02.08.2013 Views

Model Predictive Control

Model Predictive Control

Model Predictive Control

SHOW MORE
SHOW LESS

Create successful ePaper yourself

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

46 4 Linear and Quadratic Optimization<br />

c ′ z = k4<br />

(a) Linear Program -<br />

Case 1<br />

P<br />

c ′ z = k1<br />

z ∗<br />

c ′ z = k1<br />

c ′ c<br />

z = k4<br />

′ z = k3<br />

P<br />

(b) Linear Program -<br />

Case 2<br />

c ′ z = k4<br />

z ∗<br />

c ′ z = k1<br />

P<br />

(c) Linear Program -<br />

Case 3<br />

Figure 4.1 Graphical Interpretation of the Linear Program Solution, ki <<br />

ki−1<br />

not empty. Denote by J ∗ the optimal value and by Z ∗ the set of optimizers of<br />

problem (4.1)<br />

Z ∗ = argmin z∈P c ′ z<br />

Three cases can occur.<br />

Case 1. The LP solution is unbounded, i.e., J ∗ = −∞.<br />

Case 2. The LP solution is bounded, i.e., J ∗ > −∞ and the optimizer is unique.<br />

Z ∗ is a singleton.<br />

Case 3. The LP solution is bounded and there are multiple optima. Z ∗ is an<br />

uncountable subset of R s which can be bounded or unbounded.<br />

The two dimensional geometric interpretation of the three cases discussed above<br />

is depicted in Figure 4.1. The level curves of the cost function c ′ z are represented<br />

by the parallel lines. All points z belonging both to the line c ′ z = ki and to<br />

the polyhedron P are feasible points with an associated cost ki, with ki < ki−1.<br />

Solving (4.1) amounts to finding a feasible z which belongs to the level curve with<br />

the smallest cost ki. Since the gradient of the cost is c ′ , the direction of steepest<br />

descent is −c ′ .<br />

Case 1 is depicted in Figure 4.1(a). The feasible set P is unbounded. One can<br />

move in the direction of steepest descent −c and be always feasible, thus decreasing<br />

the cost to −∞. Case 2 is depicted in Figure 4.1(b). The optimizer is unique and<br />

it coincides with one of the vertices of the feasible polyhedron. Case 3 is depicted<br />

in Figure 4.1(c). The whole bold facet of the feasible polyhedron P is optimal,<br />

i.e., the cost for any point z belonging to the facet equals the optimal value J ∗ .<br />

In general, the optimal facet will be a facet of the polyhedron P parallel to the<br />

hyperplane c ′ z = 0.<br />

From the analysis above we can conclude that the optimizers of any bounded<br />

LP always lie on the boundary of the feasible polyhedron P.

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

Saved successfully!

Ooh no, something went wrong!