A modified differential evolution algorithm based on a new mutation strategy and chaos local search for optimization problems
Behrouz Ahadzadeh, Mohammad Bagher Menhaj · 2014
Differential evolution (DE) algorithm is a stochastic population-based optimization algorithm, which is widely used in solving various optimization problems. It has been shown that differential evolution (DE) algorithm is an effective, efficient, reasonably fast, reliable, and robust optimizer for many real-world applications. However, like any other evolutionary algorithms, DE does not guarantee convergence to a global optimum in a finite time. Beside this, DE also suffers from some limitations like slow convergence rate, stagnation and premature convergence. In this paper, a modified differential evolution called two-step differential evolution (2sDE) based on both a new mutation strategy (DE/best-to-pbest/1) and chaos local search is proposed, which divides DE algorithm into two stages. Firstly, modified differential evolution (2sDE) runs with the new mutation strategy (DE/best-to-pbest/1) to improve the global search ability. Secondly, 2sDE runs with chaotic local search to improve the local search ability. The proposed approach balances the exploration and exploitation abilities of DE algorithm. Comparing the results of the proposed method and other algorithms over several numerical benchmarks indicates that convergence speed and accuracy of the proposed method are better than those of the other algorithms.