25.11.2015 Views

The MOSEK Python optimizer API manual Version 7.0 (Revision 141)

Optimizer API for Python - Documentation - Mosek

Optimizer API for Python - Documentation - Mosek

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

186 CHAPTER 15. SENSITIVITY ANALYSIS<br />

f( β )<br />

f( β )<br />

β<br />

1 0<br />

β<br />

2<br />

β<br />

β<br />

1<br />

0<br />

β<br />

2<br />

β<br />

Figure 15.1: <strong>The</strong> optimal value function f l c<br />

i<br />

(β). Left: β = 0 is in the interior of linearity interval.<br />

Right: β = 0 is a breakpoint.<br />

15.4 Sensitivity analysis for linear problems<br />

15.4.1 <strong>The</strong> optimal objective value function<br />

Assume that we are given the problem<br />

z(l c , u c , l x , u x , c) = minimize c T x<br />

subject to l c ≤ Ax ≤ u c ,<br />

l x ≤ x ≤ u x ,<br />

(15.1)<br />

and we want to know how the optimal objective value changes as li<br />

c is perturbed. To answer this<br />

question we define the perturbed problem for li c as follows<br />

f l c<br />

i<br />

(β) = minimize c T x<br />

subject to l c + βe i ≤ Ax ≤ u c ,<br />

l x ≤ x ≤ u x ,<br />

where e i is the i th column of the identity matrix. <strong>The</strong> function<br />

f l c<br />

i<br />

(β) (15.2)<br />

shows the optimal objective value as a function of β. Please note that a change in β corresponds to a<br />

perturbation in l c i and hence (15.2) shows the optimal objective value as a function of lc i .<br />

It is possible to prove that the function (15.2) is a piecewise linear and convex function, i.e. the function<br />

may look like the illustration in Figure 15.1. Clearly, if the function f l c<br />

i<br />

(β) does not change much when<br />

β is changed, then we can conclude that the optimal objective value is insensitive to changes in l c i .<br />

<strong>The</strong>refore, we are interested in the rate of change in f l c<br />

i<br />

(β) for small changes in β — specificly the<br />

gradient<br />

f ′ l c i (0),<br />

which is called the shadow pricerelated to li c.<br />

<strong>The</strong> shadow price specifies how the objective value<br />

changes for small changes in β around zero. Moreover, we are interested in the linearity interval

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!