In order to solve the premature convergence problem existing in the traditional estimation of distribution algorithm (EDA)

based on the analysis of methods for diversity preservation and reasons for premature convergence

an estimation of distribution algorithm with diversity preservation (EDA-DP) is proposed. A chaotic mutation operator is introduced into EDA by taking advantage of the randomness

ergodicity

initial value sensitivity and regularity of chaos. The EDA-DP is able to adjust its mutation radius in an adaptive way according to the fitness value and the distance between each individual. Moreover

the EDA-DP is able to generate its offspring population by making use of the concentration inside the population. The EDA-DP is evaluated on a set of benchmark problems and the experimental results show that the precision of the optimal solutions and the convergence speed are improved

thanks to the EDA-DP effectively overcomes the premature convergence problem.