基于协同进化博弈的多学科设计优化

复杂系统的设计问题可以非层次分解为并行的多个子空间优化设计问题。多学科优化的迭代过程可看成子空间博弈的过程。各冲突子目标协商一致条件下,子空间合作博弈的均衡点能达成原系统的整体最优,并给出协同进化算法求解博弈的Nash均衡点的计算框架。以某型民用客机的总体优化设计为例,将其分解成气动和重量两个子空间优化。设计变量不重叠地分布于各子空间,两冲突子目标分配相同权值,线性加权组合而形成的单目标作为各子空间共同的优化目标。计算结果表明此方法是有效的。

A Coevolutionary-Game Based Method to Multidisciplinary Design Optimization

The design of complex system can be non-hierarchically decomposed into several parallel subspace design tasks,and the multidisciplinary optimization can be considered as the game of the subspaces.Based on the same object negotiated by conflict sub-objects,the equilibrium of subspace cooperative game will come to the global optimization of the original system.Coevolutionary algorithm is used to find the Nash equilibrium of the game.The preliminary design of a passenger aircraft is investigated,decomposed into two subspace design tasks ,namely,aero optimization and weight optimization.Design variables are split into each subspaee,whose object is same as linear combination of the two conflict objective functions,by setting same weight on each.The numeric result indicates that the approach is effective.