一种改进的协同优化过程研究

在分析协同优化过程基本流程的基础上,针对现有过程中系统级协调算法计算量大、容易发散等缺点,提出了动态罚因子算法,该算法计算量小,优化效率高。采用正交试验设计选取初始设计向量,加快了优化收敛速度。用经典函数和齿轮减速箱优化问题对改进后的优化过程进行测试,算例结果验证了改进措施的有效性。

Study on a kind of improved collaborative optimization

Based on the analysis of the basic process of collaborative optimization(CO),two improvements are presented.Firstly,to lessen the large computation burden and to avoid the divergence of the system-level optimization,a technique named as dynamic penalty factors is used,which has higher optimization efficiency,secondly,the orthogonal experimental design(OED) is used to choose the initial design vector to converge more quickly.Two examples are adopted to test the improved CO: a typical function and a gear reducer box design problem.The results show that the two improvements are very effective.