On Weight-Biased Mutation for Graph Problems
Many graph problems seek subgraphs of minimum weight that satisfy the problems’ constraints. Examples include the degreeconstrained minimum spanning tree and traveling salesman problems. Low-weight edges predominate in optimal solutions to these problems, and the performance of evolutionary algorithms for them is often improved by biasing their operators to favor these edges. From the distributions of edges’ ranks in optimal solutions to these two problems, we identify probabilities for edges that minimize the average expected time until mutation chooses them for inclusion in a solution. On instances of the degree-constrained minimum spanning tree problem, an evolutionary algorithm performs better with this operator than with alternative mutations. These results are not replicated on instances of the traveling salesman problem, where the inclusion of one edge in a tour requires the inclusion of another dependant edge.
KeywordsSpan Tree Minimum Span Tree Travel Salesman Problem Travel Salesman Problem Span Tree Problem
Unable to display preview. Download preview PDF.
- 2.C. Fonseca, J.-H. Kim, and A. Smith, editors. Proceedings of the 2000 IEEE Congress on Evolutionary Computation. IEEE Press, 2000.Google Scholar
- 3.J. J. Grefenstette. Incorporating problem specific knowledge into genetic algorithms. In L. Davis, editor, Genetic Algorithms and Simulated Annealing, pages 42–60. Morgan Kaufmann, 1987.Google Scholar
- 4.B. A. Julstrom. Very greedy crossover in a genetic algorithm for the Traveling Salesman Problem. In K. M. George, J. H. Carroll, E. Deaton, D. Oppenheim, and J. Hightower, editors, Proceedings of the 1995 ACM Symposium on Applied Computing, pages 324–328. ACM Press, 1995.Google Scholar
- 5.B. A. Julstrom and G. R. Raidl. Weight-biased edge-crossover in evolutionary algorithms for two graph problems. In G. Lamont, J. Carroll, H. Haddad, D. Morton, G. Papadopoulos, R. Sincovec, and A. Yfantis, editors, Proceedings of the 16th ACM Symposium on Applied Computing, pages 321–326. ACM Press, 2001.Google Scholar
- 7.M. Krishnamoorthy, A. T. Ernst, and Y. M. Sharaiha. Comparison of algorithms for the degree constrained minimum spanning tree. Technical report, CSIRO Mathematical and Information Sciences, Clayton, Australia, 1999.Google Scholar
- 8.I. Ljubic and J. Kratica. A genetic algorithm for the biconnectivity augmentation problem. In Fonseca et al. , pages 89–96.Google Scholar
- 9.G. R. Raidl. An efficient evolutionary algorithm for the degree-constrained minimum spanning tree problem. In Fonseca et al. , pages 104–111.Google Scholar
- 10.G. R. Raidl and C. Drexel. A predecessor coding in an evolutionary algorithm for the capacitated minimum spanning tree problem. In C. Armstrong, editor, Late Breaking Papers at the 2000 Genetic and Evolutionary Computation Conference, pages 309–316, Las Vegas, NV, 2000.Google Scholar
- 11.S. Thienel. ABACUS— A Branch-And-CUt System. PhD thesis, University of Cologne, Cologne, Germany, 1995.Google Scholar