30.06.2015 Views

Opportunistic Scheduling in Cellular Systems in the Presence of ...

Opportunistic Scheduling in Cellular Systems in the Presence of ...

Opportunistic Scheduling in Cellular Systems in the Presence of ...

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.

1772 IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 58, NO. 3, MARCH 2012<br />

ODEs has a unique solution for any f<strong>in</strong>ite time. Fur<strong>the</strong>rmore,<br />

<strong>the</strong> solution is bounded for any f<strong>in</strong>ite time. Indeed, one has<br />

Def<strong>in</strong>e<br />

. One can rewrite <strong>the</strong> objective<br />

function <strong>in</strong> <strong>the</strong> above optimization problem as<br />

(43)<br />

with be<strong>in</strong>g <strong>the</strong> upper bound on function , and hence<br />

From (42), we have<br />

This concludes <strong>the</strong> pro<strong>of</strong> <strong>of</strong> <strong>the</strong> lemma.<br />

APPENDIX D<br />

PROOF OF LEMMA 2<br />

Under <strong>the</strong> given hypo<strong>the</strong>sis, <strong>the</strong> BS’s policy from (30) and<br />

(31) would be . From (27), <strong>the</strong> pay<strong>of</strong>f for mobile 1<br />

under policy is<br />

Hence, <strong>the</strong> maximum <strong>of</strong> <strong>the</strong> objective function <strong>in</strong> (43) is<br />

achieved under <strong>the</strong> required constra<strong>in</strong>ts by first maximiz<strong>in</strong>g<br />

<strong>the</strong> term (<strong>the</strong> constra<strong>in</strong>ts on this alone will be less strict)<br />

subject to<br />

Def<strong>in</strong>e <strong>the</strong> probability measure<br />

for all possible result<strong>in</strong>g from and . Also<br />

def<strong>in</strong>e<br />

to obta<strong>in</strong><br />

, and <strong>the</strong>n maximiz<strong>in</strong>g <strong>the</strong> term (while<br />

ensur<strong>in</strong>g that <strong>the</strong>se variables and <strong>the</strong> optimal variables from <strong>the</strong><br />

previous step jo<strong>in</strong>tly satisfy <strong>the</strong> required constra<strong>in</strong>ts)<br />

subject to<br />

By <strong>in</strong>dependence, we are <strong>in</strong>terested <strong>in</strong> <strong>the</strong> constra<strong>in</strong>ed optimization<br />

problem:<br />

If with ,<br />

<strong>the</strong>n clearly by def<strong>in</strong>ition <strong>of</strong> <strong>the</strong> best strategies ,wehave<br />

(42)<br />

for all , and so on. Assum<strong>in</strong>g condition ,it<br />

is easily seen that (see <strong>the</strong> first equation at <strong>the</strong> top <strong>of</strong> <strong>the</strong> page).<br />

Fur<strong>the</strong>r (see <strong>the</strong> second equation at <strong>the</strong> top <strong>of</strong> <strong>the</strong> page). The<br />

above def<strong>in</strong>es <strong>the</strong> jo<strong>in</strong>t distribution with prescribed<br />

marg<strong>in</strong>als. It is now straightforward to see that <strong>the</strong> conditional<br />

distribution is <strong>in</strong>deed . This<br />

completes <strong>the</strong> pro<strong>of</strong>.<br />

REFERENCES<br />

[1] 1xEV:1x Evolution IS-856 TIA/EIA Standard Airl<strong>in</strong>k Overview, Qualcomm,<br />

Inc., 2001.<br />

[2] High Speed Downl<strong>in</strong>k Packet Access (HSDPA); Overall Description;<br />

Stage 2 Technical Specification 3rd Generation Partnership Project;<br />

Technical Specification Group Radio Access Network, 2008, 3GPP TS<br />

25.308, Release 8, V8.3.0.

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

Saved successfully!

Ooh no, something went wrong!