While coevolution has many parallels to natural evolution, methods other than those based on evolutionary principles may be used in the interactive fitness setting. In this paper we present a generalization of coevolution to co-optimization which allows arbitrary black-box function optimization techniques to be used in a coevolutionary like manner. We find that the co-optimization versions of gradient ascent and simulated annealing are capable of outperforming the canonical coevolutionary algorithm. We also hypothesize that techniques which employ non-population based selection mechanisms are less sensitive to disengagement.
T. Service and D. R. Tauritz, "Co-optimization Algorithms," Proceedings of the 10th Annual Conference on Genetic and Evolutionary Computation, Association for Computing Machinery (ACM), Jul 2008.
The definitive version is available at http://dx.doi.org/10.1145/1389095.1389166
Keywords and Phrases
Gradient Ascent; Algorithms; Coevolution; Simulated annealing (Mathematics)
Article - Conference proceedings
© 2008 Association for Computing Machinery (ACM), All rights reserved.