Interactive Evolution Strategies for Minimizing Single-objective Functions

This article discusses issues related to the application of the user interactivity in optimization process, aided by the method of the Evolution Strategies, with the objective of improve the minimization process of the evaluated functions, both in their response and in the time of obtaining, through the intervention of the user. Its development is based on the methodology of the Evolution Strategies, characterized by the concept of natural selection. However, the minimization process requires an intervention in the stagnations of local minima to get a better efficiency. To increase the applicability of the approached methodology, it is proposed the use of the user interactivity in minimization process, allowing the manipulation of mutation rate and recombination, activation of the mutation of self-adaptive parameters as well as monitoring the whole process. The proposed methodology was used in the evaluation of test functions for optimization, as well as in the evaluation of a real case.