Interactive evolutionary computation with adaptive mutation for increasing the effectiveness of advertisement texts
Quetzali Madera, Oscar Castillo, Mário Garcia, Alejandra Mancilla · 2016
Interactive evolutionary computation systems can be used to evolve advertisement texts. Google AdWords was used as the interface that users can use to see the advertisement texts, and have a chance of being persuaded by the text into clicking them. Interactive evolutionary computation systems use humans to perform fitness evaluation in the evolutionary process, which can be applied on a genetic algorithm. This work presents a comparison between two variations of the genetic algorithm: one that uses a fuzzy inference system to dynamically adapt the mutation rate, and another that uses a fixed mutation rate. The results show that the dynamically adapted genetic algorithm in the interactive evolutionary computation system has the potential to have a better performance and lead to smaller costs in advertising campaigns.