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

Create successful ePaper yourself

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

106 Eli Olinick<br />

Amaldi et al. [7] propose the following integer programming model, ACMIP, to<br />

minimize the cost of providing service to all test points in the coverage area:<br />

min ∑ aℓyℓ + λ ∑<br />

ℓ∈L m∈M<br />

dm<br />

∑ xmℓ<br />

gmℓ ℓ∈Lm<br />

gmℓ<br />

∑ ∑ dm<br />

m∈M<br />

gm j∈Lm j<br />

(5.4)<br />

∑ xmℓ = 1<br />

ℓ∈Lm<br />

∀m ∈ M, (5.5)<br />

xmℓ ≤ yℓ ∀m ∈ M,ℓ ∈ Lm, (5.6)<br />

xm j ≤ s + (1 − yℓ)βℓ ∀ℓ ∈ L, (5.7)<br />

xmℓ ∈ {0,1} ∀m ∈ M,ℓ ∈ Lm, (5.8)<br />

yℓ ∈ {0,1} ∀ℓ ∈ L. (5.9)<br />

The first term in the objective function (5.4) is the total tower usage cost <strong>and</strong><br />

the second term is the total h<strong>and</strong>set transmission power when all connections are<br />

active. The user-specified parameter λ ≥ 0 determines the trade off between infrastructure<br />

cost <strong>and</strong> power usage (energy). This trade off is discussed in Section 5.3.4.<br />

Constraint set (5.5) forces each test point to be assigned to exactly one tower location,<br />

<strong>and</strong> constraint set (5.6) links the tower location <strong>and</strong> subscriber assignment<br />

decision variables so that test points can only be assigned to selected towers. A minimum<br />

acceptable SIR is ensured for all active connections by the quality-of-service<br />

constraints (QoS) (5.7). If tower location ℓ is selected, then yℓ = 1 <strong>and</strong> (5.7) enforces<br />

the signal-to-interference requirement (5.3). However, if tower location ℓ is<br />

not selected then we do not need to enforce (5.3) at that location. Indeed, doing so<br />

may artificially constrain the solution. Therefore, the constant βℓ is selected to be<br />

large enough so that the constraint is satisfied automatically by any set of binary xmℓ<br />

values when yℓ is zero. A suitable value is<br />

where<br />

�<br />

βℓ = ∑ dm<br />

m∈M<br />

max<br />

j∈Lm\{ℓ}<br />

� gmℓ<br />

gm j<br />

max<br />

j∈Lm\{ℓ}<br />

� gmℓ<br />

gm j<br />

� �<br />

(5.10)<br />

�<br />

= 0 if Lm \ {ℓ} = /0. (5.11)<br />

Constraint sets (5.8) <strong>and</strong> (5.9) define the domains of the decision variables. Given a<br />

solution to ACMIP, we refer to the set of selected towers as ¯L ⊆ L.

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

Saved successfully!

Ooh no, something went wrong!