拉格朗日对偶问题
拉格朗日对偶函数
优化问题(不一定是凸的)的标准形式如下:
minimize f 0 ( x ) subject to f i ( x ) ≤ 0 , i = 1 , ⋯ , m h i ( x ) = 0 , i = 1 , ⋯ , p \begin{array}{lll} \text{minimize} & f_0(x) & \\ \text{subject to} & f_i(x) \leq 0, &i=1,\cdots,m \\ & h_i(x) = 0, &i=1,\cdots,p \end{array} minimizesubject tof0(x)fi(x)≤0,hi(x)=0,i=1,⋯,mi=1,⋯,p
拉格朗日的基本思想是在目标函数中考虑约束条件,将目标函数和约束函数的加权和作为最终的目标函数,称为拉格朗日函数:
L ( x , λ , ν ) = f 0 ( x ) + ∑ i = 1 m μ i f i ( x ) + ∑ i = 1 p ν i h i ( x ) \begin{array}{lll} & L(x,\lambda,\nu) =f_0(x) +\sum \limits_{i=1}^m \mu_i f_i(x)+\sum \limits_{i=1}^p \nu_i h_i(x)& \end{array} L(x,λ,ν)=f0(x)+i=1∑mμifi(x)+i=1∑pνihi(x)
μ i , ν i \mu_i,\nu_i μi,νi称为拉格朗日乘子。
拉格朗日对偶函数为拉格朗日函数关于 x x x取得的最小值:
g ( λ , ν ) = inf x ∈ D L ( x , λ , ν ) = inf x ∈ D ( f 0 ( x ) + ∑ i = 1 m μ i f i ( x ) + ∑ i = 1 p ν i h i ( x ) ) \begin{array}{lll} & g(\lambda,\nu)=\inf_{x\in \mathcal{D}}L(x,\lambda,\nu) =\inf_{x\in \mathcal{D}} \left(f_0(x) +\sum \limits_{i=1}^m \mu_i f_i(x)+\sum \limits_{i=1}^p \nu_i h_i(x) \right)& \end{array}