摘要
对于分目标为极大函数的一类不可微多目标决策问题，给出了一个非标量化可行方向法模型，并在较弱的条件下证明了其全局收敛性。以此为依据，讨论了模型中有关参量的取法，从而得到了几类可实现的具体算法。这些方法可以将决策者的偏好反映到对各目标的改善上和算法的结构上，因此不仅具有直观、计算简单等优点，而且具有应用灵活的特点。
For a class of nondifferentiable multiobjectve decision making problems with a maximalfunction for the objective,a model for non-scalarized feasible-direction methodis proposed. The global convergence of the model is proved under rather weak conditions.The selectionof the parameters in the model is discussed and several implementable algorithms have beenworked out. In the method proposed,the preference of the decision maker can be well repre-sented in the improvement of each objective and in the structure of the algorithms,
出处
《华中理工大学学报》
CSCD
北大核心
1995年第A01期75-81,共7页
Journal of Huazhong University of Science and Technology
关键词
不可微多目标
决策问题
收敛
nondifferentiable multiobjective decision making
feasible-direction method
convergence
