协进化差分进化算法及其水分布网络优化设计应用

In this paper a new DE algorithm, i.e. self-adaptive co-evolution differential evolution (SACDE), is proposed to improve the optimization performance. SACDE employs the idea of co-evolution, which means that population and control parameters are coded together. While individuals are evolving, control parameters are changing in coordinately. The automatically updated parameters can adapt to different optimization problem and thus increase the robustness of the algorithm. Simulation results show that the proposed algorithm is better than classic DE algorithm. In comparison with other self-adaptive DE algorithm, the proposed algorithm is better in some conditions. Finally, the proposed algorithm is employed to optimize a discrete complex constrained optimization problem-water distribution network problem. SACDE achieves better results in benchmark water distribution network problems than previous methods. It demonstrates that SACDE not only can handle discrete constrained optimization problem but also shows its robustness due to the exponentially increasing complexity of different water distribution networks.