Are deception and complexity conjugate variables in genetic learning?
Luis R. Lopez · 2002
This work provides an analytic starting point to the question: How deceptive is a randomly selected problem? It is shown that the bounding complexity of a large trap function is inversely proportional to the probability of a genetic algorithm encountering a fully deceptive instance, independent of problem size for gene length greater than 10/sup 4/. This result brings up interesting insights about the relationship between deception and complexity.>