Measuring and Promoting Diversity in Evolutionary AlgorithmsWhile divergence of character is a cornerstone of the Darwinian theory of natural evolution, the lack of divergence is an oppressive burden of artificial evolution. The problem has been sometimes labeled with the oxymoron premature convergence, implying that the algorithms would converge to better solutions if prevented from exploiting strongly attractive local optima. Different names have been given by different researchers, and different effects have been observed on different paradigms. Almost all deleterious. Tutorial at PPSN 2016.
MPDEA Workshop at GECCO 2016.
MPDEA Workshop at GECCO 2017.