2018-04-18 开胃学习数学系列 - Duality
2018-04-19 本文已影响0人
Kaiweio
KKT
gradient is 0
x* is the local minimizer
f(x*) = f(μ*, λ*)
= uf(x) + μ*h(x) + λ*g(x)
≤ f(x*) + μ*h(x*) + λ*g(x)
第一项是0,所以第二项也必须是0
≤ f(x*)
KKT
gradient is 0
x* is the local minimizer
f(x*) = f(μ*, λ*)
= uf(x) + μ*h(x) + λ*g(x)
≤ f(x*) + μ*h(x*) + λ*g(x)
第一项是0,所以第二项也必须是0
≤ f(x*)