一种具有全局稳定性的多学科协同优化方法

针对协同优化对初始点选取敏感和易收敛到局部极值点的问题,提出了一种具有全局稳定性的协同优化方法.该方法采用全局优化和局部优化两阶段优化策略.在全局优化阶段,系统级优化采用较大的松弛因子,在保证学科间一致性的前提下,将初始点优化到伞局极值点附近;在局部优化阶段,逐步减少松弛因子,增强学科间的一致性,得到全局极值点.两个典型的优化算例表明,优化结果不受初始点的影响,有效增强了协同优化的全局稳定性,并具有较好的可行性和收敛速度.

Improved multidisciplinary collaborative optimization with global stability

To solve the problems that the collaborative optimization results were sensitive to the initial points and usually converged to the local extremes, a new collaborative optimization method with global stability was presented. Two-phase optimization strategy was adopted in this method. In the global optimization phase, a bigger slack factor was employed in system-level optimization to reach the neighborhood of global extreme while ensuring the requirements of interdisciplinary consistency. In the local optimization phase, the slack factor was reduced gradually and the interdisciplinary compatibility was strengthened so as to reach to the global extreme. Two typical examples were adopted to test this optimization method. The results showed that optimization result was not influenced by initial point, the global stability of collaborative optimization was improved. Meanwhile, the presented approach had satisfactory feasibility and convergence speed.