Embodiment
For the ease of illustrating spectrum management method of the present invention based on the double-deck cognitive network architecture of intelligent grid with beingThe principle of system and its remarkable result brought, will be said first against based on the double-deck cognitive network architecture of intelligent grid belowIt is bright.
As shown in Figure 1.Because the peculiar communication service isomerism of electric power is strong and Qos requires that otherness is big, how with bilayerFrequency spectrum resource under the network architecture meets resource utilization and service transmission quality to give full play to the time-varying characteristics of communication serviceRequirement, one of the problem of as power telecom network urgent need to resolve.Cognitive radio technology is applied to intelligent grid Double-level ReticulatedUnder network structure, the reliability of power network various types of communication business will be effectively improved, the development of electricity automation is matched somebody with somebody to intelligent grid and is pushed awayIt is wide significant.
Spectrum management method applied to the double-deck cognition network of intelligent grid is with system, it is necessary to reference to the reality of double-layer networkThe characteristics of framework and wireless sense network, while the characteristics of intelligent grid communication service should be fully taken into account, subject matter have withUnder several aspects：1st, the double-deck cognition network of intelligent grid is the development trend of intelligent grid, but existing all kinds of cognition wirelessThe characteristics of electric spectrum technology can not completely adapt to and utilize its two-level architecture, does not only exist in actually building and fails to make full use ofThe problems such as frequency spectrum resource, longer frequency spectrum distribution time delay, also bring higher construction cost.2nd, first under double-layer network frameworkThere is intersecting area between layer Mesh neighborhood nets, from frequency spectrum angle analysis, same frequency range exists between the different nodes of intersecting areaInterference, the generation of co-channel interference phenomenon should be substantially reduced in spectrum management；From node angle analysis, intersecting area it is anyNode can access multiple Mesh neighborhoods nets, and when a certain Mesh network frequency spectrum resource is nervous, it is idle that the node can access neighborhoodMesh network, by such mode will improve whole system the availability of frequency spectrum and handling capacity.However, single Mesh netsFrequency spectrum distribution mechanism between network can not be considered in terms of interference and frequency spectrum share between net.3rd, intelligent grid communication service hasThere is the characteristics of its is obvious, such as sudden strong, isomerism is strong and Qos requires that otherness is big, in order to ensure important power information energyEnough transmitting to electric power relevant departments promptly and accurately, have higher to the transmission reliability and propagation delay time of double-deck cognition networkIt is required that.Accordingly, it would be desirable to a kind of efficient spectrum management method of research and design, on the premise of can not interfere with each other ensureing frequency spectrum,It is quick to carry out frequency spectrum distribution.It can be seen that, cognitive radio technology is applied under intelligent grid double-layer network structure, will be effectively improvedThe reliability of power network various types of communication business, it is significant to development of the intelligent grid with electricity automation and popularization.
Based on this, the present invention proposes a kind of spectrum management method based on the double-deck cognitive network architecture of intelligent grid with beingSystem.This method will use interference mitigation type graph coloring-simulated annealing blending algorithm, and combine the double-deck cognition network frame of intelligent gridFrequency spectrum resource in structure, intelligent grid carries out efficiently quickly management distribution.In actual applications, the present invention is based on intelligent gridThe spectrum management method of double-deck cognitive network architecture mainly can complete corresponding spectrum management, distribution by spectrum management server and graspMake, spectrum management server is host-guest architecture (Master-Slave), corresponding to the double-deck sensing network of intelligent grid, its is specificFramework is as shown in Figure 2.
In first layer, using each Mesh neighborhoods net as a frequency spectrum allocation unit, multiple spectrum unit compositions are double-deckThe first layer of network structure.There is Slave (a from) spectrum management service at the aggregation node of each Mesh neighborhoods netDevice, the Dynamic Spectrum Management server utilizes interference mitigation type figure by according to the idle usable spectrum inside itself Mesh networkThe graph coloring part of coloring-simulated annealing blending algorithm net the frequency spectrum distribution of interior terminal room.Slave spectrum management serversEvery sub-distribution there will be multiple feasible solutions, the first layer network, which calculates obtained frequency spectrum distribution feasible solution, will uniformly be sent to theMaster (master) spectrum management server of two layers of LTE (Long Term Evolution, Long Term Evolution) base station.MasterDynamic Spectrum Management server will Dynamic Maintenance whole network usable spectrum matrix and unit between interference matrix, usable spectrumMatrix is done to record the same frequency existed between the idle usable spectrum in each unit, unit between interference matrix record adjacent cellsDisturb situation.Eventually through moving back for Master dynamic spectrum server by utilizing interference mitigation types graph coloring-simulated annealing blending algorithmFiery part merges calculating to the feasible solution of each Mesh network spectrum unit, while being disturbed between preventing Mesh from netting, will makeTerminal node rationally accesses idle neighborhood net, and describes by object function the availability of frequency spectrum, handling capacity and time delay comprehensive networkNetwork performance, draws optimal spectrum allocation strategy and feeds back Slave spectrum service devices, finally notify terminal node.Power informationTerminal adjusts inherent parameters, uses distributed frequency range.
Spectrum management method of the present invention based on the double-deck cognitive network architecture of intelligent grid relies on double-deck cognition network frameStructure realizes the dynamic frequency spectrum deployment mechanism and algorithm of interference mitigation type graph coloring-simulated annealing fusion, compared to single figureStain types algorithm, will play the ability that simulated annealing finds global optimum, is more easy to obtain the frequency spectrum distribution of global optimum, comparesIn single simulated annealing type algorithm, graph coloring Algorithm for Solving Mesh network feasible solution is used in advance, will greatly reduce simulationThe input data scale of annealing, improves annealing efficiency and frequency spectrum distribution speed, reduces network delay.It is proposed by the present invention to be based on intelligenceThe spectrum management method of the double-deck cognitive network architecture of energy power network can be quickly obtained optimal spectrum allocation strategy, so that more preferablyUtilization intelligent grid in limited frequency spectrum resource, and optimization handling capacity and time delay while the availability of frequency spectrum is improved, mostEventually power-assisted is provided for the popularization construction of intelligent grid.
As shown in figure 3, a kind of spectrum management method based on the double-deck cognitive network architecture of intelligent grid, including step：
S100：Receive the Spectrum Conversion request that power information terminal is sent in intelligent grid.
Power information terminal sends Spectrum Conversion request in intelligent grid, receives the request.Specifically, in practical applicationIn be that Slave spectrum managements server receives request, i.e., multiple Slave spectrum managements server distribution formulas are arranged at intelligent gridIn, each Slave spectrum managements server is certain amount power information terminal service, specifically may refer to Fig. 2.
S200：It is double that intelligent grid is carried out using the graph coloring part of interference mitigation type graph coloring-simulated annealing blending algorithmThe frequency spectrum distribution of power information terminal room in layer cognitive network architecture.
The Spectrum Conversion request that response of step S100 is received, is calculated using interference mitigation type graph coloring-simulated annealing fusionThe graph coloring part of method carries out the frequency spectrum distribution of power information terminal room in the double-deck cognitive network architecture of intelligent grid.It is specific nextSay, may refer to Fig. 4, Slave spectrum managements server is schemed using interference mitigation type graph coloring-simulated annealing blending algorithmColor method obtains the local feasible solution of frequency spectrum distribution.
S300：Real time spectrum information is obtained, spectral interference matrix between usable spectrum matrix and unit is set up.
Master Dynamic Spectrum Management server obtains real time spectrum information, sets up frequency spectrum between usable spectrum matrix, unitInterference matrix.Specifically, usable spectrum matrix L={ l(m,n)|l(m,n)∈{0,1}}(M×N), when frequency spectrum n is in frequency spectrum allocation unitL when m is available(m,n)=1, otherwise l(m,n)Interference matrix C between=0 frequency spectrum allocation unit(m,j)={ c(i,k)|c(i,k)∈{0,1}}(D×D), when power information terminal i uses certain simultaneously with power information terminal k in frequency spectrum allocation unit j in frequency spectrum allocation unit mOne frequency range and c when interfering(i,k)=1, otherwise c(i,k)=0.
S400：Set up local solution list.
Master Dynamic Spectrum Management server receives local feasible solution, sets up local solution list.
Shown in Fig. 4, in one of the embodiments, set up local solution list and specifically include：
Step one：Obtain local feasible solution and collect.
Step 2：According to the local feasible solution collected, the local set of feasible solution Q of frequency spectrum is set up, wherein,D(i,j)Represent the ID, B of the power information terminal j in spectrum unit i(i,j)RepresentThe frequency range j distributed in spectrum unit i of correspondence power information Termination ID.
S500：According to spectral interference matrix between the local solution list of foundation and the unit, pass through the interference mitigationThe simulated annealing part of type graph coloring-simulated annealing blending algorithm, calculates optimal frequency spectrum distribution.
Spectral interference square between local solution list and the unit of the Master Dynamic Spectrum Management server according to foundationBattle array, by the simulated annealing part of the interference mitigation type graph coloring-simulated annealing blending algorithm, calculates optimal frequency spectrum distribution.
As shown in figure 4, in wherein one is implemented, according to spectral interference between the local solution list of foundation and the unitMatrix, by the simulated annealing part of the interference mitigation type graph coloring-simulated annealing blending algorithm, calculates optimal frequency spectrum distributionThe step of specifically include：
Step one：According to the local solution list of foundation, set of feasible solution is obtained, and choose any one from the set of feasible solutionIntersecting power information Termination ID is covered between individual net.
It is assumed that currently selected power information terminal A, its Termination ID is 1.
Step 2：Usable spectrum matrix is searched, selects the frequency range of dormant network to be divided for the power information terminal of selectionMatch somebody with somebody, obtain frequency allocation result.
Currently available spectral matrix is searched, the frequency range for the power information terminal A selection spatial networks of selection is allocated,Obtain frequency allocation result.
Step 3：Judge that frequency allocation result is disturbed between whether there is unit by spectral interference matrix between the unit,If there is interference, the lookup usable spectrum matrix is returned, the frequency range of dormant network is selected for the power information terminal of selectionIt is allocated, the step of obtaining frequency allocation result, if in the absence of interference, record electricity information terminal ID and corresponding describedFrequency allocation result.
Spectral interference matrix judges that frequency allocation result is done between whether there is unit between the unit obtained according to step S300Disturb, when interference is present, return to step two re-starts frequency allocation；When in the absence of interference, judge that this believes for electric powerThe frequency spectrum distribution for ceasing terminal A is reasonable, records power information terminal A ID and corresponding frequency allocation result, and may be used alsoDistributed with being designated the ID in request list.
Step 4：Detect that whether power information terminal is fully allocated in the set of feasible solution, if not being fully allocated, selectsThe ID of intersecting power information terminal is covered between next net in the set of feasible solution, the lookup usable spectrum matrix is returned, isThe frequency range of the power information terminal selection dormant network of selection is allocated, the step of obtaining frequency allocation result, if completeDistribution, then record epicycle frequency spectrum allocation result, draw global feasible solution.
For power information terminal in set of feasible solution using above-mentioned steps one and the frequency spectrum assigning process of step 2, and work asAfter single power information terminal spectrum is assigned, whether power information terminal is fully allocated in detection set of feasible solution, when not havingThe ID of intersecting power information terminal, and the processing of return to step two are covered when distributing, between next net in set of feasible solutionProcess, epicycle frequency spectrum allocation result when being fully allocated, then recorded, draw global feasible solution.
Step 5：Bring global feasible solution into object function, consider the availability of frequency spectrum, time delay and handling capacity becauseElement, is compared with last round of frequency spectrum allocation result.
Global feasible solution is substituted into object function F (w), the availability of frequency spectrum, time delay, handling capacity factor is considered, it is and upperOne wheel results contrast, with Probability p record compared with the figure of merit and its correspondence feasible solution, wherein Probability p is heuristic factor.The following institutes of F (w)Show, U represents the availability of frequency spectrum, T represents time delay, and Z represents handling capacity.
Step 6：Judge whether to reach annealing minimum temperature, if not up to minimum temperature, cool by annealing rule, afterThe continuous lookup usable spectrum matrix, selects the frequency range of dormant network to be allocated, obtains frequency for the power information terminal of selectionThe step of section allocation result, if having reached minimum temperature, using current solution is as optimal solution and exports.
Judge whether to reach annealing minimum temperature TminIf not up to minimum temperature, by annealing rule cooling T, continues to returnStep 2 is returned, if having reached minimum temperature, will currently solve as optimal solution and export.Minimum temperature of annealing TminIt is interference mitigationLimit value at a temperature of having been set in type graph coloring-simulated annealing blending algorithm.Cooling T is also interference mitigation in annealing ruleDefault parameter in type graph coloring-simulated annealing blending algorithm.
S600：Feed back optimal frequency spectrum allocation result power information terminal into the intelligent grid.
By each Slave spectrum services device distribute frequency spectrum allocation result, power information terminal received spectrum allocation strategy, andItself radio-frequency communication module is adjusted in time, carries out power information transmission.
Spectrum management method of the present invention based on the double-deck cognitive network architecture of intelligent grid, received spectrum conversion requests, profitCarried out with the graph coloring part of interference mitigation type graph coloring-simulated annealing blending algorithm in the double-deck cognitive network architecture of intelligent gridThe frequency spectrum distribution of power information terminal room, obtains real time spectrum information, sets up spectral interference square between usable spectrum matrix and unitBattle array, sets up local solution list, according to spectral interference matrix between the local solution list of foundation and the unit, passes through the interferenceThe simulated annealing part of abatement type graph coloring-simulated annealing blending algorithm, calculates optimal frequency spectrum distribution, feeds back optimal frequency spectrum distributionAs a result to power information terminal.In whole process, divide by spectrum management of interference mitigation type graph coloring-simulated annealing blending algorithmWith core, the optimal spectrum allocation strategy of intelligent grid double-layer network can be quickly obtained, and makes full use of the frequency in networkSpectrum resource.
In one of the embodiments, the optimal frequency spectrum allocation result of feedback power information into the intelligent grid is wholeAlso include after the step of end：
According to the optimal frequency spectrum allocation result of feedback, spectrum management is carried out to power information terminal in the intelligent grid.
As shown in figure 5, a kind of spectrum management system based on the double-deck cognitive network architecture of intelligent grid, including：
Receiving module 100, for receiving the Spectrum Conversion request that power information terminal is sent in intelligent grid.
Frequency spectrum distribute module 200, for the graph coloring part using interference mitigation type graph coloring-simulated annealing blending algorithmCarry out the frequency spectrum distribution of power information terminal room in the double-deck cognitive network architecture of intelligent grid.
Matrix sets up module 300, for obtaining real time spectrum information, sets up spectral interference between usable spectrum matrix and unitMatrix.
Module 400 is set up in solution list, for setting up local solution list.
Optimal frequency spectrum distribute module 500, for spectral interference square between the local solution list according to foundation and the unitBattle array, by the simulated annealing part of the interference mitigation type graph coloring-simulated annealing blending algorithm, calculates optimal frequency spectrum distribution.
Feedback module 600, for feeding back optimal frequency spectrum allocation result power information terminal into the intelligent grid.
Spectrum management system of the present invention based on the double-deck cognitive network architecture of intelligent grid, the received spectrum of receiving module 100Conversion requests, frequency spectrum distribute module 200 is carried out using the graph coloring part of interference mitigation type graph coloring-simulated annealing blending algorithmThe frequency spectrum distribution of power information terminal room in intelligent grid bilayer cognitive network architecture, matrix sets up module 300 and obtains real-time frequencySpectrum information, sets up spectral interference matrix between usable spectrum matrix and unit, and solution list sets up module 400 and sets up local solution list,Spectral interference matrix between local solution list and the unit of the optimal frequency spectrum distribute module 500 according to foundation, by described dryThe simulated annealing part of abatement type graph coloring-simulated annealing blending algorithm is disturbed, optimal frequency spectrum distribution is calculated, feedback module 600 is anti-Optimal frequency spectrum allocation result is presented to power information terminal.In whole process, calculated with interference mitigation type graph coloring-simulated annealing fusionMethod is that spectrum management distributes core, can be quickly obtained the optimal spectrum allocation strategy of intelligent grid double-layer network, and fullyUtilize the frequency spectrum resource in network.
In one of the embodiments, the matrix is set up module 300 and included：
Spectral matrix sets up module, for obtaining real time spectrum use information, sets up usable spectrum matrix L={ l(m,n)|l(m,n)∈{0,1}}(M×N), the l when frequency spectrum n is available in frequency spectrum allocation unit m(m,n)=1, otherwise l(m,n)=0.
Interference matrix sets up module, for setting up interference matrix C between frequency spectrum allocation unit(m,j)={ c(i,k)|c(i,k)∈{0,1}}(D×D), when simultaneously power information terminal i is used with power information terminal k in frequency spectrum allocation unit j in frequency spectrum allocation unit mA certain frequency range and c when interfering(i,k)=1, otherwise c(i,k)=0.
In one of the embodiments, module 400 is set up in the solution list includes：
Collection unit, for obtaining local feasible solution and collecting.
Set of feasible solution sets up unit, for according to the local feasible solution collected, setting up the local set of feasible solution Q of frequency spectrum, wherein,D(i,j)Represent the ID, B of the power information terminal j in spectrum unit i(i,j)Represent the frequency range j distributed in spectrum unit i of correspondence power information Termination ID.
In one of the embodiments, the optimal frequency spectrum distribute module 500 includes：
Terminal chooses unit, for the local solution list according to foundation, obtains set of feasible solution, and from the set of feasible solutionChoose between any one is netted and cover intersecting power information Termination ID.
Frequency spectrum allocation unit, for searching usable spectrum matrix, dormant network is selected for the power information terminal of selectionFrequency range is allocated, and obtains frequency allocation result.
Judging unit is disturbed, for judging frequency allocation result with the presence or absence of single by spectral interference matrix between the unitDisturbed between member, if there is interference, control the frequency spectrum allocation unit to re-execute lookup usable spectrum matrix, for the electricity of selectionThe frequency range of force information terminal selection dormant network is allocated, and obtains the operation of frequency allocation result, if in the absence of interference, rememberingRecord power information Termination ID and the corresponding frequency allocation result.
Detection unit is distributed, for detecting whether power information terminal is fully allocated in the set of feasible solution, if incompleteDistribution, then control the terminal to choose between unit reselects next net in the set of feasible solution and cover intersecting power information endThe ID at end, and control the frequency spectrum allocation unit to re-execute the lookup usable spectrum matrix, it is whole for the power information of selectionThe frequency range of end selection dormant network is allocated, and the operation of frequency allocation result is obtained, if being fully allocated, by epicycle frequency spectrumAllocation result is recorded, and draws global feasible solution.
Comparing unit, for bringing global feasible solution into object function, considers the availability of frequency spectrum, time delay and handles upAmount factor, is compared with last round of frequency spectrum allocation result.
Allocation result identifying unit, for judging whether to reach annealing minimum temperature, if not up to minimum temperature, by moving backFiery rule cooling, controls the frequency spectrum allocation unit to re-execute the lookup usable spectrum matrix, for the power information of selectionThe frequency range of terminal selection dormant network is allocated, and obtains the operation of frequency allocation result, ought if having reached minimum temperaturePreceding solution is as optimal solution and exports.
In one of the embodiments, the spectrum management system based on the double-deck cognitive network architecture of intelligent grid also includes：
Spectrum management module, for the optimal frequency spectrum allocation result according to feedback, to power information in the intelligent gridTerminal carries out spectrum management.
For further explain in detail spectrum management method of the present invention based on the double-deck cognitive network architecture of intelligent grid withThe technical scheme of system, will be illustrated using a concrete application example below.
Assuming that there are 4 cells, that is, 4 frequency spectrum allocation units are formed, respectively descending there are 4 electricity in w, x, y, z, each unitForce information terminal, terminal distribution situation are as shown in fig. 6, make usable spectrum matrix L as follows：
By distribution situation can obtain unit between interference matrix it is as follows, wherein the 2 of No. 1 terminal of w units and x unitsThere is co-channel interference in number terminal or No. 3 terminals；There is co-channel interference in No. 3 of x units or No. 4 terminals and No. 3 terminals of y units；yThere is co-channel interference with No. 1 terminal of z units in No. 4 terminals of unit.Similarly, No. 1 terminal of w units and No. 2 terminals of x unitsAnd No. 3 terminals can share w, x unit frequency spectrum；No. 3 of x units and No. 4 terminals and No. 3 terminals of y units can share x, y unit frequencySpectrum；No. 4 terminals of y units can share y, z unit frequency spectrum with No. 1 terminal of z units.
No. 1 terminal of w units, No. 2 terminals of x units is now made to be sent out with No. 4 terminals, No. 3 terminals of y units and No. 4 terminalsSpectrum Conversion request, z units is played not ask.It can then be calculated by Slave spectrum management servers after local feasible solution, onMaster spectrum service devices are reached, then can obtain local set of feasible solution Q as follows：
Result is compared phase by above-mentioned set of feasible solution Q again by Master dynamic spectrum whois lookup usable spectrum matrix LsClose interference matrix, just obtain one group of global feasible solution TEMP, according to this example, can calculate global feasible solution exist it is following6 kinds：
TEMP1=[(Dw1,Bw1),(Dx2,Bx2),(Dx4, no idle frequency spectrum), (Dy3,Bw1)]；
TEMP2=[(Dw1,Bw1),(Dx2,Bw2),(Dx4, no idle frequency spectrum), (Dy3,Bw3)]；
TEMP3=[(Dw1,Bw1),(Dx2,Bx2),(Dx4, no idle frequency spectrum), (Dy3,Bw4)]；
TEMP4=[(Dw1,Bw1),(Dx2, no idle frequency spectrum), (Dx4,Bx4),(Dy3,By4)]；
TEMP5=[(Dw1,Bw1),(Dx2, no idle frequency spectrum), (Dx4,Bx4),(Dy3,By3)]；
TEMP6=[(Dw1,Bw1),(Dx2, no idle frequency spectrum), (Dx4,Bx4),(Dy3,By1)]；
TEMP7=[(Dw1,f1),(Dx2,Bx2),(Dx4,By1),(Dy3,By3)]。
Above-mentioned feasible solution is taken in the cooling of interference mitigation type graph coloring-simulated annealing blending algorithm in optimal process, respectivelyThe availability of frequency spectrum, handling capacity, Delay Factor are weighed by calculating F (w), optimal solution is selected.Using the 7th group of feasible solution in this example,The frequency spectrum of idle Mesh neighborhoods net is taken full advantage of, object function F (w) is maximum obtained from, as optimal frequency spectrum distributionStrategy.The Master spectrum services device result will be distributed to each Slave spectrum services device, then be forwarded to power information terminal, eventuallyAfter end receives and confirmed, itself radio-frequency module is adjusted, corresponding band is changed to.
Embodiment described above only expresses the several embodiments of the present invention, and it describes more specific and detailed, but simultaneouslyCan not therefore it be construed as limiting the scope of the patent.It should be pointed out that coming for one of ordinary skill in the artSay, without departing from the inventive concept of the premise, various modifications and improvements can be made, these belong to the protection of the present inventionScope.Therefore, the protection domain of patent of the present invention should be determined by the appended claims.