An Improved Heuristic for the Bandwidth Minimization Based on Genetic Programming
In this work we develop an improved heuristic based on genetic programming (GP) for the matrix bandwidth minimization problem (MBMP). This problem consists in rearranging the rows and columns of a sparse matrix such that the non-zero elements are in a band as close as possible to the main diagonal. We evaluated our heuristic on a set of 25 benchmark instances from the literature and compared with state-of-the-art algorithms. The obtained results are very encouraging and point out that GP is an appropriate method for solving the MBMP.
Keywordsbandwidth minimization problem heuristics genetic programming
Unable to display preview. Download preview PDF.
- 2.Cuthill, E., McKee, J.: Reducing the bandwidth of sparse symmetric matrices. In: Proc. 24-th Nat. Conf., pp. 157–172. ACM, New York (1969)Google Scholar
- 3.Koohestani, B., Poli, R.: A Genetic Programming Approach to the Matrix Bandwidth-Minimization Problem. In: Schaefer, R., Cotta, C., Kołodziej, J., Rudolph, G. (eds.) PPSN XI. LNCS, vol. 6239, pp. 482–491. Springer, Heidelberg (2010)Google Scholar
- 9.Matei, O.: Evolutionary Computation: Principles and Practices. Risoprint (2008)Google Scholar