Title :
Faster Evolutionary Algorithms by Superior Graph Representation
Author :
Doerr, Benjamin ; Klein, Christian ; Storch, Tobias
Author_Institution :
Max-Planck-Inst. fur Informatik, Saarbrucken
Abstract :
We present a new representation for individuals in problems that have cyclic permutations as solutions. To demonstrate its usefulness, we analyze a simple randomized local search and a (1+1) evolutionary algorithm for the Eulerian cycle problem utilizing this representation. Both have an expected runtime of Theta(m2 log(m)), where m denotes the number of edges of the input graph. This clearly beats previous solutions, which all have an expected optimization time of Theta(m3 or worse (PPSN ´06, CEC ´04). We are optimistic that our representation also allows superior solutions for other cyclic permutation problems. For NP-complete ones like the TSP, however, other means than theoretical run-time analyses are necessary
Keywords :
computational complexity; evolutionary computation; graph theory; randomised algorithms; search problems; Eulerian cycle problem; NP-completeness; cyclic permutations; evolutionary algorithms; input graph; randomized local search; superior graph representation; Algorithm design and analysis; Approximation algorithms; Computational intelligence; Computer science; Evolutionary computation; Genetic mutations; Resonance light scattering; Runtime; Traveling salesman problems; Tree graphs;
Conference_Titel :
Foundations of Computational Intelligence, 2007. FOCI 2007. IEEE Symposium on
Conference_Location :
Honolulu, HI
Print_ISBN :
1-4244-0703-6
DOI :
10.1109/FOCI.2007.372176