20.05.2014 Views

link to my thesis

link to my thesis

link to my thesis

SHOW MORE
SHOW LESS

Create successful ePaper yourself

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

44 CHAPTER 4. GLOBAL OPTIMIZATION OF MULTIVARIATE POLYNOMIALS<br />

4.1 The global minimum of a Minkowski dominated polynomial<br />

Let q(x 1 ,...,x n ) ∈ R[x 1 ,...,x n ] be a real polynomial. We are interested <strong>to</strong> compute<br />

its infimum over R n . Let d be a positive integer such that 2d (strictly) exceeds the<br />

<strong>to</strong>tal degree of q(x 1 ,...,x n ) and consider the one-parameter family of what will be<br />

called (Minkowski) dominated polynomials:<br />

p λ (x 1 ,...,x n ):=λ(x 2d<br />

1 + ...+ x 2d<br />

n )+q(x 1 ,...,x n ), λ ∈ R + . (4.1)<br />

Note that the nomenclature for this family derives from the property that the<br />

value of p λ is dominated by the term λ(x 2d<br />

1 + ...+ x 2d<br />

n ) when the Minkowski 2d-norm<br />

‖(x 1 ,...,x n )‖ 2d =(x 2d<br />

1 + x 2d<br />

2 + ... + x 2d<br />

n ) 1/(2d) becomes large. Consequently, the<br />

polynomial p λ has a global minimum over R n for each λ ∈ R + : the presence of the<br />

dominating term λ(x 2d<br />

1 + ...+ x 2d<br />

n ) ensures that p λ has a global minimum. In fact<br />

information about the infimum (and its ‘location’) of an arbitrary real polynomial<br />

q(x 1 ,...,x n ) can be obtained by studying what happens <strong>to</strong> the global minima and<br />

the corresponding minimizing points of p λ (x 1 ,...,x n ) for λ ↓ 0, see [48], [64].<br />

Note that there may be more than one point at which the value of the global<br />

minimum is obtained.<br />

Example 4.1. The function: p λ (x 1 ,x 2 )=λ(x 6 1+x 6 2)+10x 2 1x 2 2−10x 2 1−10x 2 2+5 with λ =<br />

0.5 has multiple points at which the global minimum is obtained, as shown in Figure<br />

4.1. The global minimum value of −12.2133 is attained at 4 positions ((±1.6069, 0)<br />

and (0, ±1.6069)). The method proposed in this chapter finds at least one point in<br />

every connected component of the set of points where this global minimum is attained.<br />

Figure 4.1: Example of a polynomial with multiple locations of the global optimum

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

Saved successfully!

Ooh no, something went wrong!