Managing Diversity and Many Objectives in Evolutionary Design

by Sheikh Faishal Basher

MAP-Elite solutions after a run.

A new approach to evolving a diversity of high quality solutions for problems having many objectives is presented. Mouret and Clune’s MAP-Elites algorithm has been proposed as a way to evolve an assortment of diverse solutions to a problem. We extend MAP-Elites in a number of ways. We introduce into MAP-elites the many-objective strategy called sum-of-ranks, which enables problems with many objectives (4 and more) to be considered in the MAP. We also enhance MAPElites by extending it with multiple solutions per cell (the original MAP-Elites saves only a single solution per cell). Different ways of selecting cell members for reproduction are also considered. We test the new MAP-Elites strategies on the evolutionary art application of image generation. Using procedural textures, genetic programming is used with upwards of 15 light-weight image features to guide fitness. The goal is to evolve images that share image features with a given target image. Our experiments show that the new MAP-Elites algorithms produce a large number of diverse solutions, which can be competitive in quality to those from standard GP runs. The technique can be used in a wide variety of applications.

Downloads and publications

Managing Diversity and Many Objectives in Evolutionary Design, Sheikh Faishal Basher, MSc Thesis, 2021.
S.F. Basher and B.J. Ross,
2022 IEEE Congress on Evolutionary Computation (CEC), 2022, Padua, Italy, 2022.

Back up: http://www.cosc.brocku.ca/~bross/