大规模系统的全局优化

Global Optimization for Large Scale Systems

本文讨论了可分非凸大规模系统的全局优化控制问题 .提出了一种 3级递阶优化算法 .该算法首先把原问题转化为可分的多目标优化问题 ,然后凸化非劣前沿 ,再从非劣解集中挑出原问题的全局最优解 .建立了算法的理论基础 ,证明了算法的收敛性 .仿真结果表明算法是有效的 .

Considered is the global optimization problem for separable non-convex large-scale in this paper. A three-level hierarchical optimization algorithm is proposed. It converts the original problem into a separable multiobjective optimization problem. The noninfreior frontier is then convexified and the global optimal solution of the original problem is selected from the set of noninferior solutions of multiobjective optimization problem. Theoretical base of the algorithm is established. Simulation shows that the algorithm is effective.