Exploiting Decomposability Using Recombination in Genetic Algorithms: An Exploratory Discussion

On certain classes of problems, recombination is more effective if the parents that are being recombined share common subsolutions. These common subsolutions can be used to decompose the recombination space into linearly independent subproblems. If a problem can be decomposed into k subproblems, a single greedy recombination can select the best of 2k possible offspring. The idea of exploiting decomposability works well for the Traveling Salesman Problem, and appears to be applicable to other problems such as Graph Coloring. For Search Based Software Engineering, these ideas might be useful, for example, when applying Genetic Programming to fix software bugs in large programs. Another way in which we might achieve decomposability is by exploiting program modularity and reoccurring program patterns.

[1]  John H. Holland,et al.  Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence , 1992 .

[2]  Name M. Lastname.  Automatically Finding Patches Using Genetic Programming , 2013 .

[3]  Brian W. Kernighan,et al.  An Effective Heuristic Algorithm for the Traveling-Salesman Problem , 1973, Oper. Res..

[4]  J. K. Lenstra,et al.  Local Search in Combinatorial Optimisation. , 1997 .

[5]  G. Croes.  A Method for Solving Traveling-Salesman Problems , 1958 .

[6]  David S. Johnson,et al.  The Traveling Salesman Problem: A Case Study in Local Optimization , 2008 .

[7]  Robert Schaefer.  Parallel Problem Solving from Nature - PPSN XI, 11th International Conference, Kraków, Poland, September 11-15, 2010. Proceedings, Part II , 2010, PPSN.

[8]  Nikolaus Hansen,et al.  Adapting arbitrary normal mutation distributions in evolution strategies: the covariance matrix adaptation , 1996, Proceedings of IEEE International Conference on Evolutionary Computation.

[9]  Andrew M. Sutton,et al.  Partial neighborhoods of elementary landscapes , 2009, GECCO '09.

[10]  Forschungsinstitut für Diskrete.  Chained Lin-Kernighan for Large Traveling Salesman Problems , 2003 .

[11]  Frank Thomson Leighton,et al.  An approximate max-flow min-cut theorem for uniform multicommodity flow problems with applications to approximation algorithms , 1988, [Proceedings 1988] 29th Annual Symposium on Foundations of Computer Science.

[12]  Nikolaus Hansen,et al.  Evaluating the CMA Evolution Strategy on Multimodal Test Functions , 2004, PPSN.

[13]  Lov K. Grover.  Local search and the local structure of NP-complete problems , 1992, Oper. Res. Lett..

[14]  Claire Le Goues,et al.  A genetic programming approach to automated software repair , 2009, GECCO.

[15]  R. Tarjan,et al.  A Separator Theorem for Planar Graphs , 1977 .

[16]  Doug Hains,et al.  A Hybrid Genetic Algorithm for the Traveling Salesman Problem Using Generalized Partition Crossover , 2010, PPSN.

[17]  Claire Le Goues,et al.  Automatic program repair with evolutionary computation , 2010, Commun. ACM.

[18]  Julian F. Miller,et al.  Genetic and Evolutionary Computation — GECCO 2003 , 2003, Lecture Notes in Computer Science.

[19]  P. Stadler.  Spectral Landscape Theory , 1999 .