As performance of traditional genetic algorithm is inadequate at application, a kind of parallel genetic algorithm for multi-strategy is proposed to increase population diversity and improve robustness of the algorithm. The paper uses 8 different strategies for branch genetic manipulation of genetic algorithm to carry out group optimization, the branch groups use adaptive migration strategies to realize reasonable relocation of the best individual, and the fitness of the branch group is taken as the termination conditions for the system to realize group optimization. Experimental results shows that the robustness, convergence, stability and optimization capacity than the traditional genetic algorithm is greatly improved, it has a strong application performance.%针对传统遗传算法工程应用性能不足的特点,提出一种多策略并行遗传算法以增加算法的群体多样性及提高算法的鲁棒性；采用8个不同策略为并行遗传算法的分支遗传操作进行群体寻优,并将分支的群体采用自适应迁移策略实现最优个体的合理迁移,以分支群体的适应度为算法终止条件,实现群体寻优；利用连续单峰函数、连续多峰函数等多个函数进行平均收敛代数等参数进行算法测试,结果表明算法的鲁棒性、收敛性、稳定性及优化能力较传统遗传算法有较大的提高,该算法具有较强的工程应用性能.
展开▼
