Double-population Differential Evolution Based on Logistic Model
Huiming Xia · Journal of Information and Computational Science · 2014
For the problem that when using differential evolution to solve the optimal problems, it cannot converge so fast and will trap in a local optimal solution easily. A double-population differential evolution based on logistic model is presented. The algorithm has combined with the properties of good local searching ability, fast convergence speed of best/1 mutation strategy and good global searching ability of rand/1 mutation strategy. One population chooses rand/1 mutation strategy to be charged with exploitation tasks and the other population chooses best/1 mutation strategy to be charged with exploration tasks. The use of logistic model in the algorithm can adaptively adjust scaling factor and crossover factor during the running time, which improves global searching ability at the initial generations and local searching ability at a later time. Experiments on benchmark functions show that the proposed algorithm has advantages of fast convergence and high calculation accuracy.