改进的多岛遗传算法在动力总成悬置系统优化设计中的应用

摘要

Aiming at the demerits of Multi-Island Genetic Algorithm (MIGA), the dynamic modulation of crossover and variable probability is employed to overcome the premature phenomenon and improve the searching efficiency of this al-gorithm. Combining with the powerful local searching ability of the simulated annealing algorithm, an improved MIGA with higher efficiency is worked out. Its optimization effect is verified by simulation and experiment. This improved algorithm can be effectively applied in the optimization of powertrain mounting system.%针对多岛遗传算法（Multi-Island GA）的缺点，通过动态调节其交叉和变异概率来防止算法的早熟以及提高搜索效率，并结合模拟退火法的局部寻优能力强的优点，形成了求解效率更高的改进的多岛遗传算法，并将其应用于动力总成悬置系统的优化设计中。并通过仿真分析和实验验证了优化效果，该改进的算法可以有效应用在悬置系统的优化设计中。