稳态大系统多目标优化的双环递阶算法

Double Loop Hierarchical Algorithm for Multiobjective Optimization Problems of Steady-state Large Scale Systems

对于稳态大规模工业系统，提出了一种能够充分满足决策者的偏好克服目标函数不可分性的双环迭代算法．该算法首先根据决策人员的偏好把多目标问题聚集为不可分的单目标优化问题，再嵌入到可分的多目标优化问题中，然后从多目标的非劣解集中挑出不可分单目标的最优解，并证明了该最优解就是原问题的非劣解，算法的收敛性也有严格的数学保证．

This paper proposes a double-loop hierarchicalalgorithm for m ultiple objec- tivesoptim ization. First, the algorithm assem blesthem ultipleobjectives into oneobjec- tive by thepreference ofdecision-m aker, and then em bedsthenonseparableoneobjective optim ization problem into separablem ultiobjectiveoptim ization problem .Finally,theop- tim alsolution ofthe nonseparable problem can be obtained from the setofsolutions of the separable m ultiobjective optimization problem , this optim al solution is provided as one non-inferiorsolution ofthe originalproblem .