You are on page 1of 3

\label{7.

2}
\textbf{(7.2)}
Relationship Between optimal dual solutions of (Q) and (M)
\[
\tag{Q}
\begin{array}{rrcl}
\min & c'x &
& \\
& Ex & = h \\
& Ax & = & b ~; \\
& x & \geq & \mathbf{0}~.
\end{array}
\]
\[
Let S := { x \in \mathbb{R}^n : Ax=b , x \geq 0 } , \\
let \mathcal{X} : = { x^j : j \in \mathcal {J} } be the set of basic-feasible so
lutions S , \\
and let \mathcal {Z} := { z^k : k \in \mathcal {K} be the set of non basic-feasi
ble directions of S.
\]
Let the dual variables be $\pi$ and $y$
The dual for $ (Q) $ can be given as
\[
\tag{D_Q}
\begin{array}{rrcl}
\max & y'b + \pi' h &
& \\
& y'A + \pi'E & \leq & c'~.\\
\end{array}
\]
Now consider the master Problem
\[
\tag{M}
\begin{array}{rrcl}
\min & \sum j \in \mathcal{J} (c'{x^j}) {\lambda_j} + \sum k \in \mathcal{K}
(c'{z^k}) {\mu_k} &
& \\
& \sum j \in \mathcal{J} (E{x^j}) {\lambda_j} + \sum k \in \mathcal{K} (
E{z^k}) {\mu_k} & = h \\
& \sum j \in \mathcal{J} {\lambda_j} & = & 1 ~; \\
& {\lambda_j} & \geq & \mathbf{0} j \in \mathcal{J} ~.
& {\mu_k} & \geq & \mathbf{0} k \in \mathcal{K} ~.
\end{array}
\]
This can be written as follows :
\[
\left(
\begin{array}{c_x}
c'x^1 \\
c'x^2 \\
.
\\

.
\\
.
\\
.
\\
c'x^j \\
\end{array}
\right)
\]
\[
\left(
\begin{array}{c_z}
c'z^1 \\
c'z^2 \\
.
\\
.
\\
.
\\
.
\\
c'z^k \\
\end{array}
\right)
\]
\[
\left(
\begin{array}{x_1}
\lambda_1 \\
\lambda_2 \\
.
\\
.
\\
.
\\
.
\\
\lambda_j \\
\end{array}
\right)
\]
\[
\left(
\begin{array}{x_2}
\mu_1 \\
\mu_2 \\
.
\\
.
\\
.
\\
.
\\
\mu_k \\
\end{array}
\right)
\]
therefore the problem can be re-written as :
\[
\min & {c'}_x x_1 + {c'}_z x_2
\[ \left( \begin{array}{cc}

&

&

\\

A=\left[
\begin{array}{cc}
E{x^j} &\tempE{z^k} \\ \hline
I &\temp0
\end{array}\right]
$$\\
\end{array} \right) \times \left( \begin{array}{cc}
x_1 & \\
x_2 &
\end{array} \right) =\left( \begin{array}{cc}
h & \\
a &
\end{array} \right)\]

You might also like