Optimization Week 3: Programming (convex program, linear program)

最新推荐文章于 2024-06-06 20:24:48 发布

xiwang_chn

最新推荐文章于 2024-06-06 20:24:48 发布

阅读量238

点赞数

分类专栏： Optimization

本文链接：https://blog.csdn.net/weixin_42017454/article/details/110950643

版权

15 篇文章 1 订阅

订阅专栏

1 Convex programming

$\begin{aligned} \min_x& \quad f(x)\\ s.t.& \quad x\in C \end{aligned}$ is convex if $f$ is a convex function and $C$ is a convex set.

Local Optimality $\nRightarrow$ Global Optimality
If convex programming, Local Optimality $\Leftrightarrow$ Global Optimality
Unconstrained programming, Local Optimality $\Leftrightarrow$ $\nabla f=0$ $\Leftrightarrow$ Global Optimality

$\begin{aligned} \min_x& \quad c^Tx\\ s.t.& \quad Ax\leq b\\ & \quad Cx=d \end{aligned}$

Feasible set of LP is a Polyhedron
LP can be :
- No optimal solution
  - Infeasible
  - Unbounded
- Finite optimal solution

Feasible set $P=\{x: Ax\leq b, Cx=d\}$

Definition 1: Not the convex combination of any other two other points in the polytope.
If $\exists y,z \in P$ and $\lambda \in [0,1]$ , such that $x=\lambda y+(1-\lambda) z$ , then $x$ is not extreme.
Definition 2: It is the unique optimum for some cost vector $c$ .

Active constraints: Constraint $a_i$ is active at $x$ if $a_i^Tx=b_i$ .
Active set: $\mathcal{A}_x=\{a_i: a_i^Tx=b_i\}\cup\{d_i: c_i^Tx=d\}$ .
Basic feasible solution (BFS): $x$ is BFS is its active set $\mathcal{A}_x$ has n linearly independent vectors.

The following are equivalent:
(1) $x$ is a BFS;
(2) $x$ satisfies definition 1 of BFS;
(3) $x$ satisfies definition 2 of BFS;
Possible to have no extreme point:
P has no extreme point $\Leftrightarrow$ P contains a full line.