20.04.2014 Views

design space pruning heuristics and global optimization method for ...

design space pruning heuristics and global optimization method for ...

design space pruning heuristics and global optimization method for ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

1.2.3 Combinatorial Optimization<br />

Choosing the optimal asteroid sequence in an asteroid tour mission <strong>design</strong><br />

problem is by nature a combinatorial <strong>and</strong> integer <strong>optimization</strong> problem. The<br />

distinguishing feature of such problems is that the variables belong to a discrete set where<br />

there is not a continuum of alternatives. One can write the linear integer programming<br />

problem, which has no continuous variables, as follows 65 :<br />

{ cx : Ax b,x ∈ Z<br />

n }<br />

max ≤<br />

+<br />

(12)<br />

In Equation 12, Z n + is the set of nonnegative integral n-dimensional vectors, <strong>and</strong> x =<br />

(x 1 ,…,x n ) are the variables or unknowns. An instance of the problem is specified by the<br />

data (c, A, b), where c is an n-dimensional vector, A is an m x n matrix, <strong>and</strong> b is an m-<br />

dimensional vector. While this represents the general problem, it can be altered in a<br />

number of ways to represent myriad types of discrete <strong>optimization</strong> problems. One<br />

example of a <strong>for</strong>mulation of Equation 12 is the 0-1 knapsack problem, which is one of the<br />

most basic <strong>and</strong> common problems in combinatorial <strong>optimization</strong>. This problem deals<br />

with choosing a subset of projects to maximize the sum of their values while not<br />

exceeding some budget constraint. From Equation 12, the projects are represented by the<br />

variables x 1 ,…x n , where a value of 1 indicates that project j is in the subset <strong>and</strong> a value of<br />

0 indicates that it is not. The jth project has a cost of a j <strong>and</strong> a value of c j , <strong>and</strong> b represents<br />

the overall budget constraint.<br />

Another common type of combinatorial <strong>optimization</strong> problem, which is more<br />

directly applicable to the asteroid tour problem, is the traveling salesman problem. The<br />

classic version of this problem deals with solving the following scenario: given a set of<br />

cities <strong>and</strong> the distance between each pair of cities, determine the shortest route to visit<br />

each city exactly once while returning to the city of origin. 66<br />

Solution <strong>method</strong>s generally<br />

fall into two categories: algorithms <strong>for</strong> finding exact solutions <strong>and</strong> heuristic algorithms,<br />

27

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

Saved successfully!

Ooh no, something went wrong!