主要研究上层函数及其约束函数不要求具有凸性和可微性,下层是关于下层决策变量是凸二次规划的双层规划模型,通过Karush-Kuhn-Tucher条件转化为一个单层规划,利用下层是正定二次规划,将下层的决策变量表示为关于Lagrangian乘子的表达式,从而降低了搜索空间的维数,设计了遗传算法,并通过数值实验表明该遗传算非常有效。
In terms of the Karush-Kuhn-Tucher conditions of convex programming,a special bi-level programming problem,whose flower-level problem is a definite quadratic programming,is transformed into an equivalent single-level programming,because the quadratic programming is definite,the variables of flower-level can be solved,decrease the dimensions of the search space,a new crossover operator is designed,the experimental studies show that the new solution algorithm can be used to solve the special bilevel programming model is effective.