“…Besides, the two mentioned approaches use computational costly methods of Gibbs sampling and the belief propagation algorithm, which makes them completely incomparable in complexity with the introduced approach, which is polynomial in terms of time and number of fitness evaluations. Among all the EDA approaches, the hBOA (Pelikan, 2005) and the DSMGA (Yu and Goldberg, 2006) are selected as reference algorithms. The former is used because, first, it is a reputed and well-studied approach in the community and, second, it also learns a graphical model (Bayesian network) in order to represent the variable dependencies, so comparing the results of these two approaches may give the reader an evaluation on how these two graphical models work in contrast to each other.…”