The paper deals with the adaptive mechanisms in differential evolution (DE) algorithm. DE is a simple and effective stochastic algorithm frequently used in solving the real-world global optimization problems. The efficiency of the algorithm is sensitive to setting its control parameters. Several adaptive approaches have appeared recently in order to avoid control-parameter tuning. A new adaptive variant of differential evolution is proposed in this study. It is based on a combination of two adaptive approaches published before. The new algorithm was tested on the well-known set of benchmark problems developed for the special session of CEC2005 at four levels of population size and its performance was compared with the adaptive variants that were applied in the design of the new algorithm. The new adaptive DE variant outperformed the others in several test problems but its efficiency on average was not better.
We propose a neural network approach for global optimization with applications to nonlinear least square problems. The center idea is defined by the algorithm that is developed from neural network learning. By searching in the neighborhood of the target trajectory in the state space, the algorithm provides the best feasible solution to the optimization problem. The convergence analysis shows that the convergence of the algorithm to the desired solution is guaranteed. Our examples show that the method is effective and accurate. The simplicity of this new approach would provide a good alternative in addition to statistics methods for power regression models with large data.
The differential evolution (DE) algorithm is a powerful population-based stochastic technique to search for global optimum in the continuous search space. Success of DE algorithm strongly depends on choosing its parameters. The competition in differential evolution was shown to be an efficient instrument to avoid time-consuming process of tuning control parameters. A new variant of competitive DE algorithm, called BEBERAN, was proposed and tested on benchmark functions at four levels of the search space dimension. The BEBERAN was compared with the most promising competitive variant, DEBR18. BEBERAN, in contrast to DEBR18, includes in addition the exponential crossover.