Last Friday we were discussing the paper by Tirronen and Weber, “Sparkline Histograms for Comparing Evolutionary Algorithms“, presented at IJCCI 2010. This work proposes histograms that represent the distribution of values in the set of results for comparing the performance of Evolutionary Algorithms. Such a visual comparison provides a quick evaluation of the relative performance of the algorithms in a test set, as well as the overall performance of each one. Some patterns of the histograms permit to identify features such as lack of robustness, high rate of convergence to local optima, and high standard deviation. Altough this method does not replace numerical and statistical comparisons, it may be helpfull in the analysis of Evolutionary Algorithms, so we suggest you to take a look at the work of Tirronen and Weber.