Linear Programming Lecture Notes - Penn State Personal Web Server
Linear Programming Lecture Notes - Penn State Personal Web Server
Linear Programming Lecture Notes - Penn State Personal Web Server
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
This simplifies to:<br />
wB − vB = cB<br />
wN − vN = cN<br />
Let w = cBB −1 . Then we see that:<br />
(8.57) wB − vB = cB =⇒ cBB −1 B − vB = cB =⇒ cB − vB = cB =⇒ vB = 0<br />
Since we know that xB ≥ 0, we know that vB should be equal to zero to ensure complementary<br />
slackness. Thus, this is consistent with the KKT conditions.<br />
We further see that:<br />
(8.58) wN − vN = cN =⇒ cBB −1 N − vN = cN =⇒ vN = cBB −1 N − cN<br />
Thus, the vN are just the reduced costs of the non-basic variables. (vB are the reduced costs<br />
of the basic variables.) Furthermore, dual feasibility requires that v ≥ 0. Thus we see that<br />
at optimality we require:<br />
(8.59) cBB −1 N − cN ≥ 0<br />
This is precisely the condition for optimality in the simplex tableau.<br />
We now can see the following facts are true about the Simplex Method:<br />
(1) At each iteration of the Simplex Method, primal feasibility is satisfied. This is<br />
ensured by the minimum ratio test and the fact that we start at a feasible point.<br />
(2) At each iteration of the Simplex Method, complementary slackness is satisfied. After<br />
all, the vector v is just the reduced cost vector (Row 0) of the Simplex tableau.<br />
If a variable is basic xj (and hence non-zero), then the its reduced cost vj = 0.<br />
Otherwise, vj may be non-zero.<br />
(3) At each iteration of the Simplex Algorithm, we may violate dual feasibility because<br />
we may not have v ≥ 0. It is only at optimality that we achieve dual feasibility and<br />
satisfy the KKT conditions.<br />
We can now prove the following theorem:<br />
Theorem 8.12. Assuming an appropriate cycling prevention rule is used, the simplex<br />
algorithm converges in a finite number of iterations to an optimal solution to the linear<br />
programming problem.<br />
Proof. Convergence is guaranteed by the proof of Theorem 7.10 in which we show that<br />
when the lexicographic minimum ratio test is used, then the simplex algorithm will always<br />
converge. Our work above shows that at optimality, the KKT conditions are satisfied because<br />
the termination criteria for the simplex algorithm are precisely the same as the criteria in<br />
the Karush-Kuhn-Tucker conditions. This completes the proof. <br />
133