16.11.2012 Views

Wireless Network Design: Optimization Models and Solution ...

Wireless Network Design: Optimization Models and Solution ...

Wireless Network Design: Optimization Models and Solution ...

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.

8 The <strong>Design</strong> of Partially Survivable <strong>Network</strong>s 191<br />

M<br />

T ✇ �<br />

S<br />

O<br />

�a<br />

b<br />

c�<br />

❍❨ ❍<br />

❍<br />

✛<br />

❍<br />

�❍<br />

❜w<br />

❍❨<br />

❍❜v<br />

�<br />

❜z ✟<br />

✛✟✙ ✏ ❜y<br />

✏✏<br />

✏✏<br />

✏✮<br />

✏<br />

✛ ❜x<br />

��✐<br />

✏<br />

❅ �<br />

✏<br />

� ✏<br />

� ✏<br />

❅❅ ✏�<br />

✏<br />

✏✮<br />

Fig. 8.3 Example <strong>Network</strong><br />

home to both of these hubs. The diversity requirements of nodes v, x <strong>and</strong> w are 1, 2<br />

<strong>and</strong> 3, respectively. The arrows denote one possible assignment of cells to nodes.<br />

The formulation in Kubat et al. [9] can be expressed as (CP1) below, after defining<br />

the following. H, R, V , ci j, di, <strong>and</strong> si are as defined in Section 8.2. k j is the<br />

capacity of the path connecting hub j to the MTSO, in number of DS-0 circuits. a jl<br />

is 1 if traffic from hub j flows via hub l, <strong>and</strong> 0 otherwise. Q is a set of trees such that<br />

Ts ∩ Tt = /0 ∀Ts, Tt ∈ Q. The decision variable xi j is 1 if traffic from cell i is routed<br />

through node j, <strong>and</strong> 0 otherwise.<br />

min ∑ i∈V<br />

(CP1) ∑ i∈V<br />

∑ ci jxi j<br />

j∈R<br />

s.t. ∑ xi j = si ∀i ∈ V (8.26)<br />

j∈R<br />

∑<br />

j∈R<br />

di<br />

si<br />

a jlxi j ≤ kl ∀l ∈ H (8.27)<br />

∑ xi j<br />

j∈Tq<br />

≤ 1 ∀i ∈ V, ∀Tq ∈ Q (8.28)<br />

xi j ∈ {0, 1} ∀i ∈ V, ∀ j ∈ R (8.29)<br />

The objective function minimizes the total cost of connecting cells to hubs or the<br />

MTSO. Constraints 8.26 ensure that the diversity requirement of each cell is met<br />

by the assignments, while constraints 8.27 ensure that the capacities of the path<br />

connecting the hubs to the MTSO are not exceeded. Constraints 8.28 ensure that<br />

any given cell connects to at most one hub across all the hubs in any given tree.<br />

Kubat et al. [9] solve problem (CP1) via three heuristic techniques. Readers are<br />

referred to Kubat et al. [9] for additional details.<br />

8.4.1 A Partitioning Formulation<br />

The cell assignment problem described in this paper can be viewed as a partitioning<br />

problem where the cells are partitioned into distinct sets <strong>and</strong> assigned to various<br />

trees (or individual hubs, if there is only one hub in a given tree) in such a way that

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

Saved successfully!

Ooh no, something went wrong!