Information Technology Reference
In-Depth Information
Fig. 4. Flow chart for power allocation
Integrate the chromosome: Owing to the separation of the fitness function calcula-
tion, the chromosome must be integrated after the evaluation terminates, to be in-
cluded in the next procedure.
This step is the same as that for sub-channel allocation, including the operations of
selection, crossover, and mutation.
When the algorithm terminates, system power has been rationally allocated among
the users and the allocation scheme satisfies proportional fairness among users and
maximum system capacity. The simulation results are presented in Section 4.
4
Simulation Results
In this section, we present our simulation results to validate the effectiveness of the
proposed resource allocation strategy. A system with 64 sub-channels and 8 users was
used. The wireless channel was modeled as a 6-tap frequency-selective Rayleigh
channel employed in [8], and the total power tota P and available bandwidth were set
to 1 Watt and 1 MHz, respectively. The power spectral density for noise was set to -
80 dB/Hz. The parameters in the genetic algorithms for sub-channel allocation and
power allocation were the same, that is, W=100, Ps=0.9, Pc=0.7, Pm=0.035, and
Gen=300.
Fig. 5 shows the total system capacity in terms of sub-channel allocation, power al-
location, and maximum capacity. Owing to the fact that the sub-channel and power
are both assigned to the user with the best channel gain, the maximum capacity me-
thod achieves maximum capacity. However, it also results in other users in the system
having a zero data rate as shown in Fig. 6. The method with sub-channel allocation
achieves slightly less than maximum capacity, but it introduces a rough rate constraint
that ensures that all users can acquire sub-channels. The system capacity after power
allocation is close to the sub-channel allocation. Since the method needs to realize
proportional fairness at the expense of capacity loss, it has little effect on system ca-
pacity performance.
Fig. 6 shows the normalized ergodic capacity distribution among users for the dif-
ferent methods. Here, the data rate ratio is
γ== and
4
γγ
==== . It
...
γ
1
1
2
3
4
8
 
Search WWH ::




Custom Search