Jeffrey Horn
ABSTRACT
Recent success with a simple type of coevolution, resource defined fitness sharing (RFS), involving only pairwise interactions among species, has inspired some static analysis of the species interaction matrix. Under the assumption of equilibrium (w.r.t. selection), the matrix yields a set of linear equations. If there exists a subset of species that exactly cover the resources, then its characteristic population vector is a solution to the equilibrium equations. And if the matrix is non-singular, a solution to the equilibrium equations specifies an exact cover of the resources. This polynomial-time reduction of exact cover problems to linear equations is used in this paper to transform certain exact cover NP-complete problems to certain linear equation NP-complete problems: 0-1 Integer Programming, Minimum Weight Positive Solution to Linear Equations. While most of these problems are known to be in NP-complete, our new proof technique introduces a practical, polynomial-time heuristic algorithm for solving large instances of them.
- K. Deb and D. E. Goldberg. An investigation of niche and species formation in genetic function optimization. In J. D. Schaffer, editor, Proceedings of the Third International Conference on Genetic Algorithms, pages 42--50. Morgan Kaufman, San Mateo, CA, 1989. Google Scholar
Digital Library
- R. Dighe and M. J. Jakiela. Solving pattern nesting problems with genetic algorithms: employing task decomposition and contact detection between adjacent pieces. Evolutionary Computation, 3(3):239--266, 1996. Google ScholarDigital Library
- M. R. Garey and D. S. Johnson. Computers and Intractability. W. H. Freeman and Company, New York, 1979. Google ScholarDigital Library
- D. E. Goldberg and J. Richardson. Genetic algorithms with sharing for multimodal function optimization. In J. Grefenstette, editor, Proceedings of the Second International Conference on Genetic Algorithms, pages 41--49. Lawrence Erlbaum Associates, 1987. Google ScholarDigital Library
- J. Horn. Optimal nesting of species for exact cover of resources: Many against many. In G. Rudolph, T. Jansen, and N. Beume, editors, Parallel Problem Solving From Nature -- PPSN X, Lecture Notes in Computer Science, volume 5199, pages 438--448. Springer, 2008 Google ScholarDigital Library
- J. Horn. Pure co-evolution for shape nesting. In J. Filipe and J.Kacprzyk, editors, International Joint Conference on Computational Intelligence 2010: Proceedings of the International Conference on Evolutionary Computation, pages 255--260. SciTePress, Portugal, October 2010.Google Scholar
- J. Horn. Resource-based fitness sharing. In J.J. Merelo Guervós, P. Adamidis, H.-G. Beyer, J.-L.Fernández-Villaca\ nas, and H.-P. Schwefel, editors, Parallel Problem Solving From Nature -- PPSN VII, Lecture Notes in Computer Science, volume 2439, pages 381--390. Springer, Berlin, Germany, 2002. Google ScholarDigital Library
- J. Horn. Coevolving species for shape nesting. In J. D. Schaffer, editor, The 2005 IEEE Congress on Evolutionary Computation (IEEE CEC 2005), pages 1800--1807. IEEE Press, Piscataway, NJ, 2005.Google ScholarCross Ref
- R. M. Karp. Reducibility among combinatorial problems. In R. E. M. J. W. Thatcher, editor, Complexity of Computer Computations, pages 85--103. Plenum Press, New York, 1972.Google ScholarCross Ref
- G. Kendall. Applying Meta-Heuristic Algorithms to the Nesting Problem Utilising the No Fit Polygon. PhD thesis, University of Nottingham, 2000.Google Scholar
- W. Schudy. Is that particular case of the 'minimum weight solution to linear equations' {problem} still np-complete? In cc.complexity theory, http://cstheory.stackexchange.com/a/3721, 2010.Google Scholar
Index Terms
- NP-completeness and the coevolution of exact set covers
Recommendations
Solving a large sudoku by co-evolving numerals
GECCO '17: Proceedings of the Genetic and Evolutionary Computation Conference CompanionRecently we introduced an approach to solving Sudoku problems with co-evolution [4]: Resource-defined Fitness Sharing for Sudoku (RFSS). The idea is to find a set of non-conflicting numerals such that every cell in the puzzle is "covered" by a numeral. ...
Shape nesting by coevolving species
GECCO '05: Proceedings of the 7th annual conference on Genetic and evolutionary computationIn this paper (full version: http://cs.nmu.edu/~jeffhorn/RFS) we extend the work of [3], which introduced a new niching algorithm, resource-based fitness sharing (RFS), and demonstrated its efficacy on shape nesting problems. RFS was applied to the ...
Optimal nesting of species for exact cover of resources: two against many
GECCO '07: Proceedings of the 9th annual conference on Genetic and evolutionary computationThe application of resource-defined fitness sharing (RFS) to shape nesting problems reveals a remarkable ability to discover tilings [7, 8]. These tilings represent exact covers for a set of resources, and can be considered a maximally sized set of ...
Comments