CN101808325B - Method and device for allocating frequency spectrum - Google Patents

Abstract

The invention discloses a method and a device for allocating frequency spectrum. The method comprises the following steps of: S1, acquiring available frequency spectrum information and constructing a spectrogram model according to network topology; S2, performing spectrogram segmenting on the spectrogram model and acquiring different clusters; S3, managing the clusters obtained by the spectrogram segmenting; and S4, calculating a frequency spectrum efficiency value of nodes of the clusters, and allocating the frequency spectrum to the nodes of the clusters according to the frequency spectrum efficiency value. According to the method and the device for allocating the frequency spectrum, the use ratio of the frequency spectrum resource reaches the maximum, the allocation of the frequency spectrum resource can be balanced, and the justification of the frequency spectrum resource use is improved.

Description

Frequency spectrum distributing method and device

[technical field]

The present invention relates to a kind of frequency spectrum distributing method and device, relate in particular to a kind of frequency spectrum distributing method and device.

[background technology]

Along with society, economic and scientific and technological developing rapidly, the application of radiotechnics is more and more wider, and is especially very urgent to demands for frequency resources in the radio traffic of departments such as city emergency, subway, harbour.In the research and development of these various new technologies of while, the application of new business demands for frequency resources is increased rapidly, radio frequency problem in short supply has become the bottleneck of restriction radio communication development.

Traditional management for radio-frequency spectrum, employing be the mode of fixed allocation frequency spectrum, authorized user removes to use frequency spectrum.But the user who obtains the authorization (authorized user) is not to be round-the-clock this licensed spectrum that takies all, and some frequency ranges are not user's use in part-time, promptly can't realize frequency spectrum share, has caused the significant wastage of frequency spectrum resource.

Arise at the historic moment for solving current demand and trend.Cognitive radio is as the communication technology of new generation in the intelligent and mutual fusion process of wireless network; It is to change the self transmission parameter to satisfy the intelligent communications device of the requirement of communicating by letter according to peripheral communication environment; Make unauthorized user (cognitive user) automatically the perception authorized user use the situation of frequency spectrum; When authorized user does not use frequency spectrum, when promptly frequency spectrum is idle, to the frequency spectrum visit of selecting a good opportunity.

Yet in the spectrum management apparatus of cognitive radio networks, carry out effective spectrum allocation may, and since the appearance of authorized user, the dynamic change of the spectrum availability that the degradation factor can cause under the frequency spectrum quality, thus can cause the unbalanced of frequency spectrum resource distribution.Simultaneously because increasing of cognitive user can cause the spectrum allocation may complexity to increase with exponential.

[summary of the invention]

In view of this, be necessary to distribute the unbalanced and high technological deficiency of spectrum allocation may complexity, a kind of frequency spectrum distributing method based on the spectrogram segmenting technology is provided to frequency spectrum resource.

In addition, also be necessary to provide a kind of spectrum allocation may device based on the spectrogram segmenting technology.

The invention provides a kind of frequency spectrum distributing method, comprise the steps: step S1, obtain usable spectrum information by constructing network topology spectrogram model; Step S2 carries out spectrogram segmenting to said spectrogram model, and obtains different sub-clusterings; Step S3, the management sub-clustering that spectrogram segmenting obtained; Step S4 calculates the frequency spectrum benefit value of the node of said sub-clustering, according to the node of said frequency spectrum benefit value with spectrum allocation may to said sub-clustering; Said step S1 in turn includes the following steps: S11, obtain usable spectrum information and set attribute matrix; S12 sets up the spectrogram model according to said attribute matrix; Said spectrogram model is that (Q), V is a set of node to G=for V, E, and E is the limit collection, and Q is for depositing the usable spectrum set; Step S2 in turn includes the following steps: S21, according to the diagonal matrix of said spectral model structure spectrogram model; S22 makes up Laplce's matrix according to said diagonal matrix; S23 is according to Laplce's matrix computations characteristic vector and obtain the elements corresponding value; S24 is cut apart the spectrogram model according to the element value of said characteristic vector, obtains different sub-clusterings; Step S4 in turn includes the following steps: S41, and the node of said sub-clustering is according to the benefit matrix computations frequency spectrum benefit value in the said attribute matrix; S42, the node of said sub-clustering compares through said frequency spectrum benefit value; S43 if the benefit value that the node of said sub-clustering obtains is maximum, then obtains this frequency spectrum on usable spectrum; S44 is if the benefit value that the node of said sub-clustering obtains on usable spectrum equates that the sub-clustering node that then bandwidth is few obtains this frequency spectrum; S45, said spectrum allocation may finishes or/and said node spectrum requirement is saturated, and then spectrum allocation may stops.

Preferably, step S3 in turn includes the following steps: S31, and bunch head is selected in said sub-clustering; S32, said bunch of first membership table information of preserving said sub-clustering node; S33, said sub-clustering node perceived spectrum information and said spectrum information is sent to bunch head of sub-clustering under the said sub-clustering node; S34, the said bunch of first master controller that regularly sends to the spectrum information that receives the management sub-clustering; S35, said master controller uses information according to the frequency spectrum that spectrum information obtains sub-clustering, and uses bunch head of feedback information to sub-clustering to said frequency spectrum; S36, said node regularly use information from bunch said frequency spectrum of first reception of affiliated sub-clustering, and the new node in the said node uses the Information Selection sub-clustering according to said frequency spectrum.

Preferably, said method also comprises step S5: according to said spectrum allocation may result, the spectrum information between the node is upgraded.

Preferably, step S5 in turn includes the following steps: S51, and the node of said acquisition frequency spectrum is the usable spectrum information deletion that has distributed of said node; S52, the connected node of said acquisition frequency spectrum node is the usable spectrum information deletion that has distributed; S53, the node of the no usable spectrum of deletion; S54, the saturated node of deletion spectrum allocation may; S55 distributes frequency spectrum to give the isolated node of the node that is not connected, and until saturated, and deletes said isolated node.

The present invention also provides a kind of spectrum allocation may device based on spectrogram segmenting; At least comprise successively the spectrogram model module, spectrogram segmenting module, sub-clustering administration module and the spectrum allocation may module that connect: the spectrogram model module, be used for to usable spectrum information and according to said usable spectrum information by constructing network topology spectrogram model; Said spectrogram model is that (Q), V is a set of node to G=for V, E, and E is the limit collection, and Q is for depositing the usable spectrum set; The spectrogram segmenting module is used for said spectrogram model is carried out spectrogram segmenting, and obtains different sub-clusterings; The sub-clustering administration module is used to manage the sub-clustering that spectrogram segmenting obtains; The spectrum allocation may module is used to calculate the node frequency spectrum benefit value of said sub-clustering, and according to the node of said frequency spectrum benefit value with spectrum allocation may to said sub-clustering; Said spectrogram model module also is used to obtain usable spectrum information and sets attribute matrix, and sets up the spectrogram model according to said attribute matrix; Said spectrogram segmenting module is further used for making up diagonal matrix according to said spectral model, makes up Laplce's matrix again, and according to Laplce's matrix computations characteristic vector and acquisition elements corresponding value; And also be used for according to the element value of said characteristic vector the spectrogram model being cut apart, obtain different sub-clusterings; Said spectrum allocation may module be further used for calculating said sub-clustering node the frequency spectrum benefit value and obtain the maximum sub-clustering node of benefit value; And said spectrum allocation may given the maximum sub-clustering node of said benefit value or obtain the sub-clustering node that benefit value equates, the few sub-clustering node of bandwidth of equating to said benefit value said spectrum allocation may.

Preferably; Said sub-clustering administration module comprises master controller; Said sub-clustering administration module is further used for making said sub-clustering to select bunch head; Said bunch of first membership table information of preserving said sub-clustering node, and also be used to make the spectrum information around the said sub-clustering node perceived and said spectrum information is sent to bunch head of sub-clustering under the said sub-clustering node, the said bunch of first master controller that regularly sends to the spectrum information that receives the management sub-clustering; Said master controller is used for obtaining according to bunch first spectrum information that sends that receives the frequency spectrum use information of sub-clustering; Said master controller uses frequency spectrum on bunch head of feedback information to said sub-clustering; New node in the said node uses the Information Selection sub-clustering according to said frequency spectrum.

Preferably, said device also comprises the update module that is connected with the spectrum allocation may module, and said update module is used for according to the spectrum allocation may result spectrum information between the node being upgraded.

Preferably, said update module is further used for deleting the usable spectrum information of the node of having distributed to said acquisition frequency spectrum, distributes to the node and the saturated node of spectrum allocation may of the usable spectrum information of the connected node of said acquisition frequency spectrum node, no usable spectrum; Also be used to distribute frequency spectrum to give the isolated node of the node that is not connected,, and delete said isolated node until saturated.

Beneficial effect of the present invention has:

Frequency spectrum distributing method of the present invention at first makes up the spectrogram model; Adopt the spectrogram segmenting technology that this spectrogram model is cut apart then; Obtain different sub-clusterings and the sub-clustering of being cut apart is managed and spectrum allocation may; Thereby reached the utilance of maximization frequency spectrum resource, and distribution that can the balance frequency spectrum resource, the fairness that frequency spectrum resource utilizes improved;

When spectrum allocation may finishes or/and the node spectrum requirement is saturated, then spectrum allocation may stops; According to the node demand, restrain the computational methods of frequency spectrum fast, the concurrent spectrum allocation may of carrying out makes that the distribution of frequency spectrum resource and load are more balanced, further improves fairness and utilance that frequency spectrum resource utilizes;

The node (cognitive user) that dynamically moves, gets into or leave is selected the maximum sub-clusterings of usable spectrum and added this sub-clustering, reach the uniform distribution of frequency spectrum resource, further improve the fairness that frequency spectrum resource utilizes, further improve the utilance of frequency spectrum resource;

Network topology structure is upgraded, can be carried out the uniform distribution of frequency spectrum resource again, further improve the fairness of the utilization of frequency spectrum resource, further improve the utilance of frequency spectrum resource the node of dynamic running transform.

[description of drawings]

For further understanding characteristic of the present invention and technology contents, see also following relevant accompanying drawing of the present invention, yet institute's accompanying drawing only provides reference and explanation usefulness, is not to be used for the present invention is limited.

Fig. 1 is the flow chart of frequency spectrum distributing method of the present invention;

Fig. 2 is the sketch map of the embodiment of frequency spectrum distributing method of the present invention;

Fig. 3 is another sketch map of the embodiment of frequency spectrum distributing method of the present invention;

Fig. 4 is the block diagram of spectrum allocation may device of the present invention.

[embodiment]

Below in conjunction with accompanying drawing, describe in detail through specific embodiments of the invention, will make technical scheme of the present invention and other beneficial effects obvious.

The present invention is divided into a plurality of different sub-clusterings through spectrogram segmenting to the spectrogram model, manages through each sub-clustering autonomous selection bunch head and to sub-clustering, then the node in the sub-clustering (being cognitive user) is distributed frequency spectrum.

Referring to Fig. 1, frequency spectrum distributing method of the present invention comprises:

At first step S1 makes up the spectrogram model, obtains usable spectrum information by constructing network topology spectrogram model;

Concrete step is following:

S11, to the following attribute matrix of said usable spectrum information setting:

Spectrum availability: A={a _{I, k}| a _{I, k}∈ 0,1}} _{N * M}, it represents the availability of frequency spectrum, wherein

Interference-limited matrix: C={c _{I, j, k}| c _{I, j, k}∈ 0,1}} _{N * N * M}, representation node is interference-limited on same frequency spectrum.Wherein,

Spectrum allocation may matrix: S={s _{I, k}| s _{I, k}∈ 0,1}} _{N * M}Represent an effective spectrum allocation schemes, wherein

Matrix S need satisfy the restrictive condition of interference-limited Matrix C definition in addition, that is to say: s _{I, k}S _{J, k}=0, if c _{I, j, k}=1, $\&ForAll;i,j<N,k<M;$

Neighbours' collection: H={h _{I, k}} _{N * M}Represent neighbours (bunch head and bunch head between be node neighbours or sub-clustering between the be neighbours) number of each node i on frequency spectrum k;

Bandwidth matrices: B={b _{I, k}} _{N * M}The bandwidth that one node i successfully obtains usable spectrum k has been described;

Benefit matrix: R={r _{I, k}} _{N * M}The benefit value that representation node i uses frequency spectrum k to be obtained, i.e. r _{I, k}Be defined as: $r_{i,k}=\mathrm{Max}\frac{b_{i,k}}{(h_{i,k}+1)\&CenterDot;\underset{i=0}{\overset{M}{\mathrm{\&Sigma;}}}{a}_{i,k}\&CenterDot;b_{i,k}},$

Wherein be 0 situation, with the neighbours number h of this node i on the k frequency spectrum for fear of denominator _{I, k}Change h into _{I, k}+ 1;

The attribute matrix that sets based on above-mentioned usable spectrum information is so can set up the spectrogram model, i.e. step S12: set the spectrogram model according to said attribute matrix;

(Q), wherein V is a set of node for V, E, and E is the limit collection, and Q deposits usable spectrum set, v promptly to define a undirected spectrogram model G= _iRepresent cognitive nodes, e _IjRepresent v _iWith v _jBetween the limit;

The adjacency matrix that undirected spectrogram model G is corresponding is W=[w _Ij], w _IjBe limit e _IjWeights.w _IjRefer to v _iTo v _jBetween the number of usable spectrum.

Step S2 carries out spectrogram segmenting to said spectrogram model, and obtains different sub-clusterings;

This spectrogram segmenting techniques make use adjacency matrix and Laplce (Laplacian) matrix are represented a graph model, then Laplce (Laplacian) matrix are asked characteristic value; For the second little eigenvalue in this characteristic value ₂Pairing characteristic vector x ₂Attribute, and utilize this second little characteristic vector x ₂Element value come split spectrum figure.Characteristic vector x ₂Element represented the weights of respective nodes in the spectrogram model, so the difference between the element has reflected the difference between the node;

According to this rule, the spectrogram model is divided into two parts, wherein the second little characteristic vector x ₂In, with the value of the corresponding characteristic element of node ID less than 0, then represent the node of this sequence number will divide in a frequency spectrum subgraph model, and other will divide in another one frequency spectrum subgraph model greater than 0;

Concrete; This spectrogram segmenting technology is the process of a loop iteration; In each iterative process; Be divided into two frequency spectrum subgraph models (be sub-clustering, the pass of understanding the spectrogram model for image ties up to this step and is called frequency spectrum subgraph model) to a spectrogram model (perhaps calling frequency spectrum father graph model); According to the spectrogram model G=among the step S12 (V, E, Q) as input, operation spectrogram segmenting algorithm, the detailed step of this process is following:

S21: the diagonal matrix D=[d that makes up spectrogram model G _Ij], wherein $d_{i,j}=\left\{\begin{array}{cc}\underset{k=1}{\overset{n}{\mathrm{\&Sigma;}}}{w}_{\mathrm{Ik}},& \mathrm{If\; i}=\mathrm{<figure-callout\; id="0"\; label="j"\; filenames="HSA00000012047600012.png"\; state="\{\{state\}\}">j</figure-callout>}\\ 0,& \mathrm{Otherwise}\end{array}\right.;$

S22: make up Laplce (Laplacian) matrix L according to diagonal matrix, make L=D-W, W is the adjacency matrix among the step S1;

S23: ask the above-mentioned second little eigenvalue according to Laplce (Laplacian) matrix ₂Pairing characteristic vector x ₂

S24: according to this second little characteristic vector x ₂Element value the spectrogram model is cut apart.

Simultaneously the spectrogram model is defined internal edges number and external edge number respectively, the internal edges number refers to the fillet number between the node in the frequency spectrum subgraph model (sub-clustering), and the external edge number refers to the fillet number between the frequency spectrum subgraph model (sub-clustering);

Whether the algorithm through internal edges number and this iteration of external edge numerical control system stops; If have at least the internal edges number of a frequency spectrum subgraph model (sub-clustering) to surpass its external edge number; We claim that this frequency spectrum subgraph model (sub-clustering) does not have its frequency spectrum father graph model containing; Then stop to cut apart of this spectrogram model, and the set of the final frequency spectrum subgraph model (sub-clustering) that generates.

Step S3 divides cluster management, the management sub-clustering that spectrogram segmenting obtained; Further step is following:

S31, bunch head is selected in sub-clustering, in all nodes of sub-clustering (cognitive user), the node conduct bunch head that the limit number is maximum;

S32, bunch first membership table information of preserving node, what bunch head preserved is the membership table information of sub-clustering interior nodes under this bunch head;

S33, the sub-clustering node perceived to spectrum information be sent to bunch head of sub-clustering under the node;

S34, bunch first master controller that regularly sends to the spectrum information that receives the management sub-clustering; Bunch first regularly (this time can set up on their own) with the sub-clustering that receives spectrum information be transmitted to master controller;

S35, the spectrum information that master controller receives according to bunch head from each sub-clustering, and analyze the information that the sub-clustering frequency spectrum uses that obtains, this frequency spectrum uses information via to handle bunch head that feeds back to sub-clustering;

S36, node regularly bunch said frequency spectrum of first reception from affiliated sub-clustering of (this time can set up on their own) use information; Node uses the Information Selection sub-clustering according to said frequency spectrum, and this node comprises: move, get into or leave the new node of sub-clustering, select the maximum sub-clustering of frequency spectrum and add this sub-clustering;

Based on node move, get into or leave sub-clustering, node uses the maximum sub-clustering of Information Selection usable spectrum based on frequency spectrum and adds, improve node frequency spectrum utilization rate with improve the fairness that frequency spectrum resource utilizes.

Step S4 carries out the frequency spectrum benefit value to the node of said sub-clustering and calculates, according to said calculated value the node of spectrum allocation may to sub-clustering; Further step is following:

S41; The node of sub-clustering is according to the benefit matrix computations frequency spectrum benefit value in the said attribute matrix, for each usable spectrum of its perception calculates the different benefit value according to benefit matrix in the step 2

;

S42, and to each frequency spectrum in its usable spectrum one by one the node in the sub-clustering carry out spectrum allocation may, node is through compare with the benefit value of its neighbours' frequency spectrum the benefit value of each usable spectrum one by one;

S43, if the benefit value that this node obtains on available usable spectrum repayment is maximum among its neighbours, then this node obtains this frequency spectrum;

S44 when if identical benefit value repayment appears in this node and its neighbours, then obtains the less node of bandwidth and obtains frequency spectrum;

S45 after each node has been accomplished this spectrum allocation may, broadcasts the usable spectrum information after its spectrum allocation may information and the renewal to its neighbours, and prepares to get into next iteration cycle; Until all frequency spectrums all assigned perhaps/all saturated with the spectrum requirement of all node, then spectrum allocation may stops.

Frequency spectrum distributing method based on the spectrogram segmenting technology can minimize the weights that connect between the node; The weights that the maximization intra-node connects; Restrain the computational methods of frequency spectrum fast, the concurrent spectrum allocation may of carrying out, thereby the utilance of maximization frequency spectrum resource; And distribution that can the balance frequency spectrum resource, further improve the fairness that frequency spectrum resource utilizes.

The method of above-mentioned spectrum allocation may based on spectrogram segmenting also comprises step S5: said spectrum allocation may finishes, and the structure between the node (being network topology structure) is upgraded; Concrete update method is following successively:

The node of the said acquisition frequency spectrum of S51 is the usable spectrum information deletion that has distributed of said node;

The connected node of the said acquisition frequency spectrum of S52 node is the usable spectrum information deletion that has distributed;

The node of the no usable spectrum of S53 deletion;

The saturated node of S54 deletion spectrum allocation may;

S55 distributes frequency spectrum to give the isolated node of the node that is not connected, and until saturated, and deletes said isolated node.

This network topology structure is upgraded, can be carried out the uniform distribution of frequency spectrum resource again, further improve the fairness of the utilization of frequency spectrum resource, further improve the utilance of frequency spectrum resource the node of dynamic running transform.

Combine Fig. 2 and Fig. 3 that one embodiment is provided based on above-mentioned institute method, specific as follows:

At first, make up the spectrogram model; Set 5 usable spectrums are arranged at most (CH1, CH2, CH3, CH4, CH5), 10 nodes (SU1, SU2 ... SU10), the topological structure that they are formed is as shown in Figure 2, and the adjacency matrix of this figure can be expressed as:

$W=\left[\begin{array}{cccccccccc}0& 5& 0& 0& 0& 0& 0& 0& 0& 0\\ 5& 0& 3& 1& 1& 0& 0& 2& 0& 0\\ 0& 3& 0& 0& 0& 0& 0& 0& 0& 0\\ 0& 1& 0& 0& 5& 4& 2& 0& 0& 0\\ 0& 1& 0& 5& 0& 0& 0& 0& 0& 0\\ 0& 0& 0& 4& 0& 0& 0& 0& 0& 0\\ 0& 0& 0& 2& 0& 0& 0& 0& 0& 0\\ 0& 2& 0& 0& 0& 0& 0& 0& 3& 0\\ 0& 0& 0& 0& 0& 0& 0& 3& 0& 2\\ 0& 0& 0& 0& 0& 0& 0& 0& 2& 0\end{array}\right]$

The row and column of adjacency matrix is represented cognitive nodes SU1-10, the quantity of available spectrum between the numeric representation node of this adjacency matrix the inside.

Then, it is different sub-clusterings that frequency spectrum is cut apart this spectrogram model, includes:

Make up the diagonal matrix D=[d of spectrogram model G _Ij],

$D=\left[\begin{array}{cccccccccc}5& 0& 0& 0& 0& 0& 0& 0& 0& 0\\ 0& 12& 0& 0& 0& 0& 0& 0& 0& 0\\ 0& 0& 3& 0& 0& 0& 0& 0& 0& 0\\ 0& 0& 0& 12& 0& 0& 0& 0& 0& 0\\ 0& 0& 0& 0& 6& 0& 0& 0& 0& 0\\ 0& 0& 0& 0& 0& 4& 0& 0& 0& 0\\ 0& 0& 0& 0& 0& 0& 2& 0& 0& 0\\ 0& 0& 0& 0& 0& 0& 0& 5& 0& 0\\ 0& 0& 0& 0& 0& 0& 0& 0& 5& 0\\ 0& 0& 0& 0& 0& 0& 0& 0& 0& 2\end{array}\right]$

Make up Laplce's matrix L, make L=D-W

$L=\left[\begin{array}{cccccccccc}5& -5& 0& 0& 0& 0& 0& 0& 0& 0\\ -5& 12& -3& -1& -1& 0& 0& -2& 0& 0\\ 0& -3& -3& 0& 0& 0& 0& 0& 0& 0\\ 0& -1& 0& 12& -5& -4& -2& 0& 0& 0\\ 0& -1& 0& -5& 6& 0& 0& 0& 0& 0\\ 0& 0& 0& -4& 0& 4& 0& 0& 0& 0\\ 0& 0& 0& -2& 0& 0& 2& 0& 0& 0\\ 0& -2& 0& 0& 0& 0& 0& 5& -3& 0\\ 0& 0& 0& 0& 0& 0& 0& -3& 5& -2\\ 0& 0& 0& 0& 0& 0& 0& 0& -2& 2\end{array}\right]$

Laplce (Laplacian) matrix is asked the second little eigenvalue 2 pairing characteristic vector x ₂, x ₂=-0.0029 ,-0.0026 ,-0.0031,0.2982,0.2679,0.3355,0.3834 ,-0.2864 ,-0.4331 ,-0.5569

This characteristic vector x ₂In divide a frequency spectrum subgraph model (sub-clustering) G into less than 0 element corresponding nodes ₁([v ₁... v ₃, v ₈... v ₁₀]), divide another one frequency spectrum subgraph model (sub-clustering) G into greater than 0 element corresponding nodes in addition ₂([v ₄... v ₇]); G ₁The internal edges number be 5, the external edge number is 2; G ₂The internal edges number be 3, the external edge number is 2; Therefore the internal edges number of two sub-graphs is proceeded to divide all greater than the external edge number.

Carry out the loop iteration process of spectrogram segmenting.

Make up frequency spectrum subgraph model G ₁The adjacency matrix of model

${\mathrm{<figure-callout\; id="1"\; label="W"\; filenames="HSA00000012047600012.png"\; state="\{\{state\}\}">W</figure-callout>}}_1=\left[\begin{array}{cccccc}0& 5& 0& 0& 0& 0\\ 5& 0& 3& 2& 0& 0\\ 0& 3& 0& 0& 0& 0\\ 0& 2& 0& 0& 3& 0\\ 0& 0& 0& 3& 0& 2\\ 0& 0& 0& 0& 2& 0\end{array}\right]$

Make up frequency spectrum subgraph model G ₁Diagonal matrix

${\mathrm{<figure-callout\; id="1"\; label="D"\; filenames="HSA00000012047600012.png"\; state="\{\{state\}\}">D</figure-callout>}}_1=\left[\begin{array}{cccccc}5& 0& 0& 0& 0& 0\\ 0& 10& 0& 0& 0& 0\\ 0& 0& 3& 0& 0& 0\\ 0& 0& 0& 5& 0& 0\\ 0& 0& 0& 0& 5& 0\\ 0& 0& 0& 0& 0& 2\end{array}\right]$

Make up subgraph G ₁Laplce (Laplacian) matrix

${\mathrm{<figure-callout\; id="1"\; label="L"\; filenames="HSA00000012047600012.png"\; state="\{\{state\}\}">L</figure-callout>}}_1=\left[\begin{array}{cccccc}5& -5& 0& 0& 0& 0\\ -5& 10& -3& -2& 0& 0\\ 0& -3& 3& 0& 0& 0\\ 0& -2& 0& 5& -3& 0\\ 0& 0& 0& -3& 5& -2\\ 0& 0& 0& 0& -2& 2\end{array}\right]$

Ask the second little eigenvalue ₂Pairing characteristic vector x ₂

x ₂＝-0.3813、-0.3219、-0.4348、0.1214、0.3854、0.6313

With the second little eigenvalue ₂Pairing characteristic vector x ₂Element corresponding nodes less than 0 in the characteristic vector divides a frequency spectrum subgraph model (sub-clustering) G into ₃([v ₁... v ₃]), divide another one frequency spectrum subgraph model (sub-clustering) G into greater than 0 element corresponding nodes in addition ₄([v ₈... v ₁₀]); G ₃The internal edges number be 2, the external edge number is 1; G ₄The internal edges number be 2, the external edge number is 1; Because G ₂, G ₃, G ₄Not containing after cutting apart, thus we can to obtain the final result that this figure cuts apart be G ₂([v ₄... v ₇]), G ₃([v ₁... v ₃]), G ₄([v ₈... v ₁₀]).

Then to a minute cluster management, comprising: a bunch first-selection is selected, and in order better to adapt to this variation, proposes one to converge to the spectrum allocation may algorithm of target fast.In an iterative process of spectrum allocation may, as much as possible usable spectrum is distributed to node, make and use few iterations to accomplish the spectrum allocation may process of one whole as far as possible; Concurrent execution in a plurality of sub-clusterings that this algorithm obtains in the step spectrogram segmenting can greatly be accelerated the convergence rate of spectrum allocation may;

Bunch head of each sub-clustering independently carries out the management to node; When new node or node occurring and moving, the frequency spectrum that the periodic perception master controller of this node sends uses information, selects the maximum sub-clustering of usable spectrum; Add this sub-clustering, and by bunch first sub-clustering information of giving.

At last, spectrum allocation may;

Because the spectrum allocation may algorithm is that to cut apart this spectrogram model at the step frequency spectrum be concurrent execution in all sub-clusterings that obtain after spectrogram segmenting is carried out in different sub-clusterings;

Get final product so enumerate the spectrum allocation may situation of a sub-clustering, below with G ₄([v ₈... v ₁₀]) carry out spectrum allocation may for example.The bandwidth of supposing each frequency spectrum is 0.5M, and the usable spectrum tabulation of SU8 is: (CH5), bandwidth demand is 0.8M for CH1, CH3; The usable spectrum tabulation of SU9 is: (CH1, CH3, CH4, CH5), bandwidth demand is 0.9M; The usable spectrum tabulation of SU10 is: (CH1, CH2, CH3, CH4), bandwidth demand is 1.0M.The frequency spectrum that then distributes to SU8 is: (CH1, CH3), the frequency spectrum that distributes to SU9 is: (CH4 CH5), owing to do not have the frequency spectrum conflict between SU9 and the SU8, therefore can use identical frequency spectrum simultaneously, and the frequency spectrum that then distributes to SU9 is: (CH1, CH2).So far three nodes are all saturated, then distribute to stop.

The device that the present invention is based on the spectrum allocation may of spectrogram segmenting comprises spectrogram model module, spectrogram segmenting module, sub-clustering administration module and the spectrum allocation may module that connects successively at least;

The spectrogram model module, be used for to usable spectrum information by constructing network topology spectrogram model, simultaneously to spectrogram model specification spectral properties matrix, and set up undirected spectrogram model G=(V, E, Q);

The spectrogram segmenting module is used for said spectrogram model is carried out spectrogram segmenting, and obtains different sub-clusterings; Be used for said spectrogram model is carried out spectrogram segmenting, and obtain different sub-clusterings; This spectrogram segmenting module makes up diagonal matrix according to said spectral model, makes up Laplce's matrix again, and according to Laplce's matrix computations characteristic vector and acquisition elements corresponding value; And also be used for according to the element value of said characteristic vector the spectrogram model being cut apart, obtain different sub-clusterings.

The sub-clustering administration module is used to manage the sub-clustering that spectrogram segmenting obtains; The sub-clustering administration module is used to make said sub-clustering to select bunch head; The first membership table information of preserving said sub-clustering node of this bunch; And also be used to make the spectrum information around the said sub-clustering node perceived and spectrum information is sent to bunch head of sub-clustering under the sub-clustering node, bunch first master controller that regularly sends to the spectrum information that receives the management sub-clustering; This master controller is used to receive a bunch first spectrum information that sends, and uses information according to the frequency spectrum of sub-clustering that spectrum information obtains; This master controller uses frequency spectrum feedback information to said bunch of head; New node in the said node uses the Information Selection sub-clustering according to said frequency spectrum.

The spectrum allocation may module is used to calculate the node frequency spectrum benefit value of said sub-clustering, and according to the node of said frequency spectrum benefit value with spectrum allocation may to said sub-clustering; This spectrum allocation may module be used to calculate sub-clustering node the frequency spectrum benefit value and obtain the maximum sub-clustering node of benefit value; And give the maximum sub-clustering node of said benefit value or for obtaining the sub-clustering node that benefit value equates with spectrum allocation may, the few sub-clustering node of bandwidth of equating to said benefit value spectrum allocation may.

The device of above-mentioned spectrum allocation may based on spectrogram segmenting also comprises the update module that is connected with the spectrum allocation may module, and this update module is used for according to the spectrum allocation may result, and the spectrum information between the node (being network topology structure) is upgraded; This update module is further used for deleting the usable spectrum information of having distributed to the node that obtains frequency spectrum, distributes to the node and the saturated node of spectrum allocation may of the usable spectrum information of the connected node of said acquisition frequency spectrum node, no usable spectrum; Also be used to distribute frequency spectrum to give the isolated node of the node that is not connected, until saturated, and deletion does not have this isolated node of usable spectrum.

The above embodiment has only expressed execution mode of the present invention, and it describes comparatively concrete and detailed, but can not therefore be interpreted as the restriction to claim of the present invention.Should be pointed out that for the person of ordinary skill of the art under the prerequisite that does not break away from the present invention's design, can also make some distortion and improvement, these all belong to protection scope of the present invention.Therefore, the protection range of patent of the present invention should be as the criterion with accompanying claims.

Claims (8)

1. a frequency spectrum distributing method comprises the steps:

Step S1 obtains usable spectrum information by constructing network topology spectrogram model;

Step S2 carries out spectrogram segmenting to said spectrogram model, and obtains different sub-clusterings;

Step S3, the management sub-clustering that spectrogram segmenting obtained;

Step S4 calculates the frequency spectrum benefit value of the node of said sub-clustering, according to the node of said frequency spectrum benefit value with spectrum allocation may to said sub-clustering;

Said step S1 in turn includes the following steps:

S11 obtains usable spectrum information and sets attribute matrix;

S12 sets up the spectrogram model according to said attribute matrix;

Said spectrogram model is that (Q), V is a set of node to G=for V, E, and E is the limit collection, and Q is for depositing the usable spectrum set;

Step S2 in turn includes the following steps:

S21 is according to the diagonal matrix of said spectrogram model construction spectrogram model;

S22 makes up Laplce's matrix according to said diagonal matrix;

S23 is according to Laplce's matrix computations characteristic vector and obtain the elements corresponding value;

S24 is cut apart the spectrogram model according to the element value of said characteristic vector, obtains different sub-clusterings;

Step S4 in turn includes the following steps:

S41, the node of said sub-clustering is according to the benefit matrix computations frequency spectrum benefit value in the said attribute matrix;

S42, the node of said sub-clustering compares through said frequency spectrum benefit value;

S43 if the benefit value that the node of said sub-clustering obtains is maximum, then obtains this frequency spectrum on usable spectrum;

S44 is if the benefit value that the node of said sub-clustering obtains on usable spectrum equates that the sub-clustering node that then bandwidth is few obtains this frequency spectrum;

S45, said spectrum allocation may finishes or/and said node spectrum requirement is saturated, and then spectrum allocation may stops.

2. frequency spectrum distributing method according to claim 1 is characterized in that step S3 in turn includes the following steps:

S31, bunch head is selected in said sub-clustering;

S32, said bunch of first membership table information of preserving said sub-clustering node;

S33, said sub-clustering node perceived spectrum information and said spectrum information is sent to bunch head of sub-clustering under the said sub-clustering node;

S34, the said bunch of first master controller that regularly sends to the spectrum information that receives the management sub-clustering;

S35, said master controller uses information according to the frequency spectrum that spectrum information obtains sub-clustering, and uses bunch head of feedback information to sub-clustering to said frequency spectrum;

S36, said node regularly use information from bunch said frequency spectrum of first reception of affiliated sub-clustering, and the new node in the said node uses the Information Selection sub-clustering according to said frequency spectrum.

3. according to each described frequency spectrum distributing method in claim 1 or 2, it is characterized in that said method also comprises step S5:, the spectrum information between the node is upgraded according to said spectrum allocation may result.

4. frequency spectrum distributing method according to claim 3 is characterized in that step S5 in turn includes the following steps:

S51, the node of said acquisition frequency spectrum is the usable spectrum information deletion that has distributed of said node;

S52, the connected node of said acquisition frequency spectrum node is the usable spectrum information deletion that has distributed;

S53, the node of the no usable spectrum of deletion;

S54, the saturated node of deletion spectrum allocation may;

S55 distributes frequency spectrum to give the isolated node of the node that is not connected, and until saturated, and deletes said isolated node.

5. the spectrum allocation may device based on spectrogram segmenting is characterized in that, comprises the spectrogram model module, spectrogram segmenting module, sub-clustering administration module and the spectrum allocation may module that connect successively at least:

The spectrogram model module, be used for to usable spectrum information and according to said usable spectrum information by constructing network topology spectrogram model;

The spectrogram segmenting module is used for said spectrogram model is carried out spectrogram segmenting, and obtains different sub-clusterings;

The sub-clustering administration module is used to manage the sub-clustering that spectrogram segmenting obtains;

The spectrum allocation may module is used to calculate the node frequency spectrum benefit value of said sub-clustering, and according to the node of said frequency spectrum benefit value with spectrum allocation may to said sub-clustering;

Said spectrogram model module also is used to obtain usable spectrum information and sets attribute matrix, and sets up the spectrogram model according to said attribute matrix; Said spectrogram model is that (Q), V is a set of node to G=for V, E, and E is the limit collection, and Q is for depositing the usable spectrum set;

Said spectrogram segmenting module is further used for making up Laplce's matrix again according to said spectrogram model construction diagonal matrix, and according to Laplce's matrix computations characteristic vector and acquisition elements corresponding value; And also be used for according to the element value of said characteristic vector the spectrogram model being cut apart, obtain different sub-clusterings;

Said spectrum allocation may module is further used for calculating the frequency spectrum benefit value of the node of said sub-clustering; Obtain the maximum sub-clustering node of benefit value; And with the sub-clustering node of said spectrum allocation may to said benefit value maximum; Or if the benefit value that the node of said sub-clustering obtains on usable spectrum equates then to obtain the sub-clustering node that benefit value equates, and give said benefit value the equal few sub-clustering node of bandwidth said spectrum allocation may.

6. spectrum allocation may device according to claim 5; It is characterized in that; Said sub-clustering administration module comprises master controller; Said sub-clustering administration module is further used for making said sub-clustering to select bunch head; Said bunch of first membership table information of preserving said sub-clustering node, and also be used to make the spectrum information around the said sub-clustering node perceived and said spectrum information is sent to bunch head of sub-clustering under the said sub-clustering node, the said bunch of first master controller that regularly sends to the spectrum information that receives the management sub-clustering; Said master controller is used for obtaining according to bunch first spectrum information that sends that receives the frequency spectrum use information of sub-clustering; Said master controller uses frequency spectrum on bunch head of feedback information to said sub-clustering; New node in the said node uses the Information Selection sub-clustering according to said frequency spectrum.

7. spectrum allocation may device according to claim 5 is characterized in that said device also comprises the update module that is connected with the spectrum allocation may module, and said update module is used for according to the spectrum allocation may result spectrum information between the node being upgraded.

8. based on the described spectrum allocation may device of claim 7; It is characterized in that said update module is further used for deleting the usable spectrum information of the node of having distributed to said acquisition frequency spectrum, distributes to the node and the saturated node of spectrum allocation may of the usable spectrum information of the connected node of said acquisition frequency spectrum node, no usable spectrum; Also be used to distribute frequency spectrum to give the isolated node of the node that is not connected,, and delete said isolated node until saturated.

Images