of sparse linear systems. A reduction in the bandwidth allows to attenuate the phenomenon of the fill-in and yields advantages both in memory requirements and in computational time.