interior point method
motivation:大型计算时
不是第一个多项式时间算法,
如果看到很好的方向,可以直接过去,(沿边走、方向被限制),但是看多个方向,计算量增加
Key knowledge • 1.Who is in the interior? - Initial solution • 2. How do we know a current solution is optimal? - Optimality condition • 3. How to move to a new solution? - Which direction to move? (good feasible direction) - How far to go? (step-length)
good enough 时,到边附近足够多,就跳上边界点,进化purification(identify谁是nbv 设为0)
1. Who is in the interior?
2. How do we know a current solution is optimal?
in simplex method:
Q2 - How do we know a current solution is optimal?
• Basic concept of optimality:
A current feasible solution optimal if and only if
no feasible direction at this point is a good direction.
• In other words, “every feasible direction is not a good direction to move!”
在该点,能走的都不是好方向
Feasible direction • In an interior-point method, a feasible direction at a current solution is a direction allows it to take a small movement while staying to be interior feasible. 本身是一个interior feasible solution - There is no problem to stay interior if the step-length is small enough. - To maintain feasibility, we need
移动一点点,linear constraint不feasible 才能约束
A矩阵作用在向量上 =0,数学上叫做 向量落在矩阵A的零空间里,
Good direction
• In an interior-point method, a good direction at a current solutio