CN106604401A - Resource allocation method in heterogeneous network - Google Patents

Abstract

The invention discloses a resource allocation method in a heterogeneous network. The method comprises the following steps: carrying out subchannel allocation on macrocell user equipment MUEs through a difference algorithm; allocating power for each subchannel through a water-filling algorithm; carrying out grouping on Femtocell through an ant colony algorithm; and carrying out channel and power allocation on femtocell user equipment FUEs through a heuristic algorithm and a distributed power allocation algorithm. The method can effectively suppress cross-layer interference and same-layer interference between a macrocell user equipment layer and a femtocell user equipment layer, so that network spectrum efficiency is improved effectively, and QoS of the MUEs and the FUEs is ensured.

Description

A kind of resource allocation methods in heterogeneous network

Technical field

The present invention relates to wireless communication technology field, the resource allocation methods in more particularly to a kind of heterogeneous network.

Background technology

Mobile communication technology experienced first generation simulation mobile communication technology, the arrowband based on voice of second generation numeral After the wide-band mobile communication technology of mobile communication technology, the third generation for the purpose of high speed internet business and multimedia service, with LTE/LTE-A is that the forth generation broadband access of representative and Distributed Network System have been put into commercialization.However, contrast LTE/LTE-A System and 3G systems, though the carrier frequency utilization rate of LTE/LTE-A systems has been raised, its path loss compares 3G systems Increase.Therefore, LTE/LTE-A systems can not well cover interior.There are some researches show, nearly 90% data service and 60% Speech business occurs indoors, so the in-door covering problem urgent need to resolve of LTE/LTE-A systems.

Femto cell (Femtocell) technology is obtained in recent years as one of most promising technology of indoor wireless communication Extensive research is arrived.Femtocell is exactly the cell that covered of Home eNodeB of short distance, low-power, low cost, Ke Yitong Digital Subscriber Line (Digital Subscriber Line, DSL) or optical fiber and macro-cell communication are crossed, and it is shared with macrocell Band resource, not only can provide the user preferably interior experience, additionally it is possible to unload macrocell network flow, and increase net Network coverage.It is grand compared to traditional however, Femtocell also brings new problem while lift system capacity It is different that cellular network, macro base station (Macro Base Station, MBS) and Femtocell Base Station (FBS) coexist Interference environment in network forming network is more complicated.Not only there is the same layer interference between original MBS in heterogeneous network, add again The same layer interference between cross-layer interference and FBS between MBS and FBS.With the raising of FBS density, these interference can become tighter Weight.

For the problem with present on, presently relevant document has been proposed for some for reducing cross-layer interference and doing with layer The method disturbed, wherein have scholar have studied a kind of interference management scheme and resource allocation policy for double-layer network, is reducing MBS Under the premise of the interference being subject to, and it is excellent to have carried out power distribution to femto user (Femtocell User Equipment, FUE) Change.Also scholar proposes the subchannel under the premise of FUE service quality (Quality of Service, QOS) is ensured Optimize allocation algorithm with power joint, and be provided with interference threshold protecting grand user.But, these methods are not all accounted for Whether in cross-layer interference impact can be produced to the normal transmission of grand user (Macro User Equipment, MUE).Someone carries A kind of Resource Allocation Formula for being intended to minimize the same layer interference between Femtocell is gone out, it is ensured that the QOS of FUE, but has not examined Consider the importance of MUE.There is scholar to carry out federated resource distribution with reference to access control mechanism on the basis of packet, to ensure to use The QoS at family, but have ignored the problem of fairness.

Inventor has found, with the increase of FBS deployment densities, the computation complexity of the Resource Allocation Formula of centralized cooperation Lifted therewith, would become hard to realize in the scene of FBS dense distributions, but, if being effectively grouped to FBS, can be very This problem is solved well.

The content of the invention

For above the deficiencies in the prior art, present invention discusses the resource in Macrocell-Femtocell networks point With problem, a kind of resource allocation algorithm being grouped based on Femtocell is proposed, can effectively suppress macrocell (Macrocell) Cross-layer between client layer and Femtocell client layers is disturbed and disturbed with layer, effectively lifts network spectrum efficiency.

Resource allocation methods in a kind of heterogeneous network of the present invention, comprise the following steps：

Subchannel distribution is performed to grand user MUEs using difference algorithm；

Water-filling algorithm is adopted for each MUEs subchannel distribution through-put power；

Femto cell Femtocell is grouped using ant group algorithm；

Channel distribution is carried out to femto user FUEs using heuritic approach；

Power distribution is carried out to FUEs.

Preferably, the employing difference algorithm performs subchannel distribution and includes to MUEs：

Build first time iteration beneficial matrixWithFor target letter Number distribution MUEs subchannels, wherein,The transmission power for being macro base station in subchannel k, ξ_m.k∈ { 0,1 }, represents subchannel k The state for distributing to grand user；It is Signal to Interference plus Noise Ratio of the MUEm in subchannel k, M is grand total number of users, and K is that subchannel is total Number, if also, K<M, then add M-K virtual MUEs, M × M rank square formations is built, if K>M, then add K-M virtual subchannel, Build K × K rank square formations.

Preferably, the employing ant group algorithm carries out packet to Femtocell includes：

Define the interference collection W={ w of femto base station FBS_ij| i, j ∈ { 1 ..., F } }, w_ijRepresent FBS_iThe user for being authorized Receive FBS_jThe mean value of reference signal power, F is the number of FBS；

Define FBS set S={ s_i| i ∈ { 1 ..., F } }, s_iI-th femto base station for representing, packet space is Groups={ G_x| x=1,2,3 ..., Z }, G_xWhat is represented is specific packet mode, wherein x be with specific ant x as starting point, What wherein Z was represented is the number of packet mode；

Weight coefficient f is set₁And f₂, for adjusting expected heuristic value and information heuristic greedy method；

Calculate interference collection W；

R ant is thrown in, Z kind packet modes are formed, the packet mode for selecting W minimum；

Adjacent difference DELTA W for being grouped interference value twice is calculated, if Δ W is less than a predefined threshold value, by f₁And f₂Adjust Expected heuristic value and information heuristic greedy method simultaneously update global information element；

When a predefined condition is reached, calculate and terminate.

Preferably, the employing heuritic approach carries out channel distribution to FUEs includes：

Assume the power averaging distribution on per sub-channels, and different Femtocell is given by orthogonal subchannel distribution Packet, with FBSs in group identical subchannel can be multiplexed；

On the basis of each FUEs rate requirement is met, power system capacity is maximized,

Wherein, D_jThe set of FUEs, C are serviced by j-th FBS_lL groups FBS distributed by k-th channel, constrain bar Part represents that all of subchannel all can be assigned in a certain group, λ_k,lRepresent whether channel k distributes to l groups,The data-rate requirements of FUEs are represented,Refer to the FUEn's that j-th FBS is serviced Data-rate requirements, K represents subchannel number, and B represents the bandwidth of channel, and L represents the number of packet,Represent that letter is dry to make an uproar Than.

Preferably, the power distribution that carries out to FUEs also includes adopting distributed power distributing method：

Wherein, p_ijK () is j-th femto base station s_jTo the transmission power that femto user j is walked in kth, γ_i(k) table Show that the Signal to Interference plus Noise Ratio SINR that femto user i is walked in kth, β are proportionality coefficients, γ is the SINR threshold values for arranging.

Compared with the prior art, the present invention can effectively reduce interference of the FBSs to MUEs, it is ensured that the normal transmission of MUEs, carry Rise the capacity of whole system.

Description of the drawings

Fig. 1 present invention is for the resource allocation methods preferred embodiment flow chart in Macrocell-Femtocell networks；

It is used for MUEs subchannels in the resource allocation methods that Fig. 2 present invention is used in Macrocell-Femtocell networks The embodiment flow chart of assigning process；

Fig. 3 present invention emulates comparison diagram with prior art outage probability；

Fig. 4 present invention and prior art average throughput comparison diagram；

Fig. 5 present invention emulates comparison diagram with prior art spectrum efficiency；

Fig. 6 present invention emulates comparison diagram with prior art fairness；

Fig. 7 present invention emulates comparison diagram with prior art satisfaction.

Specific embodiment

Must become more apparent to express the object, technical solutions and advantages of the present invention, below in conjunction with the accompanying drawings and specifically Case study on implementation is described in further details to the present invention.

Fig. 1 show the present invention for the resource allocation methods preferred embodiment in Macrocell-Femtocell networks Flow chart, the method is comprised the following steps：

Step 101, using difference algorithm to MUEs perform subchannel distribution；

Step 102, water-filling algorithm is adopted for each MUEs subchannel distribution through-put power；

Step 103, Femtocell is grouped using ant group algorithm；

Step 104, channel distribution is carried out to FUEs using heuritic approach；

Step 105, power distribution is carried out to FUEs.

Fig. 2 is the present invention for believing for MUEs in the resource allocation methods in Macrocell-Femtocell networks The embodiment flow chart of road assigning process, including：

It is assumed that MBS transmission powers on each of the sub-channels are equal.Signal to Interference plus Noise Ratio SINRs of the MUEm in subchannel k It is calculated as follows：

Wherein,WithThe transmission power of respectively MBS and FBSj in subchannel k.δ=1,2 ..., K } represent The set of subchannel, k ∈ δ.What is represented is the set of MUE, WithRespectively grand base Stand and femto base station j it is determined that subchannel k under to MUEm channel gain, φ={ 1,2 ..., F } for FBS value model Enclose, σ²Refer to noise power.

Convolution (1), under conditions of MUEs meets data-rate requirements, plans as follows by subchannel distribution problem：

ξ_m.k∈{0,1} (5)

Wherein, B represents the bandwidth of subchannel, R_mRepresent the data-rate requirements of MUEm, ξ_m,kRepresent the distribution of subchannel k Situation.When subchannel k distributes to MUEm, ξ_m,kIt is otherwise 0 for 1.

Before subchannel distribution is carried out to MUE, one interference threshold is set in the subchannel distributed by MUE, to protect The normal transmission of card MUE.Therefore MUEm needs to meet in given subchannel k：

Represent grand user m can it is determined that subchannel k on can normal transmission a jamming margin.

Known by formula (2), it is exactly that a M appointment between grand user and K sub-channels is asked that subchannel distribution is carried out to MUE Topic.For the assignment problem of MUEs subchannels, inventor employs difference algorithm.As follows is planned to MUEs subchannels problem：

Wherein, M refers to the sum of grand user, and K refers to the sum of subchannel.Formula (7) is the optimization aim letter for needing to solve Number.For convenience of understanding, present invention assumes that transmission power of the macro base station on each sub-channels is equal, can be known by formula (7) Objective matrix required for roadThe algorithm is as shown in Fig. 2 specifically include：

101A：According to it is in need distribution subchannel MUEs channel gain, build first time iteration needed for benefit Matrix

101B：If K<M, i.e. subchannel number are less than MUE numbers, then add M-K virtual subchannel, and objective matrix is become For the square formation of M × M ranks；

101C：If K>When M, i.e. MUE numbers are more than subchannel number, the virtual MUE of addition K-M, by objective matrix It is changed into the square formation of K × K ranks；

101D：After changing into the Assignment Problems of balance, using the difference method of standard, the allocation strategy of subchannel is obtained；

101E：If end condition is accomplished, program is terminated at once；Otherwise, first by the allocated subchannel from treating Remove in the sets of sub-channels of distribution, remove in the MUE set for then distributing subchannel from needs by the MUE for meeting rate requirement again Go, finally build new objective matrix using new sets of sub-channels and MUE set, go to step 101B and start the cycle over；Work as institute The rate requirement for having MUE is satisfied or all subchannels are allocated finishes, and terminates algorithm.

It is each subchannel distribution through-put power in step 102, its concrete methods of realizing is：

After the subchannel distribution of MUEs terminates, the power of mean allocation at the beginning is divided again with water-filling algorithm Match somebody with somebody, so as to the capacity of further lift system.The planning for carrying out power distribution to MUEs is as follows：

Wherein,The gain interference ratio in given subchannel k is referred to, m is MUE, p that finger has determined that when subchannel distribution_kFor the power in subchannel k, P_totFor total transmission power, bar The transmission power that part (10) seeks to meet in all subchannels is not more than total transmission power.

According to above-mentioned condition, power adjustment is carried out by water-filling algorithm.Understand, obtained by lagrange's method of multipliers：

Wherein, ζ is Lagrange multiplier, is a constant.Calculate local derviationEach is just obtained Through-put power in subchannel：

Wherein,η=B/ ζ ln2 are water line.Every height letter can be quickly obtained using said method Through-put power on road, further improves the total handling capacity of system.

Further, it is specific as follows for the FBS grouping process described in Fig. 1 flow chart steps 103：

It is identical with packet problem model of studying in coordination that FBS is grouped problem.It is herein that this point is solved based on ant group algorithm Group problem.Define the interference collection W={ w of FBS_ij| i, j ∈ { 1 ..., F } }, F is the number of FBS.w_ijRepresent femto base station i institutes The user of mandate receives the mean value of femto base station j reference signal powers.In the same manner, w_jiRefer to that femto base station j is authorized The mean value of femto base station i reference signal powers that receives of user.Represented using value larger among both herein Disturbed condition between femto base station i and femto base station j, i.e. w_ij=w_ji=max (w_ij w_ji).Define the set S=of FBS {s_i| i ∈ { 1 ..., F } }, packet space is Groups={ G_x| x=1,2,3 ..., Z }, what wherein Z was represented is the individual of packet mode Number, the initial FBS that each ant selects is different, represents a kind of packet mode G_x, have：

Wherein K is packet number.For packet mode G_xIf, s_jIt has been assigned to groupThenEach FBS During a group can only be all assigned to simultaneously, it is impossible to which FBS can be allocated in two groups, and all of FBS, must be In some group.Therefore, forWithHave：

The packet problem definition of FBS is：

W×S×Groups→G_x (16)

Packet i.e. for FBS is namely based on interference W and FBS set S and selects interference minimum in packet space Groups Packet mode G_x.Problem planning is as follows：

s.t.：

In object function, Q values represent given packet mode G_xThe interference value sum of middle each group.Constraints (18), (19) ensure that F FBS can be allocated and finish with (20), and can only be in a group.Therefore, FBS packets problem is exactly to look for To the minimum of a value of object function Q values, packet mode G now_x, it is exactly group result that we obtain, the frequency spectrum effect after packet Rate can be significantly improved.

Ant group algorithm is utilized herein, W is disturbed as heuristic information using FBS, and have in searching process for ant group algorithm The phenomenon of local optimum may be absorbed in, be that pheromone concentration arranges a maximin, limit a scope, it is to avoid when initial Pheromone concentration be 0 path on always without ant optimizing, so as to the degree of accuracy for reducing being grouped.And two weight coefficient f are set₁ And f₂, dynamic regulation is carried out to expected heuristic value and information heuristic greedy method, to change the important of information content and desired value Degree to pheromones being updated.Algorithm initializes first relevant parameter, calculates the interference W of F FBS.Select most optimal sorting Group process be：According to the characteristic of ant group algorithm, R ant is selected to throw in, R ant selects the FBS of starting from set S, And the FBS of ant next step arrival is selected by state transition probability formula, until F FBS has been traveled through, select each time Starting point FBS is different, represents different packet modes, in the Z kind packet modes for being formed, the packet for selecting interference total value W minimum Mode.Calculate the adjacent interference value that is grouped twice and obtain difference DELTA W, if less than a certain threshold value of setting, then by f₁And f₂To move State adjusts heuristic factor and expecting factor, then global information element is updated.When the number of times of iteration is more than the maximum time for setting Number t_max, stops iteration.Detailed process is as follows：(1) initiation parameter.Initialization relevant parameter t_max, ant number R, information Element initialization matrix MartrixN × N, pheromones volatilization probability ρ etc..

(2) after ant is thrown in, the initial FBS for selecting ant to set out records the FBS volumes that every ant sets out with array p Number.Taboo list (tabu list) is set up, the FBS selected by ant is recorded, it is to avoid repeatedly selected during optimizing Select same FBS.Taboo table record has selected the process of FBS as follows：

1. for the ant thrown in, a s is randomly selected from FBS set S_iAs the initial starting point that ant sets out；

If 2. now in array p without s_iNumbering i, numbering i is stored in p；Make p=[i]；Remember in taboo list The lower s of record_iNumbering i；Tabu=[s_i], represent that this is with s_iFor the packet mode G that starting point is set up_x；

If 3. having s in array p_iNumbering i, choose a femto base station s again from set S_jFor rising that ant sets out Point, then jump to and 2. rejudged.

(3) path is selected according to transition probability formula.F is separately added in state transition probability formula₁And f₂With two power Weight coefficient, when the difference of adjacent two solution that object function is solved, that is, threshold of interference difference DELTA W for solving less than setting When value q, then by changing weight coefficient f₁And f₂Dynamically to adjust heuristic factor α and expecting factor β, to change ant The significance level of residual risk amount and desired value in walking.Transition probability formula is as follows：

(4) wherein,Represent in t ant k by s_iIt is transferred to s_jProbability；allowed_kRepresent ant k next The FBS that step is allowed to select；Martrix (i, j) represents the pheromones value on current path.

(5) ant will select and current s according to formula (21)_iDisturb less s_jIt is mobile.At it to s_jMobile process In, can be to s_jPacket numbering is carried out, the process of FBS distribution group numbers is as follows：

If 1.Then next femto base stationR=1,2 ... K r ≠ y；

2. otherwise

3. until every ant has traveled through all FBS, i.e.,All it is allocated and finishes, it is inevitable at a certain groupIn, now form packet space Groups={ G_x| x=1,2 ..., M }.

(6) local information element is updated.ForAccording to W values, according to formula (22) local message is updated Element：

(7) compareMiddle W is worth size, the packet mode G for selecting W values minimum_xAs the optimal solution of this iteration.

(8) global information element is updated according to formula (23), it is to avoid the pheromones of ant residual are excessively flooded and opened Photos and sending messages：

Matrix (t+1)=(1- ρ) × Matrix (t)+Δ τ (t) (23)

(9) pass through without ant at the beginning due to being likely to occur some paths, cause pheromones to be 0, make this paths one Directly select without ant, such group result degree of accuracy can decrease.For this phenomenon, the maximin of pheromones is set A scope is put, is both avoided being grouped and extremely unbalance situation is occurred, and do not affect it to look for outstanding solution during optimizing, and and Increased the diversity of packet mode.

(10) if Δ W<Q, then adjust f₁'=f₁+Δf₁, f₂'=f₂+Δf₂；Otherwise, f₁、f₂Keep constant.

(11) t=t+1 is made, works as t<During t_max, the increment of pheromones returns 0, and jumping to step (3) carries out Pheromone update； Otherwise, algorithm terminates, and selects the now minimum packet mode G of interference value_bestAs our optimal solution.

Further, channel distribution is carried out to femto user using a kind of heuritic approach in step 104, milli can be being met Realize under the data-rate requirements of picocell user, detailed process is：

It is assumed that the power averaging distribution on per sub-channels.Match somebody with somebody orthogonal subchannel to different component, can with FBSs in group To be multiplexed identical subchannel.On the basis of each FUEs rate requirement is met, power system capacity is maximized, problem is planned such as Under：

Wherein, D_jFor the set of j-th FBS service FUEs.Refer to s_jThe data-rate requirements of the FUEn of service.When λ_k,lWhen equal to 1, represent that subchannel k distributes to a group C_lUse, otherwise λ_k,lFor 0.Formula (25) represents all of subchannel all Can be assigned in a certain group.What formula (26) was represented is the data-rate requirements of femto user.

Channel distribution is carried out to user using a kind of heuristic channel allocation algorithm, algorithm flow is as follows：

1. per group of Mean Speed demand R is calculated_l,|C_l| FBSs in represent l groups Number.

2. per group of subchannel number N for needing distribution is determined_l,

3. for each available subchannels, its SINR at per group is calculated successively.For example subchannel k is in l groups SINR isk∈δ。

4. determine that subchannel k is maximum in the SINR which is organized, if this group is not allocated to enough subchannels by sub- letter Road k distributes to this group.

5. δ and per group of allocated number of subchannels are updated.Repeat step is 3. 4. until distributing whole subchannels.

Wherein, oneself assigned subchannel of group can only be used per the FBS in group, to eliminate in different groups Interference between Femtocell.

Power distribution is carried out to femto user in step 105, including using distributed power allocation algorithm, its concrete reality Now process is：

Power distribution is carried out as the following formula：

Wherein, p_ijK () is s_jTo the transmission power that femto user j is walked in kth.γ_iK () represents that femto user i exists The SINR of kth step.β is proportionality coefficient, span for (0,1].γ is the SINR threshold values for arranging, and the SINR of femto user puts down Weigh on γ or γ, to avoid affecting the QoS of MUE.

The channel gain that concrete parameter used is shown in Table 1, this paper in system emulation mainly considers that path loss, shade decline Fall, wall penetration loss and antenna gain.

The simulation parameter of table 1

Fig. 3 shows that indoors ratio is the outage probability under 10%-100% to MUE.In simulations, the deployment density of FBS It is set as outage threshold for 100% and by -6dB.From analogous diagram, as the ratio of indoor MUE increases, the interruption of MUE Probability is constantly being lifted.This is because channel condition of the interior between MUE and macro base station is poor and the interference of FBS that be subject to is very tight Weight, the channel quality for causing indoor MUE is difficult to ensure that.The Outage probability of distributed antenna of algorithm described herein becomes compared with RRA algorithms It is more and more superior.Because this paper algorithms are by avoiding FBS and neighbouring MUE from using same sub-channel and reduce the modes such as transmission power Significantly reduce interference of the FBS to MUE, it is ensured that the minimum SINR demands of MUE.It can be seen that as MUEs exists Indoor ratio increases, and the MUEs outage probabilities that RRA algorithms are obtained increase to always close 100%, but carried algorithm is obtained MUEs outage probabilities always below 10%.Therefore, this paper algorithms compare other algorithms and can preferably eliminate FBSs to MUEs Interference, it is ensured that the normal transmission of MUEs.

Fig. 4 describes to show average throughputs of the indoor MUE in the case where FBS deployment densities are 10%-100%.Show in figure Show, with the increase of the Femtocell density of deployment, the average throughput of MUE is constantly declining.This is because what MUE was subject to Cross-layer interference is increased as the Femtocell density of deployment becomes big.Max carrier to interference algorithm, improved difference method in figure Obtain in the case where any interference management is not carried out with the handling capacity of RRA algorithms.It is improved for this performance Difference method is more less better than max carrier to interference algorithm, but it considers the fairness between MUE, can meet the normal of more MUE Transmission.This paper algorithms 1 are on the basis of using improved difference algorithm to MUE distribution subchannels, to add again to its power Adjustment.Understand be improved can the performance of system by power adjustment from simulation performance curve.This paper algorithms 2 are to adopt The interference management strategy for being carried herein, significantly reduces interference of the FBSs to MUEs, so as to improve the property of whole system Energy.

Fig. 5 describes FBSs under different deployment densities, the spectrum efficiency change of femtocell.This paper algorithms 1 are to be directed to A kind of grouping algorithm that FBS is proposed.Compared with other several algorithms, algorithm 1 can preferably eliminate the interference between FBS, improve The SINR of FUE, so that spectrum efficiency gets a promotion.Due to the FBSs numbers in each group it is unequal, by group in just After handing over packet, frequency band is caused sufficiently to be utilized, so as to reduce spectrum efficiency.HCFM reduces FFI, but does not account for Cross-layer interference between Macrocell-Femtocell, to ensure the service quality of MUEs.This paper algorithms 2 are in algorithm 1 pair Again the adjustment to its power is added on the basis of FBS packets, further improve the capacity of whole system.

Fig. 6 describes the fairness [16] between FUEs.Show in figure, with the increase of Femtocell deployment densities, Fairness between FUE constantly declines.This is because the increase of the Femtocell density with deployment, the difference of FUE channel qualities Also become increasing.This paper algorithms 1 are that this chapter carries grouping algorithm.As can be seen that the public affairs between the FUE that obtains of this paper algorithms 1 Levelling performance is only than this paper algorithm 2 almost.FBS numbers in each group that other grouping algorithms are got are unbalanced, cause not The interference difference being subject to the FUE in group is larger.Non- grouping algorithm RRA is randomly assigned, and what some FUE may be subject to disturbs It is more serious, cause SINR relatively low.When FBSs dense deployments, packet is easily trapped into local most to Hopfield algorithms Excellent, the packet degree of accuracy is not high.This paper algorithms 2 carry out power adjustment on the basis of algorithm 1 is to FBS packets, further reduce dry Disturb and improve the SINR of FUE.

The satisfaction of FUEs is as shown in fig. 7, contrast other algorithms, only carrying algorithm herein protects can the satisfaction of FUEs Hold in a higher level.It is an iteration searching process to carry grouping algorithm herein, can be adaptive according to FBSs deployment densities Ground is answered to adjust the FBSs numbers in each group, packet performance is improved constantly, and can preferably eliminate interference.This paper algorithms 2 be This paper algorithms 1 on the basis of FBS packets for carrying out power adjustment so that more FUEs can meet rate requirement.

The present invention can effectively suppress the cross-layer between macrocell user layer and Femto cell client layer to disturb and same Layer interference, effectively lifts network spectrum efficiency, it is ensured that the QoS of FUEs and MUEs.

The present invention has been carried out further for embodiment or embodiment to the object, technical solutions and advantages of the present invention Detailed description, should be understood that embodiment provided above or embodiment be only the preferred embodiment of the present invention and , not to limit the present invention, all any modifications made for the present invention within the spirit and principles in the present invention, equivalent are replaced Change, improve, should be included within the scope of the present invention.

Claims (5)

1. resource allocation methods in a kind of heterogeneous network, it is characterised in that comprise the following steps：

Subchannel distribution is performed to grand user MUEs using difference algorithm；

Water-filling algorithm is adopted for each MUEs subchannel distribution through-put power；

Femto cell Femtocell is grouped using ant group algorithm；

Channel distribution is carried out to femto user FUEs using heuritic approach；

Power distribution is carried out to FUEs.

2. method according to claim 1, is further characterized in that, the employing difference algorithm performs subchannel point to MUEs With including：

Build first time iteration beneficial matrixWithFor object function point With MUEs subchannels, wherein,The transmission power for being macro base station in subchannel k, ξ_m.k∈ { 0,1 }, represents dividing for subchannel k The state of the grand user of dispensing；It is Signal to Interference plus Noise Ratio of the MUEm in subchannel k, M is grand total number of users, and K is subchannel sum, If also, K<M, then add M-K virtual MUEs, M × M rank square formations is built, if K>M, then add K-M virtual subchannel, structure Build K × K rank square formations.

3. method according to claim 1, is further characterized in that, the employing ant group algorithm is carried out point to Femtocell Group includes：

Define the interference collection W={ w of femto base station FBS_ij| i, j ∈ { 1 ..., F } }, w_ijRepresent FBS_iThe user for being authorized receives To FBS_jThe mean value of reference signal power, F is the number of FBS；

Define FBS set S={ s_i| i ∈ { 1 ..., F } }, s_iI-th femto base station for representing, packet space is Groups= {G_x| x=1,2,3 ..., Z }, G_xWhat is represented is specific packet mode, and wherein x is wherein Z generations with specific ant x as starting point Table be packet mode number；

Weight coefficient f is set₁And f₂, for adjusting expected heuristic value and information heuristic greedy method；

Calculate interference collection W；

R ant is thrown in, Z kind packet modes are formed, the packet mode for selecting W minimum；

Adjacent difference DELTA W for being grouped interference value twice is calculated, if Δ W is less than a predefined threshold value, by f₁And f₂Adjust and expect Heuristic greedy method and information heuristic greedy method simultaneously update global information element；

When a predefined condition is reached, calculate and terminate.

4. method according to claim 1, is further characterized in that, the employing heuritic approach carries out channel point to FUEs With including：

Assume the power averaging distribution on per sub-channels, and by orthogonal subchannel distribution to different Femtocell point Group, with FBSs in group identical subchannel can be multiplexed；

On the basis of each FUEs rate requirement is met, power system capacity is maximized,

$max\underset{k=1}{\overset{K}{\&Sigma;}}\underset{l=1}{\overset{L}{\&Sigma;}}\underset{j\&Element;C_l}{\&Sigma;}\underset{n\&Element;D_j}{\&Sigma;}{\mathrm{Blog}}_2(1+{\mathrm{\&gamma;}}_{j,k,n}^F){\mathrm{\&lambda;}}_{k,l}$

$\begin{array}{cc}s.t.& \underset{l=1}{\overset{L}{\&Sigma;}}{\mathrm{\&lambda;}}_{k,l}=1,k\&Element;\{1,2,\mathrm{...},K\}\end{array}$

$\underset{k=1}{\overset{K}{\&Sigma;}}{\mathrm{Blog}}_2(1+{\mathrm{\&gamma;}}_{j,k,n}^F){\mathrm{\&lambda;}}_{k,l}\&GreaterEqual;R_{j,n}^F$

Wherein, D_jThe set of FUEs, C are serviced by j-th FBS_lL groups FBS distributed by k-th channel, constraints table Show that all of subchannel all can be assigned in a certain group, λ_k,lRepresent whether channel k distributes to l groups,The data-rate requirements of FUEs are represented,Refer to the FUEn's that j-th FBS is serviced Data-rate requirements, K represents subchannel number, and B represents the bandwidth of channel, and L represents the number of packet,Represent that letter is dry to make an uproar Than.

5. method according to claim 1, is further characterized in that, the power distribution that carries out to FUEs also includes employing point Cloth power distribution method：

$p_{ij}(k+1)=\&lsqb;(1-\mathrm{\&beta;})+\mathrm{\&beta;}\frac{\mathrm{\&gamma;}}{{\mathrm{\&gamma;}}_i\left(k\right)}\&rsqb;p_{ij}\left(k\right)$

Wherein, p_ijK () is j-th femto base station s_jTo the transmission power that femto user j is walked in kth, γ_iK () represents milli The Signal to Interference plus Noise Ratio SINR that slightly user i is walked in kth, β are proportionality coefficients, and γ is the SINR threshold values for arranging.