Research Article A Resource Allocation Evolutionary ...
Transcript of Research Article A Resource Allocation Evolutionary ...
Hindawi Publishing CorporationMathematical Problems in EngineeringVolume 2013, Article ID 406143, 8 pageshttp://dx.doi.org/10.1155/2013/406143
Research ArticleA Resource Allocation Evolutionary Algorithm for OFDM Basedon Karush-Kuhn-Tucker Conditions
Hai-Lin Liu and Qiang Wang
School of Applied Mathematics, Guangdong University of Technology, Guangzhou 510520, China
Correspondence should be addressed to Qiang Wang; [email protected]
Received 20 January 2013; Accepted 26 March 2013
Academic Editor: Yuping Wang
Copyright Β© 2013 H.-L. Liu and Q. Wang. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.
For orthogonal frequency division multiplexing (OFDM), resource scheduling plays an important role. In resource scheduling,power allocation and subcarrier allocation are not independent. So the conventional two-step method is not very good for OFDMresource allocation. This paper proposes a new method for OFDM resource allocation. This method combines evolutionaryalgorithm (EA) with Karush-Kuhn-Tucker conditions (KKT conditions). In the optimizing process, a set of subcarrier allocationprograms are made as a population of evolutionary algorithm. For each subcarrier allocation program, a power allocation programis calculated through KKT conditions.Then, the system rate of each subcarrier allocation program can be calculated.The fitness ofeach individual is its system rate.The information of optimizing subcarrier and power allocation can be interacted with each other.So, it can overcome the shortcoming of the two-step method. Computer experiments show the proposed algorithm is effective.
1. Introduction
Orthogonal frequency division multiplexing (OFDM) hasattracted more and more research interest because it is apotential solution for high-rate data service demands [1].In OFDM system, the high speed data flow is dividedinto a lot of slow speed data flows. So, the inters symbolinterference (ISI) can be reduced greatly.Meanwhile differentsubcarrierβs channel fading condition is independent, so, theOFDM resource allocation has high flexibility. The flexibilityof multiuser access can also be improved. The principle ofOFDM is described in Figure 1.
The resource allocation problems in OFDM system havetwo different types of analytical perspectives: one is tomaximize the system capacity (RA) when the emission poweris limited; the other is to minimize the transmit power (ME)under the condition that the system capacity is limited.However, in the fourth-generation mobile communicationsystem, the main business is the users experience. Therefore,the former perspective causes more attention.
Resource allocation algorithms can be classified into twocategories: (a) the model can be solved directly. However,it may get an exact result; the computational complexity
is too high. (b) The solving process is divided into severalsteps through some executive strategies.Though it cannot getsuch an exact solution, it is simple and easy to handle. Afteroptimization, the progress shows the proposed solution is ofacceptable accuracy. The two-step method is described as aneffective algorithm. In a two-stepmethod, we first consideratesubcarrier allocation then the power allocation.This methodmay result in low complexity and fast solution speed.
The recent research on the two-step method focuseson the rules of subcarrier allocation with the purpose totake capacity of system and fairness into account. But mostattention is paid to find a good algorithm to optimize thepower allocation so as to maximize the system capacity.For the optimization of the power distribution, the mainmethods are classical deterministic optimization algorithmand intelligent optimization algorithms.
At present, most algorithms for OFDM resource allo-cation focus on how to improve solution efficiency basedon system fairness. Algorithms proposed in [2β5] werebased on the gradient information. In early research, theaffusion algorithm attracted the attention of most scholars.It can maximize the system but make the system fairnessworse. Making use of the gradient information, article [2]
2 Mathematical Problems in Engineering
defined a kind of system throughout increase. For each powerallocation, enhanced power was allocated to the user whohad the maximum throughout. By introducing the constraintconditions according to the definition of throughout increase,the system can enhance throughout and meet the fairness. Inpaper [3], a framework of algorithm based on gradient infor-mation was presented. The model was expressed as a graphoptimization problem, and it resolves the problem with dualdecomposition. Considering that OFDM resource allocationmodel is a nonlinear optimization problem, the literature in[5β10] uses Lagrange method to measure resource allocation.Based on Lagrange method, the literature in [11] introduceda self-adaptive iterative algorithm in order to meet thesystem fairness. In paper [8], the model was converted intoproblem in graph theory and was solved with graph theorycombined with Lagrange method, which gave a satisfactoryresult. In paper [9], the model was converted into a convexoptimization problem, which could be efficiently solved byLagrange method. Also, many algorithms were put forwardby introducing some specific strategies to deal with OFDMresource allocation problem. In these algorithms, classicalor intelligent optimization algorithms did not appear, aspresented in the literature in [12β20]. By introducing theconcept of pareto efficiency, paper [18] presented the trade-offprogram between operators and usersβ service. The literaturein [21β27] solved OFDM resource allocation problem withintelligent algorithm which takes the advantages of simpleoperation and convenient calculation. By introducing weightfactors, paper [25] combined transmit power and systemcapacity to a cost function. A modified PSO algorithm waspresented to solve the problem and got good performance.Paper [26] presented amultiobjective optimizationmodel fordetermining the transmit power and system capacity, respec-tively, and used NSGA-2 algorithm to solve the problem.Due to the mutual restrictive relationship between transmitpower and system capacity, a variable was optimized whilethe other performed worse. Therefore, it is feasible to usemultiobjective algorithm for optimization.
Above, most of the literature is based on two-stepmethod. Firstly, subcarrier is allocated, and then, the modelis optimized. The two-step method can significantly reducethe complexity of the algorithm. But because the subcarrierallocation and power allocation are dependent, it is unrea-sonable to allocate subcarrier and power separately. So, theefficiency of the algorithm is not very high. However, ifsubcarrier allocation and power allocation are optimized atthe same time, the complexity of the algorithm is so high thatit does not meet the instantaneous requirements for mobilecommunication system. Therefore, it is expected to improvethe accuracy and efficiency of algorithm if the information ofsubcarrier allocation and power allocation is able to interactwith each other in each iteration.
A hybrid evolutionary algorithm for OFDM resourceallocation is proposed in this paper. In this algorithm,the subcarrier and power allocation is optimized by turns.Because the OFDM resource allocation is a hybrid opti-mization problem, the discrete and continuous variables canbe optimized by evolutionary algorithm and KKT condi-tions, respectively. This algorithm, firstly, generates a group
b(1)
b(2)
b(3)
Subc
arrie
rs,
bit,
and
pow
eral
loca
tion
IFFT
/cyc
licpr
efix
add
Channel estimations
H(1)
H(2)
User 1terminal
User 2terminal
terminalb(π)
H(π) User π
Figure 1: System model of the downlink MU-OFDM.
subcarrier allocation program, and then, the correspondingpower allocation program is generated by KKT conditions.The system rate of each subcarriers can be obtained. For theevolutionary algorithm, the set of subcarrier is regarded as apopulation.The system rate is the fitness of each individual. Inthe optimizing process, the subcarrier and power allocationinformation can be used with each other. So this algorithm ismore efficient.
The remainder of the paper is organized as follows:the model of OFDM resource allocation and algorithm isdescribed at Sections 2 and 3, computer simulation results areshown at Section 4, and conclusions are drawn at Section 5.
2. The Model of OFDM Resource Allocation
Considering that spectrum resources are limited, OFDMresource allocation plays an important role in the 4th gen-eration mobile communication system. Each eNodeB has alot of orthogonal subcarriers allocated to each user. The coreproblem of OFDM resource allocation is which subcarriersshould be allocated to a user and howmuch each subcarrierβspower should be.
Assume that there are π subcarriers shared by πΎ users;define the channel gain of the πth user on the πth subcarrierππ,π; the noise power is π
2= (π0π΅)/π, where π0 is noise
power spectral density. The signal-to-noise ratio (SNR) isβπ,π = π
2
π,π/π2. The capacity of user π on subcarrier π is
normalized by ππ,π = (1/π) ln(1 + (ππ,πβπ,π/Ξ)), where Ξ isa constant. So, data rate of user π can be written as
π π =
π
β
π=1
π€π,π
πln(1 +
ππ,πβπ,π
Ξ) . (1)
Themathematic model of resource allocation problem can bedescribed as follows:
maxπΉ (π, π) =
πΎ
β
π=1
π
β
π=1
π€π,π
πln(1 +
ππ,πβπ,π
Ξ) (2)
Mathematical Problems in Engineering 3
s.tπΎ
β
π=1
π
β
π=1
ππ,π β€ ππ
πΎ
β
π=1
π€π,π β€ 1 βπ
ππ,π β₯ 0, π€π,π β₯ 0, βπ, βπ,
(3)
where π€π,π β {0, 1}. It means the πth subcarrier is allocatedto the πth user when π€π,π is 1. If π€π,π is 0, it means the πthsubcarrier is not allocated to the πth user. ππ is the totaltransmit power. π = (ππ,π)πΎΓπ
, π = (π€π,π)πΎΓπ.
3. Karush-Kuhn-Tucker Theorem
KKT conditions are proposed by Karush and Kuhn andTucker independently. KKT conditions are first-order neces-sary conditions to solve optimization problem with equalityand inequality constraints.
Consider the following maximization problem
max π (π₯)
s.t ππ (π₯) β₯ 0, π = 1, 2, . . . , π
βπ (π₯) = 0, π = 1, 2, . . . , π
(4)
with π : π π
β π , ππ : π π
β π π, βπ : π
πβ π π being
continuously differentiable functions.Kuhn-Tucker Theorem. If π₯
β is a (local) optimum of theproblem
max π (π₯) (5)
s.t ππ (π₯) β₯ 0, π = 1, 2, . . . , π (6)
βπ (π₯) = 0, π = 1, 2, . . . , π (7)
βπ (π₯β) + βπ (π₯
β) πβ+ ββ (π₯
β) πβ= 0,
β (π₯β) = 0, π (π₯
β) β₯ 0,
πββ₯ 0, π
βπ (π₯β) = 0.
(8)
Equation (10) means βπ is the linear combination of βπ
and ββ, where πβ and π
β are called Lagrange multiplier.Equation (11) means the optimal point must meet all theconstraints including equality and inequality constraints.That is to say the optimal point should be a feasible solution.
4. Making Use of KKT Conditions
For OFDM resource allocation problem, after the subcarrierassignment is completed, the above model becomes
max πΉ (π, π) =
πΎ
β
π=1
π
β
π=1
π€π,π
πln(1 +
ππ,πβπ,π
Ξ) (9)
s.tπΎ
β
π=1
π
β
π=1
ππ,π β€ ππ
ππ,π β₯ 0, π = 1, 2, . . . , πΎ, π = 1, 2, . . . , π,
(10)
where π, π€π,π are known quantity. Equation (15) is anonlinear equation, and it is continuous and differentiable.When ππ,π > 0, ππ,π = βπ,π/Ξ, a stagnation point of theequation can be obtained by KKT conditions:
ππΏ
πππ,π
=ππ,π
1 + ππ,πππ,π
β π β€ 0, (11)
ππ,π (ππ,π
1 + ππ,πππ,π
β π)=0, π=1, 2, . . . , πΎ, π=1, 2, . . . , π
πππΏ
ππ= 0
(12)
Because ππ,π β₯ 0, from (10), we can get
ππ,π
1 + ππ,πππ,π
β π = 0,
πΎ
β
π=1
π
β
π=1
ππ,π β ππ = 0
(13)
by (12). ππ,π can be calculated by
ππ,π (π) =ππ + β
πΎ
π=1(π€π,π/π)β
π
π=1ππ,π
πΎ Γ πβ
1
ππ,π
. (14)
5. Hybrid Evolutionary Algorithm forOFDM Resource Allocation
Goldberg proposed the evolutionary algorithmβs commonlyform.The evolutionary algorithm (EA) simulates the naturalselection in biological evolution. It is essentially a randomsearch algorithm. EAβs global search ability is excellent, andEA dose not demand the objective function continuous ordifferentiable. In (2), the variable π€π,π is discrete, so EAcan be used to solve subcarrier allocation. Meanwhile, KKTconditions are used to solve the variable ππ,π by (14). Sincethe allocation of subcarrier and power is not independent, itis rational to combine these two algorithms to solve resourceallocation problem.
ForOFDMresource allocation, the two-stepmethod can-not get a good solution because the allocation of subcarrierand power is not independent. In this paper, we propose
4 Mathematical Problems in Engineering
Subcarrier
allocation
Power
allocation
Figure 2: The information of optimizing subcarrier and powerallocation can interact with each other.
a new resource allocation program based on KKT condi-tions. This new program combines evolutionary algorithmand KKT conditions. Firstly, we randomly generate somesubcarrier allocation program π1,π2, . . . ,ππ. For each ππ,there is πΉ(ππ, π), (π = 1, 2, . . . , π), and each πΉ(ππ, π)
has only one variable π. πΉ(ππ, π) can be solved by (14).Assume the solution is ππ; for each ππ, we can get a ππ,(π = 1, 2, . . . , π). When ππ and ππ are known, we can getπΉπ. πΉπ is the system rate corresponding to ππ. As previouslydescribed, πΉπ is not a good result. Then, the proposedalgorithm optimizes ππ, (π = 1, 2, . . . , π) using evolutionaryalgorithm. Assume that π1(π‘),π2(π‘), . . . ,ππ(π‘) is π‘ gener-ation population of evolutionary algorithm; through (14),theπ1(π‘), π2(π‘), . . . , ππ(π‘) can be calculated. Select individualsfromπ1(π‘),π2(π‘), . . . ,ππ(π‘) to cross over and mutate.Then,new individuals are generated. Each new individualβs powerallocation can be calculated trough (14). According to everynew individualβs power allocation, select next generationindividuals π1(π‘ + 1),π2(π‘ + 1), . . . ,ππ(π‘ + 1). So, theinformation of power and subcarrier allocation canmake fulluse of each other. It is described in Figure 2.
5.1. Fitness Function. A hybrid evolutionary algorithm forOFDM resource allocation is proposed in this paper. Thisalgorithm firstly generates a set of subcarrier allocation pro-grams π1,π2, . . . ,ππ, where ππ is πth subcarrier allocationprogram, π = 1, 2, . . . , π. For each ππ, its correspondingpower allocation ππ can be obtained through (14), then, thefitness function of ππ = (π€
π
π,π)πΎΓπ can be calculated as
follows:
πΉ (ππ, π) =
πΎ
β
π=1
π
β
π=1
π€π
π,π
πln(1 +
ππ,πβπ,π
Ξ) . (15)
5.2. Encoding of Solutions. Evolutionary algorithm hasattracted more and more attentions because of its goodperformance in the engineering field. For evolutionary algo-rithm, encoding is one of the most important parts in EA.An excellent encoding method cannot only avoid generatingillegal solutions but also improve the performance of thealgorithm.
The subcarrier allocation program π is a πΎ Γ π matrixas follows:
π = (
(
1 0 0 . . . 0 1 0
0 1 1 . . . 0 0 1
0 0 0 . . . 1 0 0
0 0 0 . . . 0 0 0
. . . . . . . . .
0 0 0 . . . 0 0 0
)
)
. (16)
It is clear that there is only one 1 in each column, whichmeanseach subcarrier can only be allocated to one user.WhenπΎ andπ are very large, the computational process expends a lot oftime. What is more, the process of crossover and mutationwould be very complex. So, it is necessary to encode π forsimple operation. In this paper, π is encoded as follows:
π = (π€1, π€2, . . . , π€π) , (17)
where π€π indicates the πth subcarrier allocated to π€πth user.For example, ifπ€π is 5, it means the πth subcarrier is allocatedto the 5th user.
π is also a πΎ Γ π matrix as follows:
π = (
(
π1,1 π1,2 π1,3 . . . π1,π
π2,1 π2,2 π2,3 . . . π2,π
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
ππΎ,1 ππΎ,2 ππΎ,3 . . . ππΎ,π
)
)
, (18)
where ππ,π is the power of the πth subcarrier allocated to πthuser. When π is known, π can be simplified into
π = (π1, π2, . . . , ππ) , (19)
where ππ is the power of πth subcarrier. For example, if ππ =0.5, it means the πth subcarrierβs power is 0.5 w.
5.3. The Proposed Algorithmβs Framework. The hybrid evolu-tionary algorithmproposed in this paper firstly generates a setof subcarrier allocation programsπ1,π2,π3, . . . ,ππ. Then,we can obtain the power allocation program correspondingto each subcarrier allocation program from (14). Substitutingππ and ππ in (14), we can get each individualβs fitness πΉ(ππ, ππ)
of ππ. These subcarrier allocation programs are regarded asa population, and each subcarrier allocation program is anindividual. The fitness of each individual is the system rateobtained by (15). After crossover and mutation, we can geta new population constituted by new subcarrier allocationprograms. A new power allocation program and system ratecan be calculated. Repeat the above steps; the informationoptimizing power and the subcarrier allocation can interactwith each other. Therefore, the performance of the algorithmis improved greatly. According to the above description,the framework of hybrid evolutionary algorithm for OFDMresource allocation is as follows.The Proposed Algorithm
Step 1. Generate a set of subcarrier allocation programs{π1,π2, . . . ,ππ} as the initial population, denoted as pop.
Mathematical Problems in Engineering 5
2 4 6 8 10 122
4
6
8
10
12
14
16
Users
Max
imum
bit
rate
(bits
\s\H
z)
Our algorithmMPSOPFA
Figure 3: System throughput versus number of users.
Step 2. Calculate the fitness of each individual according to(15).
Step 3. Select individuals from pool to crossover, and gener-ate offspring population, denoted as child.
Step 4. Select individuals from child to mutate in mutationprobability and update child.
Step 5. Renew pop with child.
Step 6. If the stopping criteria are unsatisfied, go to Step 2;otherwise, stop.
6. Computer Simulation
We compare the proposed algorithm with [10, 25] in theaspects of the transmit power and system capacity in casethe number of users changes from 2 to 12. In the numericalsimulations, the channel is modeled as slow-varying Rayleighand its components have independent identically distributedcomplex values with zero-mean and unit variance. Totaltransmit power, BER, and the total bandwidth are set as 0.1 w,10β3, and 1MHZ, respectively. The number of subcarrier is
64, and the crossover probability andmutation probability are0.8 and 0.01.
Figure 3 shows the performance of the maximum datarate with growth of the number of users. Figure 4 also showsthat the system total rate calculated by the three algorithms isalmost the same when the number of users is less. However,as the number of users increases, the performance of thealgorithm presented by this paper is better than that of theother two algorithms. Then, we test the proposed algorithmβsminimum transmit power and compare it with MPSO [25]and PFA [10]. As it is shown in Figure 4, when the number ofusers is less than 4, the proposed algorithmβs performance isnot better than that of the other two algorithms. However,
4
5
6
7
8
9
10
11
12
Min
imum
tran
smit
pow
er (d
Bm)
2 4 6 8 10 12Users
Our algorithmMPSOPFA
Figure 4: Minimum transmit power versus number of users.
2 4 6 8 10 121
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
Number of users
Min
imum
use
r cap
acity
(bits
/s/H
z)
Our algorithmMPSOPFA
Figure 5: minimum user capacity for multiuser OFDM versusnumber of users for proposed algorithm and other methods.
since the proposed algorithm combines evolutionary algo-rithm with KKT conditions, it becomes better than MPSOand PFA with the users number increasing. It is clear that theproposed algorithm has superior performance.
Figure 5 shows the minimum user capacity of the pro-posed algorithm, MPSO, and PFA with growth of the totalusers number. If the minimum user capacity is increased, thetotal system rate can also increase. From Figure 5, MPSOβsminimum user capacity is much larger than that of PFA, andthe proposed algorithmβs minimum user capacity is muchlarger than MPSO.
6 Mathematical Problems in Engineering
5 10 15 20 250
1
2
3
4
5
6
SNR (dB)
Tota
l cap
acity
(bits
/s/H
z)
Our algorithm 4 usersMPSO 4 usersPFA 4 users
Figure 6: Capacity comparison of our algorithm and other twoalgorithms as the distributions of 4 users are different.
10 15 20 25 30 35 40 45 502.5
3
3.5
4
4.5
5
5.5
6
Distance (km)
Max
imum
bit
rate
(bits
/s/H
z)
Our algorithm 4 usersMPSO 4 usersPFA 4 users
Figure 7: Capacity comparison of our algorithm and another twoalgorithms with different distance.
Then, this paper compares the capacity of the proposedalgorithm,MPSO, and PFA under different SNRwith 4 users.The comparison result is shown in Figure 6. From Figure 6, itis clear that the proposed algorithmβs performance is betterthan the other two algorithms. When SNR is less than 15 dB,the performance of the proposed algorithm improves slowly,although it is also better than the other two algorithms. Withthe increasing of SNR, the proposed algorithmβs performanceenhances more and more fast. Figure 8 shows the minimumtransmit power of the proposed algorithm, MPSO, andPFA under different SNR with 8 users. From Figure 8, the
5 10 15 20 254.5
5
5.5
6
6.5
7
7.5
SNR (dB)
Min
imum
tran
smit
pow
er (d
Bm)
Our algorithm 8 usersMPSO 8 usersPFA 8 users
Figure 8: Minimum transmit power of our algorithm and anothertwo algorithms as the distributions of 8 users are different.
Table 1
Run index Proposed algorithm MPSO PFA1 1.407 3.012 1.112
2 1.507 3.105 1.203
3 1.426 3.125 1.235
4 1.620 3.106 1.135
5 1.526 2.988 1.125
6 1.458 3.241 1.109
7 1.635 3.075 1.246
8 1.425 3.005 1.106
9 1.326 3.276 1.011
10 1.427 3.006 1.189
Table 2: System rate variance for 100 experiments.
Algorithm 2 users 4 users 6 usersProposed algorithm 0.00000 0.00060 0.00150
MPSO 0.00000 0.00220 0.00760
PFA 0.00000 0.00120 0.00550
proposed algorithmβs minimum transmit power is less thanthat of the other two algorithms.
Figure 7 shows the capacity comparison of the proposedalgorithm and the other two algorithms with different dis-tance, and the number of user is 4. From Figure 7, thetotal capacity is reducing with the distance becoming longerand longer. And the proposed algorithm shows a betterperformance against the increase of distance.
InTable 1, we calculate the execution time to reach the tar-get capacity needed by the proposed algorithm, MPSO, andPFA, respectively. The target capacity is preset to 14 bit/s/Hzwith 12 users. It is clear that the proposed algorithm can reachthe target capacity more fast. It is more efficient to run the
Mathematical Problems in Engineering 7
Table 3: System rate variance for 100 experiments.
Algorithm 8 users 10 users 12 usersProposed algorithm 0.00230 0.00700 0.00960
MPSO 0.00940 0.01200 0.01820
PFA 0.00770 0.00850 0.00920
Table 4: Minimize power variance for 100 experiments.
Algorithm 2 users 4 users 6 usersProposed algorithm 0.00000 0.00000 0.00080
MPSO 0.00000 0.00060 0.00240
PFA 0.00000 0.00000 0.00120
Table 5: Minimize power variance for 100 experiments.
Algorithm 8 users 10 users 12 usersProposed algorithm 0.00085 0.00090 0.00130
MPSO 0.00270 0.00320 0.00390
PFA 0.00160 0.00200 0.00340
proposed algorithm than to run other two algorithms. It ismainly because the proposed algorithm optimizes subcarrierand power allocation at the same time, and in this algorithm,the information of subcarrier and power allocation can inter-act with each other. So, the proposed algorithmβs convergenceis better.
Tables 2 and 4 are system rate variance for 100 exper-iments and minimize power variance for 100 experiments,respectively, when the users number changes from 2 to 6.Tables 3 and 5 are system rate variance for 100 experimentsand minimizing power variance for 100 experiments respec-tively, when the users number changes from 8 to 12. Theyreflect the the stability of algorithm.
These four tables show the stability of the proposedalgorithm is similar to PFA and is better than MPSO in thecase of 2 users and 4 users. In the case of more than 6 users,the stability of our proposed algorithm is better than that ofMPSO and PFA.
7. Conclusions
Since the subcarrier and power allocation are dependent, itis reasonable to optimize them at the same time. However,because the subcarrier allocation is a discrete problemand thepower allocation is a continuous problem, it is very difficultto optimize subcarrier and power allocation at the sametime. This paper proposes a hybrid evolutionary algorithmfor OFDM resource allocation. This algorithm can optimizesubcarrier allocation and transmit power alternately. Theprocess optimizing power allocation can take advantage ofthe information of subcarrier allocation, and the processoptimizing subcarrier allocation can also take advantage ofthe information of power allocation. Therefore, the informa-tion optimizing power and subcarrier allocation can interactwith each other, and the efficiency of the algorithm can be
improved significantly. The computer simulation shows thatthe proposed algorithm increases the system rate apparently.
Acknowledgments
This work was supported in part by the Natural Sci-ence Foundation of China (60974077), the Natural Sci-ence Foundation of Guangdong Province (S2011030002886,S2012010008813), Programme of Science and Technologyof Guangdong Province (2012B091100033), Programme ofScience and Technology of the Department of Educationof Guangdong Province (2012KJCX0042), and ZhongshanProgramme of science and technology (20114A223).
References
[1] S. Sadr, A. Anpalagan, and K. Raahemifar, βRadio resourceallocation algorithms for the downlink of multiuser OFDMcommunication systems,β IEEE Communications Surveys andTutorials, vol. 11, no. 3, pp. 92β106, 2009.
[2] H. Seo and B. G. Lee, βA proportional-fair power allocationscheme for fair and efficient multiuser OFDM systems,β inProceedings of the IEEE Global Telecommunications Conference(GLOBECOM β04), pp. 3737β3741, Dallas, Tex, USA, December2004.
[3] J. Huang, V. Subramanian, R. Agrawal, and R. Berry, βJointscheduling and resource allocation in uplinkOFDMsystems forbroadband wireless access networks,β IEEE Journal on SelectedAreas in Communications, vol. 27, no. 2, pp. 226β234, 2009.
[4] D. Wang, Z. Li, and X. Wang, βJoint optimal subcarrier andpower allocation for wireless cooperative networks over OFDMfading channels,β IEEE Transactions on Vehicular Technology,vol. 61, no. 1, pp. 249β257, 2012.
[5] J. Huang, V. Subramanian, R. Agrawal, and R. Berry, βJointscheduling and resource allocation in uplinkOFDMsystems forbroadband wireless access networks,β IEEE Journal on SelectedAreas in Communications, vol. 27, no. 2, pp. 226β234, 2009.
[6] A. Maiga, J. Y. Baudais, and J. F. Helard, βSubcarrier, bit andtime slot allocation for multicast precoded OFDM systems,β inProceedings of the IEEE International Conference on Communi-cations (ICC β10), pp. 1β6, May 2010.
[7] X. Wang and G. B. Giannakis, βResource allocation for wirelessmultiuser OFDM networks,β IEEE Transactions on InformationTheory, vol. 57, no. 7, pp. 4359β4372, 2011.
[8] A. N. Zaki and A. O. Fapojuwo, βOptimal and efficient graph-based resource allocation algorithms for multiservice frame-based OFDMA networks,β IEEE Transactions on Mobile Com-puting, vol. 10, no. 8, pp. 1175β1186, 2011.
[9] W. W. L. Ho and Y. C. Liang, βOptimal resource allocation formultiuser MIMO-OFDM systems with user rate constraints,βIEEE Transactions on Vehicular Technology, vol. 58, no. 3, pp.1190β1203, 2009.
[10] T. D. Nguyen and Y. Han, βA proportional fairness algorithmwith QoS provision in downlink OFDMA systems,β IEEECommunications Letters, vol. 10, no. 11, pp. 760β762, 2006.
[11] A. S. Mahmoud, A. Y. Al-Rayyah, and T. R. Sheltami, βAdaptivepower allocation algorithm to support absolute proportionalrates constraint for scalable OFDM systems,β in Proceedings ofthe 71st IEEE Vehicular Technology Conference (VTC β10), pp. 1β4, Taipei, Taiwan, May 2010.
8 Mathematical Problems in Engineering
[12] M. S. Maw and I. Sasase, βEfficient resource allocation formultiuser MIMO-OFDM Uplink system to guarantee theproportional data rate fairness among users in a system,β inProceedings of the 1st International Symposium on Access Spaces(ISAS β11), pp. 132β137, Yokohama, Japan, 2011.
[13] Y. Li, N. Lu, M. Peng, and W. Wang, βMultiuser resourceallocation for OFDM downlink with terminal bandwidth limi-tation,β in Proceedings of the IEEEWireless Communications andNetworking Conference (WCNC β10), pp. 2378β2390, April 2010.
[14] P. D. Morris and C. R. N. Athaudage, βFairness based resourceallocation for multi-user MIMO-OFDM systems,β in Proceed-ings of the 63rd IEEE Vehicular Technology Conference (VTCβ06), pp. 314β318, July 2006.
[15] A. Fahmi, M. Asvial, and D. Gunawan, βUplink resourceallocation algorithms with fractional power control as powerconstraints for OFDMA system,β in Proceedings of the IEEERegion 10 Conference (TENCON β11), pp. 592β596, 2011.
[16] H.-G. Chen, L.-P. Lu, and M.-Y. Zhai, βA new OFDM resourceallocation algorithm based on the bit exchange scheme,β inProceedings of the IEEE International Conference on OxideMaterials for Electronic Engineering, pp. 837β840, 2012.
[17] Z. Han, T. Himsoon, W. P. Siriwongpairat, and K. J. R.Liu, βResource allocation for multiuser cooperative OFDMnetworks: who helps whom and how to cooperate,β IEEETransactions on Vehicular Technology, vol. 58, no. 5, pp. 2378β2391, 2009.
[18] K. D. Lee and T. S. P. Yum, βOn pareto-efficiency between profitand utility in OFDM resource allocation,β IEEE Transactions onCommunications, vol. 58, no. 11, pp. 3277β3285, 2010.
[19] M. Li, X. Wang, H. Zhang, and M. Tang, βResource allocationwith subcarrier cooperation in OFDM-based wireless multicastsystem,β in Proceedings of the 73rd IEEE Vehicular TechnologyConference (VTC β11), pp. 1β5, Budapest, Hungary, 2011.
[20] X. Zhao, L. Zheng, M. Huang, and J. Wang, βProportionalresource allocation for multi-user OFDM systems,β in Proceed-ings of the IEEE International Conference on InformationTheoryand Information Security (ICITIS β10), pp. 1100β1103, December2010.
[21] D. Perea-Vega, J. F. Frigon, and A. Girard, βNear-optimaland efficient heuristic algorithms for resource allocation inMISO-OFDM systems,β in Proceedings of the IEEE InternationalConference on Communications (ICC β10), pp. 1β6, Cape Town,South Africa, May 2010.
[22] Q. Wang, H. Liu, and J. Li, βResource allocation for OFDMsystem based on evolutionary algorithm,β in Proceedings ofthe 8th International Conference on Wireless ComputationalIntelligence and Security (CIS β12), pp. 77β81, 2012.
[23] N. Ksairi, P. Bianchi, P. Ciblat, and W. Hachem, βResourceallocation for downlink cellular OFDMA systemsβpart II:practical algorithms and optimal reuse factor,β IEEE Transac-tions on Signal Processing, vol. 58, no. 2, pp. 735β749, 2010.
[24] N. A. Odhah, M. I. Dessouky, W. E. Al-Hanafy, and F. E. A.El-Samie, βGreedy power allocation algorithm for proportionalresource allocation in multi-user OFDM systems,β in Proceed-ings of the 29th National Radio Science Conference (NRSC β12),pp. 421β428, Cairo, Egypt, 2012.
[25] S. Sadeque, I. Ahmed, and R. Vaughan, βImpact of individualand joint optimizations in multi-user OFDM resource alloca-tion by modified PSO,β in Proceedings of the 24th CanadianConference on Electrical andComputer Engineering (CCECE β11),pp. 001233β001237, Niagara Falls, Canada, 2011.
[26] R. Annauth and H. C. S. Rughooputh, βEvolutionary multi-objective approach for resource allocation in OFDM systems,βin Proceedings of the 4th International Joint Conference onComputational Sciences and Optimization, pp. 195β199, 2011.
[27] I. Ahmed and S. P. Majumder, βAdaptive resource allocationbased on modified genetic algorithm and particle swarmoptimization for multiuser OFDM systems,β in Proceedings ofthe 5th International Conference on Electrical and ComputerEngineering (ICECE β08), pp. 211β216, December 2008.
Submit your manuscripts athttp://www.hindawi.com
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
MathematicsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Mathematical Problems in Engineering
Hindawi Publishing Corporationhttp://www.hindawi.com
Differential EquationsInternational Journal of
Volume 2014
Applied MathematicsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Probability and StatisticsHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
OptimizationJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
CombinatoricsHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
International Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Operations ResearchAdvances in
Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Function Spaces
Abstract and Applied AnalysisHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
International Journal of Mathematics and Mathematical Sciences
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
The Scientific World JournalHindawi Publishing Corporation http://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Algebra
Discrete Dynamics in Nature and Society
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Decision SciencesAdvances in
Discrete MathematicsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com
Volume 2014 Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Stochastic AnalysisInternational Journal of