CN107613554B - A kind of aware distributed robust method for controlling downlink power of interference - Google Patents
A kind of aware distributed robust method for controlling downlink power of interference Download PDFInfo
- Publication number
- CN107613554B CN107613554B CN201710902375.0A CN201710902375A CN107613554B CN 107613554 B CN107613554 B CN 107613554B CN 201710902375 A CN201710902375 A CN 201710902375A CN 107613554 B CN107613554 B CN 107613554B
- Authority
- CN
- China
- Prior art keywords
- base station
- small base
- interference
- power level
- mean field
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Mobile Radio Communication Systems (AREA)
Abstract
The invention proposes a kind of aware distributed robust method for controlling downlink power of interference, it is intended in the uncertain super-intensive network environment of channel information, realize the robust distributed power control of small base station.Realize step are as follows: small base station calculates the approximate interference mean field of itself, it sets the time interval of power control and obtains disturbance state space, discretization is carried out respectively to it, and relevant iteration step length is calculated, the matrix of the interference mean field of building, Lagrangian and power level is initialized；Derive the more new formula of interference mean field, Lagrangian and power level；Small base station will interfere mean field to be assigned to termination variable, and is utilized respectively three more new formulas and is updated respectively to interference mean field, Lagrangian and power level, judge whether power level restrains, if so, stopping iteration, if it is not, repeating the above steps.The present invention improves the robustness and system energy efficiency of power control under the premise of guaranteeing user communication quality.
Description
Technical field
The invention belongs to wireless communication technology fields, are related to a kind of aware distributed robust downlink power control side of interference
Method.It can be used in the interference management of the network communicating system of the small base station deployment of super-intensive.
Background technique
With the development of wireless communication technique and the explosive growth of mobile application, traditional honeycomb is increasingly
Difficulty meets the growing data traffic requirement of people.Home eNodeB isomery Cellular Networks are disposed with the super-intensive that traditional base station coexists
It can be multiplexed limited frequency resource completely, i.e. Home eNodeB and macro base station uses identical frequency band, can not only promote frequency spectrum
Efficiency, and the distance between user and access point can be shortened, lifting system handling capacity, but it is dry to exist simultaneously serious people having a common goal
It disturbs, internal system interference is caused to have complicated interference structure, this interference is more serious in super-intensive environment, gives network
Interference management bring challenges, this become restrict network performance an important factor for.For this purpose, how effectively to manage dense deployment isomery
Interference in cellular network becomes the key for improving network performance.
Interference management can be divided into three types according to the processing mode to interference problem: interference eliminates, AF panel with
And interference coordination.Interference cancellation techniques are that signal of communication reaches the letter that will be received behind receiving end by demodulated or decoded information
The technology that interference in number is directly eliminated, interference cancellation techniques generally can not to the more demanding of receiving device, user equipment
Meet condition, the communication commonly used in user equipment in uplink to base station.Interference mitigation technology is by beam forming, does
The methods of randomization, precoding is disturbed to share the progress interference signal neutralization of known interference problem or multi-user jointly to reach
Inhibit the purpose of interference.And interference coordination, also known as jam avoidance technology then refer to through the distribution to downlink resource such as frequency
Or the setting of transmission power is coordinated, the technology of representative has the division of the resources such as power control, time domain frequency domain, wherein being based on power control
The main realization principle of the interference coordination schemes of system is: in home base station network, guaranteeing that femtocell user reaches user
Under the premise of the communication quality of demand, the transmission power of Home eNodeB is minimized as far as possible, and each Home eNodeB is reduced with this
Interference coordination is realized in interference to peripheral base station.
It is exactly power control, root that an important means of interference coordination is carried out in dense deployment Home eNodeB isomery Cellular Networks
Distributed and Centralized Power Control can be divided into according to power control mode classification.Centralized Power Control in order to make receive signal
Signal to Interference plus Noise Ratio (SINR) it is identical, gain and base station received signal power according to each link carry out transmitting function to transmitting terminal
The adjustment of rate.Centralized Power Control can uniformly be adjusted base station mobile subscriber, optimize its transmission power, can
Deteriorated caused by system performance with improving feedback.Such as a kind of non intelligent adaptive algorithm, Power Control Problem is configured to
One betting model designs the cost function of game person under the demand for guaranteeing user's communication link SINR, in the system of minimum
The closed solutions of power level are derived under the target of intraformational interference.Centralized Power Control needs explicitly to recognize all chains
Road information, time delay is big, and computation complexity is high, is difficult to implement in systems in practice.
Distributed power control refers to that the SINR to make receiving end receive signal is equal, according to link gain and receives
The power of signal, the method that individually transmission power of itself is adjusted.When designing distributed power control method, usually
Power Control Problem is configured to a betting model, game person formulates certainly according to oneself state and the strategy of other game persons
Body strategy, so that game person needs to carry out with other game persons when formulating itself strategy big when the quantity of game person is very big
The information exchange of amount reduces the efficiency of power control, and existing distributed power control method is caused to be difficult to apply in super-intensive
In network environment.And there are uncertain informations in actual channel information, but existing distributed power control method is false mostly
If channel information is known completely, cause the robustness of power control poor.Such as application publication number is CN 104469857
A, the patent application of entitled " a kind of energy-saving power control method that the home base station network of dense deployment has QoS to ensure " are public
In the home base station network for having opened a kind of dense deployment, there is the distributed energy-saving power control method of time delay QoS guarantee.This method
Realization step are as follows: all subchannels in home base station cells are numbered；Judge whether power distribution reaches Na Shijun
Weighing apparatus or more than the number of iterations the upper limit, if so, terminate iteration, if it is not, each terminal reports it to be subjected on each of the sub-channels
Same layer is interfered to Home eNodeB；Home eNodeB is calculated in the transmission power of a certain sub-channels and the line of the total transmission power in base station
Sexual intercourse calculates Home eNodeB and corresponds to transmission power level in the subchannel；Calculate the transmitting of the every sub-channels of Home eNodeB
Linear relationship between power determines the transmission power of Home eNodeB on each of the sub-channels；If hair is calculated in previous step
It penetrates power and is greater than the maximum transmission power that single subchannel allows, then the transmission power in the subchannel is single channel permission
Maximum transmission power, otherwise, maximum transmission power in the subchannel is to calculate resulting transmission power level, is transferred to step
Two；Iteration is terminated, each Home eNodeB completes radio resource allocation, and waiting is dispatched next time.The disadvantage of this method is that
It is simple to assume that channel information is known completely, cause the robustness of power control poor.
Summary of the invention
It is an object of the invention to overcome the problems of the above-mentioned prior art, a kind of aware distributed Shandong of interference is proposed
Stick method for controlling downlink power, it is intended in the uncertain super-intensive network environment of channel information, realize the robustness point of small base station
Cloth power control.
Technical thought of the invention is: weakening mean field theory of games frame using interference, the control of small base station power is asked
Topic is configured to a betting model, introduces indeterminate in the interference dynamical state equation of game person, is guaranteeing user's communication
Under conditions of link SINR demand, the corresponding cost function with robustness is designed, optimal power control problem is converted
As the smallest optimization problem of cost function is made in the case where maximizing indeterminate, derive that the Hamilton-of system is refined
Comparable-Bellman equation and Fu Ke-Planck-Ke Ermoge love equation, then can be by the mean field game mould of system complex
Type is expressed as the combination of above-mentioned two equation, to quickly derive the equilibrium solution of game, by finite-difference algorithm and
Discrete Lagrange relaxation method solves above-mentioned equation respectively, to obtain a kind of distributed interference perception downlink power control plan
Slightly.
According to above-mentioned technical thought, the technical solution for realizing that the object of the invention is taken includes the following steps:
(1) small base station i calculates the approximate interference mean field of itself:
(1a) in the scene of N number of small base station, all small base station transmitting powers areTest signal a；
The user UE of (1b) small base station i serviceiMeasure the power of received signal and noiseAnd it willIt feeds back to
Small base station i；
(1c) small base station i utilizes the power for testing signal aPowerWith small base station i to user UEiChannel gain
gI, i, calculate the average interference channel gain of small base station iMean field is interfered with approximation
(2) time interval [0, T] of small base station i setting power control, and obtain power control disturbance state space [0,
Imax]:
The time interval [0, T] of (2a) small base station i setting power control, wherein the maximum value of T expression time interval；
(2b) small base station i obtains the disturbance state space [0, I of power controlmax]: its in scene in addition to small base station i
He emits signal, user UE at small base station full poweriGive the Feedback of Power being interfered to small base station i, small base station i is by jamming power
As maximum interference state Imax, constitute the disturbance state space [0, I of power controlmax]；
(3) small base station i is to time interval [0, T] and disturbance state space [0, Imax] discretization is carried out, and calculate time dimension
The iteration step length δ of degreetWith the iteration step length δ of disturbance state dimensionI:
(3a) small base station i is to time interval [0, T] and disturbance state space [0, Imax] discretization is carried out simultaneously, it is wrapped
Discrete grid block containing X period and Y disturbance state section；
(3b) small base station i calculates the iteration step length δ of time dimension by Xt, the iteration of disturbance state dimension is calculated by Y
Step-length δI；
(4) small base station i building interference mean fieldLagrangianAnd power levelAnd it is carried out initial
Change:
(4a) small base station i constructs matrix behavior X+1 in discrete grid block, is classified as the interference mean field of Y+1Lagrange
OperatorAnd power level
Small base station number in (4b) small base station i statistics scene in each disturbance state section accounts for the ratio of small base station total number
Example obtains disturbance state distribution；
(4c) small base station i is by interference by disturbance state distribution to interference mean fieldIt is initialized, is obtained initial
Interference mean field after changeSimultaneously to LagrangianAnd power levelIt is initialized, is initialized respectively
Lagrangian afterwardsWith the power level after initialization
(5) interference mean field is derivedMore new formula m, LagrangianMore new formula λ and power levelIt updates public
Formula p:
Weaken mean field theory of games using interference, designs the cost function with robustness, and derive game playing system
Hamilton Jacobi Bellman equation and Fu Ke-Planck-Ke Ermoge love equation, are asked using different methods
Solution, concrete methods of realizing are as follows:
(i) Fu Ke-Planck-Ke Ermoge love equation is solved using Upwind finite difference method, it is flat obtains interference
Equal fieldMore new formula m；
(ii) discrete Lagrange relaxation method is used, Hamilton Jacobi Bellman equation and Fu Ke-Pu Lang are passed through
Ke-Ke Ermoge love equation, solves LagrangianMore new formula λ and power levelMore new formula p:
By Hamilton Jacobi Bellman equation and Fu Ke-Planck-Ke Ermoge love it is equations turned be corresponding
Optimization problem enables Lagrange's equation to interference using the Lagrange's equation of Lagrangian method building optimization problem
The result of mean field derivation is equal to zero, solves LagrangianMore new formula λ, enable Lagrange's equation to power water
The result of flat derivation is equal to zero, solves power levelMore new formula p；
(6) small base station i will interfere mean fieldIt is assigned to termination variableAnd to interference mean fieldIt is updated
Obtain matrixRealize step are as follows:
(6a) small base station i will interfere mean fieldIt is assigned to termination variableTerminate variable
(6b) small base station i is to interference mean fieldIt is updated:
Small base station i utilizes interference mean fieldMore new formula m, pass through power levelTo interference mean fieldIn member
Element is updated, and obtains updated matrix
(7) small base station i is by matrixIt is assigned to interference mean fieldInterfere mean field
(8) small base station i is to LagrangianUpdate obtains matrixAnd by matrixIt is assigned to LagrangianRealize that steps are as follows:
(8a) small base station i utilizes LagrangianMore new formula λ, pass through power levelTo LagrangianIn element be updated, obtain updated matrix
(8b) small base station i enables Lagrangian
(9) small base station i is by power levelIt is assigned toAnd to power levelIt is updated, obtains matrixIt is real
It is existing that steps are as follows:
(9a) small base station i is by power levelIt is assigned toI.e.
(9b) small base station i is to interference mean fieldIt is updated:
Small base station i utilizes power levelMore new formula p, pass through interference mean fieldAnd LagrangianTo function
Rate is horizontalIn element be updated, obtain updated matrix
(10) small base station i is by matrixIt is assigned to power levelThat is power level
(11) small base station i judges power levelWhether restrain:
Small base station i is calculatedWith termination variableThe absolute value of difference, and whether judge all elements in the value
Less than preset threshold value, if so, power levelConvergence, and by the power levelAs the power level after convergenceIf it is not, repeating step (6)~step (11).
Compared with the prior art, the invention has the following advantages:
1, the present invention weakens mean field theory of games using interference, and external environment simplifies the complicated interference of small base station
To interfere mean field, a small amount of information exchange is only needed since small base station calculates itself interference mean field, is greatly reduced ultra dense
Collect the information exchange in scene between small base station, to improve the efficiency of power control, is especially suitable for super-intensive network scenarios.
2, the present invention passes through the disturbance state dynamical equation in game person when using decrease mean field theory of games is interfered
Middle introducing indeterminate devises the corresponding cost function with robustness, and using the thought of worst case, derives function
The more new formula of rate level improves the robustness of power control to obtain a kind of Poewr control method with robustness.
3, the present invention, only by local information, utilizes interference mean field and Lagrangian when updating power level
Power level is updated, without carrying out power control by central node, so being a kind of distributed power control
Method processed.
Detailed description of the invention
Fig. 1 is schematic diagram of a scenario used in the embodiment of the present invention；
Fig. 2 is implementation flow chart of the invention；
Fig. 3 is the present invention and non intelligent adaptive algorithm system average energy efficiency cumulative distribution function simulation comparison figure.
Specific embodiment
Below in conjunction with the drawings and specific embodiments, the present invention will be described in further detail.It should be appreciated that this place is retouched
The specific embodiment stated is only used to explain the present invention, is not intended to limit the present invention.
Referring to Fig.1, the scene that the present embodiment uses, including 1 macro base station MeNB, 6 Home eNodeB SeNB, multiple families
Base station user SUE.It is the scene that one family base station SeNB services a user SUE by System Model Reduction for convenience of analysis,
And Home eNodeB and user are uniformly distributed in the scene in the scene, choose one of Home eNodeB SeNB here1Carry out function
Rate control, namely SeNB1As small base station i, the wherein frequency spectrum of the Home eNodeB full rate multiplexing macro base station of dense deployment, this
In consider intraformational interference the case where, then for user SUE1, remove small base station SeNB1The signal that other outer small base stations are sent to it all
It is interference signal, wherein pre-set the SINR threshold value of user's communication requirement, wherein SINR > 0.This method is capable of providing
A kind of effective robust method for controlling downlink power for carrying out interference coordination.
Referring to Fig. 2, a kind of aware distributed robust method for controlling downlink power of interference, comprising the following steps:
The small base station i of step 1) calculates the approximate interference mean field of itself:
Step 1a) small base stationAccording to the signaling that small base station i is issued, transmission power isTest signal a；
Step 1b) small base station i service user UEiMeasure the power of received signal and noiseAnd it willFeedback
To small base station i；
Step 1c) small base station i using test signal a powerPowerWith small base station i to user UEiChannel increase
Beneficial gI, i, calculate the average interference channel gain of small base station iMean field is interfered with approximationWherein, calculate small base station i's
Average interference channel gainMean field is interfered with approximationCalculation formula be respectively as follows:
Wherein, N indicates the number of small base station in scene,Indicate the power of test signal a,Indicate user UEiIt receives
Signal and noise power,Indicate small base station i to user UEiChannel gain；
The time interval [0, T] of the small base station i setting power control of step 2), and obtain the disturbance state space of power control
[0, Imax]:
Step 2a) small base station i setting power control time interval [0, T], wherein the maximum value of T expression time interval；
Step 2b) small base station i obtains the disturbance state space [0, I of power controlmax]: in scene in addition to small base station i
Other small base station full powers emit signal, user UEiGive small base station i, small base station i that will interfere function the Feedback of Power being interfered
Rate is as maximum interference state Imax, constitute the disturbance state space [0, I of power controlmax]；
The small base station i of step 3) is to time interval [0, T] and disturbance state space [0, Imax] discretization is carried out, and when calculating
Between iteration step length δtWith the iteration step length δ of disturbance stateI:
Step 3a) small base station i uses finite difference calculus, to time interval [0, T] and disturbance state space [0, Imax] simultaneously
Discretization is carried out, the discrete grid block comprising X period and Y disturbance state section is obtained；
Step 3b) small base station i passes through the iteration step length δ that X calculates time dimensiont, changing for disturbance state dimension is calculated by Y
Ride instead of walk long δI, wherein the iteration step length δ of time dimensiontThe iteration step length δ of disturbance state dimensionICalculation formula be respectively as follows:
The small base station i building interference mean field of step 4)LagrangianAnd power levelAnd it is carried out just
Beginningization:
Step 4a) small base station i constructs matrix behavior X+1 in discrete grid block, it is classified as the interference mean field of Y+1Glug
Bright day operatorAnd power level
Step 4b) the small base station number in small base station i statistics scene in each disturbance state section accounts for small base station total number
Ratio obtains disturbance state distribution；
Step 4c) small base station i by interference by disturbance state distribution to interfering mean fieldIt is initialized, is obtained just
Interference mean field after beginningizationSimultaneously to LagrangianAnd power levelIt is initialized, is obtained initial respectively
Lagrangian after changeWith the power level after initializationRealize step are as follows:
Step 4c1) small base station i enablesAnd disturbance state distribution is assigned to interference mean fieldThe first row in remove
Other elements outside first column element, and enable interference mean fieldIn be equal to zero except the every other element of the first row, obtain
Interference mean field after initialization
Step 4c2) small base station i enables LagrangianIn all elements be equal to zero, the drawing after being initialized
Ge Lang operator
Step 4c3) small base station i enables power levelIn all elements be equal to the minimum emissive power p of small base station imin
With the maximum transmission power p of small base station imaxBetween any value, the power level after being initialized
Step 5) derives interference mean fieldMore new formula m, LagrangianMore new formula λ and power levelMore
New formula p:
Weaken mean field theory of games using interference, designs the cost function with robustness, and derive game playing system
Hamilton Jacobi Bellman equation and Fu Ke-Planck-Ke Ermoge love equation, are asked using different methods
Solution, concrete methods of realizing are as follows:
(i) Fu Ke-Planck-Ke Ermoge love equation is solved using Upwind finite difference method, it is flat obtains interference
Equal fieldMore new formula m；
(ii) discrete Lagrange relaxation method is used, Hamilton Jacobi Bellman equation and Fu Ke-Pu Lang are passed through
Ke-Ke Ermoge love equation, solves LagrangianMore new formula λ and power levelMore new formula p:
By Hamilton Jacobi Bellman equation and Fu Ke-Planck-Ke Ermoge love it is equations turned be corresponding
Optimization problem enables Lagrange's equation to interference using the Lagrange's equation of Lagrangian method building optimization problem
The result of mean field derivation is equal to zero, solves LagrangianMore new formula λ, enable Lagrange's equation to power water
The result of flat derivation is equal to zero, solves power levelMore new formula p；
Wherein, mean field is interferedMore new formula m, LagrangianMore new formula λ and power levelIt updates public
Formula p, expression formula are respectively as follows:
Interfere mean fieldMore new formula m:
WhereinIt is the interference mean field at time t and disturbance state I, whereinWhereinIt is that small base station i exists
Power level at time t and disturbance state I, σ2It is small base station i to user UEiO-U channel model in preset variance,
WhereinWherein Wherein κJ, iAnd κI, iIt is respectively
Small base station j to user UEiWith small base station i to user UEiO-U channel model in preset mean value,It is that approximate interference is average
, gI, iIt is small base station i to user UEiChannel gain, wherein ξiIt is preset uncertain in the dynamical state equation of small base station i
Amount, the value being arranged in a certain range is bigger, and the robustness of system is better, but is more than that a certain range will affect power control
Accuracy, wherein δtIt is the iteration step length of time, δIIt is the iteration step length of disturbance state；
LagrangianMore new formula λ:
WhereinIt is the Lagrangian at time t and disturbance state I,WhereinSmall base station i when
Between power level at t and disturbance state I, σ2It is preset small base station i to user UEiO-U channel model in variance,
InWherein Wherein κJ, iAnd κI, iIt is pre- respectively
If small base station j to user UEiWith small base station i to user UEiO-U channel model in mean value,It is that approximate interference is average
, gI, iIt is small base station i to user UEiChannel gain,It is power water of the small base station j at time t and disturbance state I
It puts down, wherein ξiIt is the Uncertainty in the dynamical state equation of preset small base station i, wherein δtIt is the iteration step length of time dimension,
δIIt is the iteration step length of disturbance state dimension, whereinIt is cost of the small base station i in t moment
Function, γthIt is preset SINR threshold value, It is the Background Noise Power of t moment, ρ2It is pre-
If level of robustness, indicate the robustness size of system,It is uncertain in the dynamical state equation of preset small base station i
Amount；
Power levelMore new formula p:
WhereinIt is power level of the small base station i at time t and disturbance state I, whereinIt is time t and disturbance state
Interference mean field at I, whereinIt is the Lagrangian at time t and disturbance state I, σ2It is small base station i to user UEi
O-U channel model in preset variance, whereinWherein κI, iIt is small base station i to user UEiO-U letter
Preset mean value, g in road model equationI, iIt is small base station i to user UEiChannel gain,Small base station j in time t and
Power level at disturbance state I,It is approximate interference mean field, wherein δtIt is the iteration step length of time, δIIt is disturbance state
Iteration step length, wherein γthIt is preset SINR threshold value,It is the Background Noise Power of t moment.
The small base station i of step 6) will interfere mean fieldIt is assigned to termination variableAnd to interference mean fieldIt carries out
Update obtains matrixRealize step are as follows:
Step 6a) small base station i will interfere mean fieldIt is assigned to termination variableTerminate variable
Step 6b) small base station i is to interference mean fieldIt is updated:
Small base station i utilizes interference mean fieldMore new formula m, pass through power levelTo interference mean fieldIn member
Element is updated, and obtains updated matrix
The small base station i of step 7) is by matrixIt is assigned to interference mean fieldInterfere mean field
The small base station i of step 8) is to LagrangianUpdate obtains matrixAnd by matrixIt is assigned to Lagrangian calculation
SonRealize that steps are as follows:
Step 8a) small base station i utilizes LagrangianMore new formula λ, pass through power levelLagrange is calculated
SonIn element be updated, obtain updated matrix
Step 8b) small base station i enables Lagrangian
The small base station i of step 9) is by power levelIt is assigned toAnd to power levelIt is updated, obtains matrix
Realize that steps are as follows:
Step 9a) small base station i is by power levelIt is assigned toI.e.
Step 9b) small base station i is to interference mean fieldIt is updated:
Small base station i utilizes power levelMore new formula p, pass through interference mean fieldAnd LagrangianTo function
Rate is horizontalIn element be updated, obtain updated matrix
The small base station i of step 10) is by matrixIt is assigned to power levelThat is power level
The small base station i of step 11) judges power levelWhether restrain:
Small base station i is calculatedWith termination variableThe absolute value of difference, and whether judge all elements in the value
Less than preset threshold value, if so, power levelConvergence, and by the power levelAs the power level after convergenceIf it is not, repeating step 6)~step 11).
Effect of the invention is described further below in conjunction with analogous diagram.
One, simulated conditions:
Simulation object: a kind of aware distributed robust method for controlling downlink power of interference, and non intelligent adaptive algorithm,
It is compared in energy efficiency.Power Control Problem is configured to a betting model, protected by wherein non intelligent adaptive algorithm
Under the demand for demonstrate,proving user's communication link SINR, the cost function of game person is designed, under the target for minimizing system intraformational interference
Derive the closed solutions of power level, in dense deployment the effect of power control is poor in network environment.
Simulation parameter: assume that macro base station MeNB coverage area is 500m × 500m, uniformly divides in coverage area in scene
The small base station of cloth, we choose the small base station of different number here, calculate separately out corresponding system in the case of every kind small base station density
Unite average EE, and the system under different small base station densities that finally obtains is averaged the corresponding CDF of EE, we choose the quantity of small base station here
It is chosen according to 4,9,16,25 ..., 144 rule, wherein each small base station services a user SUE, chooses one in the scene
A small base station carries out power control using this method, finally obtains the average energy efficiency of system as small base station i, small base station i
(EE), according to the distribution of the macrocellular covering radius of selection and small base station, other small bases can be then calculated by path loss model
The interference stood to small base station i, wherein the channel fading model of same layer link is PL=127+30log10The unit of D, D are km, small
Base station j to user UEiWith the channel gain of small base station iIt is that small base station i is arrived
User UEiO-U channel model in variance, κJ, iAnd κI, iIt is preset small base station j to user UE respectivelyiWith small base station i to use
Family UEiO-U channel model in mean value, enable κ hereJ, i=4 × 10-9, κI, i=1 × 10-9, mixing is listed in table 1 here
Simulated environment parameter setting under cellular network.
1 hybrid cellular network simulated environment parameter setting of table
Two, emulation content and analysis:
Under above-mentioned simulating scenes, statistics obtains the cumulative distribution letter of system average energy efficiency under different small base station densities
Number, as a result as shown in figure 3, ordinate is the accumulation of system average energy efficiency if figure abscissa is system average energy efficiency
Distribution function, the value that identical ordinate value corresponds to abscissa is bigger, then system average behavior is better.As seen from the figure, this hair
Bright performance is better than non intelligent adaptive algorithm.
Claims (4)
1. a kind of aware distributed robust method for controlling downlink power of interference, comprising the following steps:
(1) small base station i calculates the approximate interference mean field of itself:
(1a) in the scene of N number of small base station, all small base station transmitting powers areTest signal a；
The user UE of (1b) small base station i serviceiMeasure the power of received signal and noiseAnd it willFeed back to small base
Stand i；
(1c) small base station i utilizes the power for testing signal aPowerWith small base station i to user UEiChannel gain gi,i, meter
Calculate the average interference channel gain of small base station iMean field is interfered with approximation
(2) time interval [0, T] of small base station i setting power control, and obtain power control disturbance state space [0,
Imax]:
The time interval [0, T] of (2a) small base station i setting power control, wherein the maximum value of T expression time interval；
(2b) small base station i obtains the disturbance state space [0, I of power controlmax]: other in scene in addition to small base station i are small
Base station full power emits signal, user UEiBy the Feedback of Power being interfered give small base station i, small base station i using jamming power as
Maximum interference state Imax, constitute the disturbance state space [0, I of power controlmax]；
(3) small base station i is to time interval [0, T] and disturbance state space [0, Imax] discretization is carried out, and calculate time dimension
Iteration step length δtWith the iteration step length δ of disturbance state dimensionI:
(3a) small base station i is to time interval [0, T] and disturbance state space [0, Imax] discretization is carried out simultaneously, it obtains comprising X
The discrete grid block of period and Y disturbance state section；
(3b) small base station i calculates the iteration step length δ of time dimension by Xt, the iteration step length of disturbance state dimension is calculated by Y
δI；
(4) small base station i building interference mean fieldLagrangianAnd power levelAnd it is initialized:
(4a) small base station i constructs matrix behavior X+1 in discrete grid block, is classified as the interference mean field of Y+1LagrangianAnd power level
Small base station number in (4b) small base station i statistics scene in each disturbance state section accounts for the ratio of small base station total number, obtains
It is distributed to disturbance state；
(4c) small base station i is distributed by disturbance state to interference mean fieldIt is initialized, the interference after being initialized is flat
Equal fieldSimultaneously to LagrangianAnd power levelIt is initialized respectively, the Lagrange after being initialized
OperatorWith the power level after initialization
(5) interference mean field is derivedMore new formula m, LagrangianMore new formula λ and power levelMore new formula
P:
Weaken mean field theory of games using interference, designs the cost function with robustness, and derive Hami of game playing system
- Jacobi-Bellman equation and Fu Ke-Planck-Ke Ermoge love equation, it is solved, is had using different methods
Body implementation method are as follows:
(I) solves Fu Ke-Planck-Ke Ermoge love equation using Upwind finite difference method, obtains interference mean fieldMore new formula m；
(II) uses discrete Lagrange relaxation method, passes through Hamilton Jacobi Bellman equation and Fu Ke-Planck-Ke
Er Moge love equation, solves LagrangianMore new formula λ and power levelMore new formula p:
By Hamilton Jacobi Bellman equation and Fu Ke-Planck-Ke Ermoge love it is equations turned be corresponding optimal
Change problem enables Lagrange's equation average to interference using the Lagrange's equation of Lagrangian method building optimization problem
The result of field derivation is equal to zero, solves LagrangianMore new formula λ, enable Lagrange's equation to seek power level
The result led is equal to zero, solves power levelMore new formula p；
(6) small base station i will interfere mean fieldIt is assigned to termination variableAnd to interference mean fieldIt is updated
To matrixRealize step are as follows:
(6a) small base station i will interfere mean fieldIt is assigned to termination variableTerminate variable
(6b) small base station i is to interference mean fieldIt is updated:
Small base station i utilizes interference mean fieldMore new formula m, pass through power levelTo interference mean fieldIn element
It is updated, obtains updated matrix
(7) small base station i is by matrixIt is assigned to interference mean fieldInterfere mean field
(8) small base station i is to LagrangianUpdate obtains matrixAnd by matrixIt is assigned to Lagrangian
Realize that steps are as follows:
(8a) small base station i utilizes LagrangianMore new formula λ, pass through power levelTo LagrangianIn
Element be updated, obtain updated matrix
(8b) small base station i enables Lagrangian
(9) small base station i is by power levelIt is assigned toAnd to power levelIt is updated, obtains matrixRealize step
It is rapid as follows:
(9a) small base station i is by power levelIt is assigned toI.e.
(9b) small base station i is to interference mean fieldIt is updated:
Small base station i utilizes power levelMore new formula p, pass through interference mean fieldAnd LagrangianTo power water
It is flatIn element be updated, obtain updated matrix
(10) small base station i is by matrixIt is assigned to power levelThat is power level
(11) small base station i judges power levelWhether restrain:
Small base station i is calculatedWith termination variableThe absolute value of difference, and judge whether all elements in the value are less than
Preset threshold value, if so, power levelConvergence, and by the power levelAs the power level after convergence
If it is not, repeating step (6)~step (11).
2. a kind of aware distributed robust method for controlling downlink power of interference according to claim 1, which is characterized in that step
Suddenly the iteration step length δ of time dimension is calculated described in (3b)tWith the iteration step length δ of disturbance state dimensionI, calculation formula difference
Are as follows:
Wherein, X indicates the number of time dimension discrete point in discrete grid block, and Y indicates that disturbance state dimension is discrete in discrete grid block
The number of point, T indicate the maximum value of time interval, ImaxIndicate maximum interference state.
3. a kind of aware distributed robust method for controlling downlink power of interference according to claim 1, which is characterized in that step
Suddenly small base station i described in (4c) is distributed by disturbance state to interference mean fieldIt is initialized, after being initialized
Interfere mean fieldSimultaneously to LagrangianAnd power levelIt is initialized respectively, the drawing after being initialized
Ge Lang operatorWith the power level after initializationRealize step are as follows:
(4c1) small base station i is enabledAnd disturbance state distribution is assigned to interference mean fieldThe first row in remove first row
Other elements outside element, and enable interference mean fieldIn be equal to zero except the every other element of the first row, initialized
Interference mean field afterwards
(4c2) small base station i enables LagrangianIn all elements be equal to zero, after being initialized Lagrange calculate
Son
(4c3) small base station i enables power levelIn all elements be equal to the minimum emissive power p of small base station iminWith small base
Stand the maximum transmission power p of imaxBetween any value, the power level after being initialized
4. a kind of aware distributed robust method for controlling downlink power of interference according to claim 1, which is characterized in that step
Suddenly interference mean field described in (5)More new formula m, LagrangianMore new formula λ and power levelIt updates public
Formula p, expression formula are respectively as follows:
Interfere mean fieldMore new formula m:
WhereinIt is the interference mean field at time t and disturbance state I, whereinWhereinSmall base station i when
Between power level at t and disturbance state I, σ2It is preset small base station i to user UEiO-U channel model in variance,
InWherein Wherein κj,iAnd κi,iIt is pre- respectively
If small base station j to user UEiWith small base station i to user UEiO-U channel model in mean value,It is that approximate interference is average
, gi,iIt is small base station i to user UEiChannel gain, whereinBe in the dynamical state equation of preset small base station i not
Determine amount, wherein δtIt is the iteration step length of time dimension, δIIt is the iteration step length of disturbance state dimension；
LagrangianMore new formula λ:
WhereinIt is the Lagrangian at time t and disturbance state I,WhereinIt is small base station i in time t
With the power level at disturbance state I, σ2It is preset small base station i to user UEiO-U channel model in variance, whereinWherein Wherein κj,iAnd κi,iIt is default respectively
Small base station j to user UEiWith small base station i to user UEiO-U channel model in mean value,It is that approximate interference is average
, gi,iIt is small base station i to user UEiChannel gain,It is power water of the small base station j at time t and disturbance state I
It puts down, wherein ξiIt is the Uncertainty in the dynamical state equation of preset small base station i, wherein δtIt is the iteration step length of time dimension,
δIIt is the iteration step length of disturbance state dimension, whereinIt is cost of the small base station i in t moment
Function, γthIt is preset SINR threshold value, It is the Background Noise Power of t moment, ρ2It is pre-
If level of robustness, indicate the robustness size of system,It is uncertain in the dynamical state equation of preset small base station i
Amount；
Power levelMore new formula p:
WhereinIt is power level of the small base station i at time t and disturbance state I, whereinIt is at time t and disturbance state I
Interference mean field, whereinIt is the Lagrangian at time t and disturbance state I, σ2It is small base station i to user UEiO-
Preset variance in U channel model, whereinWherein κi,iIt is small base station i to user UEiO-U channel mould
Preset mean value, g in type equationi,iIt is small base station i to user UEiChannel gain,It is small base station j in time t and interference
Power level at state I,It is approximate interference mean field, wherein δtIt is the iteration step length of time, δIIt is changing for disturbance state
It rides instead of walk length, wherein γthIt is preset SINR threshold value,It is the Background Noise Power of t moment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710902375.0A CN107613554B (en) | 2017-09-29 | 2017-09-29 | A kind of aware distributed robust method for controlling downlink power of interference |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710902375.0A CN107613554B (en) | 2017-09-29 | 2017-09-29 | A kind of aware distributed robust method for controlling downlink power of interference |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107613554A CN107613554A (en) | 2018-01-19 |
| CN107613554B true CN107613554B (en) | 2019-10-25 |
Family
ID=61057978
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710902375.0A Active CN107613554B (en) | 2017-09-29 | 2017-09-29 | A kind of aware distributed robust method for controlling downlink power of interference |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107613554B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111031600A (en) * | 2019-12-22 | 2020-04-17 | 贵州师范大学 | OFDMA network distributed power robustness control algorithm for cluster flight spacecraft |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104640193A (en) * | 2015-01-30 | 2015-05-20 | 中国联合网络通信集团有限公司 | Interference coordination method and device |
| CN106332257A (en) * | 2016-09-14 | 2017-01-11 | 西安电子科技大学 | Distributed power control method based on interference mean field in super-dense D2D network |
| CN106455032A (en) * | 2016-09-14 | 2017-02-22 | 西安电子科技大学 | Distributed power control method for major interference in ultra-dense network |
| CN107093002A (en) * | 2017-03-02 | 2017-08-25 | 平顶山天安煤业股份有限公司 | A kind of bore closed quality classification and hazard assessment system based on cloud computing |
-
2017
- 2017-09-29 CN CN201710902375.0A patent/CN107613554B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104640193A (en) * | 2015-01-30 | 2015-05-20 | 中国联合网络通信集团有限公司 | Interference coordination method and device |
| CN106332257A (en) * | 2016-09-14 | 2017-01-11 | 西安电子科技大学 | Distributed power control method based on interference mean field in super-dense D2D network |
| CN106455032A (en) * | 2016-09-14 | 2017-02-22 | 西安电子科技大学 | Distributed power control method for major interference in ultra-dense network |
| CN107093002A (en) * | 2017-03-02 | 2017-08-25 | 平顶山天安煤业股份有限公司 | A kind of bore closed quality classification and hazard assessment system based on cloud computing |
Non-Patent Citations (1)
| Title |
|---|
| Downlink Power Control in Self-Organizing Dense Small Cells Underlaying Macrocells: A Mean Field Game;Prabodini Semasinghe, et al.;《IEEE Transactions on Mobile Computing》;20150330;第5卷(第2期);第350-363页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107613554A (en) | 2018-01-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | Interference management in ultra-dense networks: Challenges and approaches | |
| TWI553566B (en) | A self-optimizing deployment cascade control scheme and device based on tdma for indoor small cell in interference environments | |
| Andrews et al. | Autonomous spectrum sharing for mixed LTE femto and macro cells deployments | |
| Su et al. | LTE-U and Wi-Fi coexistence algorithm based on Q-learning in multi-channel | |
| CN102573033B (en) | Multi-Femtocell downlink power interference control method based on game theory | |
| CN102098680B (en) | Dynamic frequency spectrum management method and system | |
| Rengarajan et al. | Self-organizing dynamic fractional frequency reuse on the uplink of OFDMA systems | |
| CN103856947B (en) | The disturbance coordination method that a kind of combined channel selects and power controls | |
| Deruyck et al. | Reducing the power consumption in LTE-Advanced wireless access networks by a capacity based deployment tool | |
| US20180124713A1 (en) | Intelligent deployment cascade control device based on an fdd-ofdma indoor small cell in multi-user and interference environments | |
| CN106454920A (en) | Resource allocation optimization algorithm based on time delay guarantee in LTE (Long Term Evolution) and D2D (Device-to-Device) hybrid network | |
| CN103369542A (en) | Game theory-based common-frequency heterogeneous network power distribution method | |
| CN103648102B (en) | Heterogeneous network interference coordination method based on dynamic zone expansion and power control | |
| Yu et al. | Dynamic resource allocation in TDD-based heterogeneous cloud radio access networks | |
| CN110248370A (en) | The small disturbance coordination method for distinguishing cluster in a kind of dynamic TDD network | |
| CN107613554B (en) | A kind of aware distributed robust method for controlling downlink power of interference | |
| CN106412988A (en) | Improved weighted graph-based super dense heterogeneous network interference coordination method | |
| CN103338453A (en) | Dynamic frequency spectrum access method and system for hierarchical wireless network | |
| CN103428843B (en) | Power coordinating method integrating effectiveness of near field users and effectiveness of distant filed users | |
| CN105764068A (en) | Small base station capacity and coverage optimization method based on tabu search | |
| Marshoud et al. | Macrocell–femtocells resource allocation with hybrid access motivational model | |
| Liu et al. | Distributed enhanced inter-cell interference coordination (eICIC) in LTE-Advanced HetNets: A potential game approach | |
| Li et al. | Distributed power control for two-tier femtocell networks with QoS provisioning based on Q-learning | |
| Zhang et al. | On stochastic geometry modeling of WLAN capacity with dynamic sensitive control | |
| TWI624168B (en) | An intelligent deployment cascade control device based on an fdd-ofdma indoor small cell in multi-user and interference environments |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |