CN111182571B - Long-term joint optimization method for base station activation control and beam forming - Google Patents

Long-term joint optimization method for base station activation control and beam forming Download PDF

Info

Publication number
CN111182571B
CN111182571B CN202010031744.5A CN202010031744A CN111182571B CN 111182571 B CN111182571 B CN 111182571B CN 202010031744 A CN202010031744 A CN 202010031744A CN 111182571 B CN111182571 B CN 111182571B
Authority
CN
China
Prior art keywords
base station
network
processing unit
updating
auxiliary
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
Application number
CN202010031744.5A
Other languages
Chinese (zh)
Other versions
CN111182571A (en
Inventor
马孟园
林静然
杨健
利强
邵怀宗
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Electronic Science and Technology of China
Original Assignee
University of Electronic Science and Technology of China
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by University of Electronic Science and Technology of China filed Critical University of Electronic Science and Technology of China
Priority to CN202010031744.5A priority Critical patent/CN111182571B/en
Publication of CN111182571A publication Critical patent/CN111182571A/en
Application granted granted Critical
Publication of CN111182571B publication Critical patent/CN111182571B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/02—Arrangements for optimising operational condition
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/06—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station
    • H04B7/0613—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission
    • H04B7/0615—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission of weighted versions of same signal
    • H04B7/0617—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission of weighted versions of same signal for beam forming
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/08—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the receiving station
    • H04B7/0837—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the receiving station using pre-detection combining
    • H04B7/0842—Weighted combining
    • H04B7/086—Weighted combining using weights depending on external parameters, e.g. direction of arrival [DOA], predetermined weights or beamforming
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/06—Testing, supervising or monitoring using simulated traffic
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Abstract

The invention discloses a long-term joint optimization method for base station activation control and beamforming, which comprises the following steps: initializing a central processing unit, a network control unit and an auxiliary processing unit in the MISO network, wherein the network control unit collects and processes information such as channels in the network, service states of base stations and the like, the central processing unit performs corresponding analysis according to received current environment information, the network control unit opens or closes a designated base station, then a forming beam is calculated and configured by combining channel information, meanwhile, a power consumption indicating quantity is fed back, the auxiliary processing unit receives and stores historical data of system operation, the data stored by the auxiliary processing unit is utilized to assist the central processing unit to update the weight of the neural network, and the auxiliary system operates normally. The invention provides a self-adaptive online solution, which reduces network power and solves the problems of low environmental adaptability, complex operation, poor practicability, high operation and maintenance cost and large capital investment in the prior art.

Description

Long-term joint optimization method for base station activation control and beam forming
Technical Field
The invention belongs to the technical field of wireless communication, and particularly relates to a long-term joint optimization method for base station activation control and beam forming.
Background
In the context of heterogeneous networks, many different types of base stations are often deployed within a cell, and maintaining the operation of these base stations consumes a significant amount of energy. Base station activation control reduces network energy consumption by shutting down operating base stations that are serving redundancies within a cell, a technique that has received significant attention in recent decades. Base station active control and beamforming typically work together to force the network to serve as many users as possible within an acceptable range of QoS. Since these problems are NP-hard (NP-hard refers to the absence of an algorithm with polynomial time to solve the problem) mixed integer programming, researchers have often adopted convex interval relaxation or applied sparsity constraints to achieve acceptable complexity processing.
Although the existing schemes have differences in formula derivation and algorithm design, the past research is to activate and deactivate the base station based on the instant Channel State Information (CSI), so the prior art has the following problems:
(1) the operation state of the base station can be changed at any time, so that the management of the network tends to be complex in actual operation, and the practicability of the activation control technology is low;
(2) the burden overhead of the system due to management is heavy, the operation and maintenance cost of the system is high, and the capital investment is large.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a long-term joint optimization method of base station activation control and beam forming, which is simple and convenient to operate, reduces operation and maintenance cost and saves capital investment, and the base station activation control in a period of time is considered to control the operation state switching frequency of a base station so as to maintain the flexibility of base station activation and simultaneously keep the stability of the operation state of the base station to a certain degree, thereby improving the practicability and solving the problems of low environmental adaptability, complex operation, low practicability, high operation and maintenance cost and large capital investment in the prior art.
In order to achieve the purpose of the invention, the invention adopts the technical scheme that:
a long-term joint optimization method for base station activation control and beam forming comprises the following steps:
s1, initializing parameters and configurations of a central processing unit, a network control unit and an auxiliary processing unit in the MISO network;
s2, collecting and processing the service state information of the channel and the base station in the network by using the network control unit, and outputting the processed information to the central processing unit;
s3, analyzing by using the central processing unit according to the received current environment information, and returning the base station activation and closing scheme to the network control unit;
s4, the network control unit is used for turning on or off the appointed base station according to the received scheme, then the shaped beam is calculated and configured by combining the channel information, and the power consumption indication quantity in the network is fed back;
s5, utilizing the auxiliary processing unit to receive and save the history data of the system operation, including environment information, base station configuration scheme and power consumption indication amount;
s6, assisting the central processing unit to update the weight of the neural network by using the data stored by the auxiliary processing unit, and assisting the normal operation of the system;
and S7, judging whether the network service is finished, if not, returning to the step S2, otherwise, finishing the process.
Further, in step S1, initializing parameters and configurations of the central processing unit, the network control unit and the auxiliary processing unit in the MISO network includes:
setting the number K of base stations, the number M of antennas at a transmitting end, the number I of downlink users and the number T of division time slots in a network;
setting the maximum operating power P of the base stationmaxNoise power σ2Lowest signal-to-noise ratio gamma of user link and control quantity lambda for maintaining operation overhead of base station1Control quantity lambda for switching working state of base station2；
Setting the neural network structure of the central processing unit, initializing the network weight theta and controlling parameters alpha and XfInitializing an auxiliary processing unit memory upper limit D and an auxiliary control parameter Xb。
Further, in step S2, the network control unit is used to collect and process the current channel in the network and the service status information of the base station at the previous time, and output the processed information to the central processing unit, which is represented as:
st＝φ(Ht,αt-1)
wherein s istThe current environment information after being processed; ht,αt-1Respectively serving current channel information and last time base station service state information in the networkPhi (-) is a mapping function within the network control unit.
Further, in step S4, the network control unit is utilized to turn on or off the designated base station according to the received scheme, and then combine the channel information HtCalculating and configuring shaped beam WtWhile feeding back the power consumption indicator r in the networktThe method comprises the following steps:
s4-1, setting auxiliary variables and intermediate parameters;
s4-2, initializing iteration control parameters;
s4-3, updating the current objective function value, wherein the updating formula is as follows:
Figure BDA0002364565660000031
wherein the content of the first and second substances,
Figure BDA0002364565660000032
in order to update the value of the objective function,
Figure BDA0002364565660000033
the value is the next generation objective function value, and n is the iteration number indicating quantity;
s4-4, calculating the next generation beam forming matrix according to the auxiliary variable and the intermediate parameter, and updating the auxiliary variable and the intermediate parameter;
s4-5, calculating a next generation objective function value, updating an inner layer iteration parameter, judging whether the next generation objective function value, the current objective function value and an inner layer iteration number indicating quantity meet requirements, if so, outputting a beam forming matrix and entering the step S4-6, otherwise, entering the step S4-3;
s4-6, according to the calculated beam forming matrix WtLast time base station service state indication quantity alphat-1And current base station service state indication quantity alphatAn indication of the power consumption r in the output networkt。
Further, in step S4-4, the method calculates a next generation beamforming matrix according to the auxiliary variable and the intermediate parameter, and updates the auxiliary variable and the intermediate parameter, including the following sub-steps:
s4-4-1, calculating the next generation beam forming matrix of the inner layer according to the current auxiliary variable and the current intermediate parameter, wherein the calculation formula is as follows:
Figure BDA0002364565660000041
wherein, WkA beamforming matrix for a base station K to users in the network;
s4-4-2, updating an auxiliary variable F, wherein the updating formula is as follows:
definition of
Figure BDA0002364565660000042
Figure BDA0002364565660000043
Wherein f isiIs the i-th row of the F,
Figure BDA0002364565660000044
is fiThe (i) th element of (a),
Figure BDA0002364565660000045
is fiRemoving
Figure BDA0002364565660000046
Vector of remaining elements, yiIs the i-th row of the Y,
Figure BDA0002364565660000047
is yiThe (i) th element of (a),
Figure BDA0002364565660000048
is yiRemoving
Figure BDA0002364565660000049
Vectors of the remaining elements, beta being a parameter satisfying a constraint condition [ ·]+Taking a non-negative value, wherein 1 is a column vector with all elements being 1;
s4-4-3, updating the auxiliary variable U, wherein the updating formula is as follows:
Figure BDA0002364565660000051
wherein, FuAnd phiuAre respectively Fn+1And phinRemoving the left sub-matrix of the last column;
s4-4-4, updating the intermediate parameters { Ψ, Φ }, wherein the updating formula is as follows:
Figure BDA0002364565660000052
further, in step S4-5, the formula for determining whether the next-generation objective function value, the current objective function value, and the indication quantity of the number of iterations in the inner layer meet the requirement is:
Figure BDA0002364565660000053
wherein the content of the first and second substances,
Figure BDA0002364565660000054
in order to update the value of the objective function,
Figure BDA0002364565660000055
and the value is an objective function of the next generation, N is an iteration number indicating quantity, N is the maximum iteration number, and err is the difference value of the objective function before and after iteration.
Further, in the step S4-6, the beamforming matrix W is calculated according to the calculated beamforming matrix WtLast time base station service state indication quantity alphat-1And current base station service state indication quantity alphatAn indication of the power consumption r in the output networktThe calculation formula of (2) is as follows:
Figure BDA0002364565660000056
Δ＝max(0,γ-γmm)
Pall(t)＝Pt(t)+λ1Pm(t)+λ2Ps(t)
wherein, C1As a feedback quantity amplitude control factor, C2Penalty factor, P, for violating user link quality of service requirementst(t),Pm(t),Ps(t) Total Transmission Power, maintenance base station operating Power and base station State switching Power, λ, of base stations in the slotted t network, respectively1,λ2Respectively maintaining the control quantity of the base station operation overhead and the control quantity of the base station working state switching, wherein gamma is a preset lower limit of user service quality, and gamma ismmAnd taking the upper limit value of the user service quality in the network when the decision a is taken for the time slot t.
Further, in step S5, the receiving and storing the history data of the system operation by using the auxiliary processing unit specifically includes:
the environmental information, the base station configuration scheme, the power consumption indication amount and the environmental information at the next time are recorded in the memory of the auxiliary processing unit as a tuple each time.
Further, in step S6, the data stored in the auxiliary processing unit is used to assist the central processing unit in updating the neural network weight, so as to assist the normal operation of the system, which specifically includes the following sub-steps:
s6-1, uniformly sampling X from the memory by using the auxiliary processing unitbThe experience bar group (s, a, r, s') is output to the central processing unit;
s6-2, preparing a label, which is expressed as:
Figure BDA0002364565660000061
s6-3, defining a loss function loss | | | z-B (S | θ) | ceiling2Updating the weight of the neural network B (s | theta) to be estimated according to the loss function, and expressing as follows:
Figure BDA0002364565660000062
s6-4, every interval X of central processing unitfStep updating step, target neural network B-(s|θ-) The weight is updated to be the weight theta of the neural network B (s | theta) to be estimated-←θ。
The invention has the following beneficial effects:
(1) the invention has obvious advantages in the aspects of base station management control and communication link switching in the network, ensures the QoS condition, simultaneously reduces the system overhead and the extra resource cost, reduces the operation and maintenance cost of the system, and saves the capital investment;
(2) the invention divides a period into a plurality of time slots, models the channel into a block fading channel, and approximates the channel in the time slot to a flat fading channel, thereby greatly simplifying the processing of the subsequent problems, further reducing the operation complexity and improving the practicability.
Drawings
Fig. 1 is a flow chart of a long-term joint optimization method of base station activation control and beamforming according to the present invention;
FIG. 2 is a schematic of the overall design of the present invention;
fig. 3 is a graph comparing the results of power consumption simulation for the present invention and the conventional scheme.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
As shown in fig. 1 and fig. 2, an embodiment of the present invention provides a long-term joint optimization method for base station active control and beamforming, including the following steps S1 to S7:
s1, initializing parameters and configurations of a central processing unit, a network control unit and an auxiliary processing unit in the MISO network;
in this embodiment, initializing parameters and configurations of a Central Processing Unit (CPU), a Network Control Unit (NCU), and an Auxiliary Processing Unit (APU) in an MISO (multiple-input-single-output) network according to the present invention includes:
setting the number K of base stations, the number M of antennas at a transmitting end, the number I of downlink users and the number T of division time slots in a network;
setting the maximum operating power P of the base stationmaxNoise power σ2Lowest signal-to-noise ratio gamma of user link and control quantity lambda for maintaining operation overhead of base station1Control quantity lambda for switching working state of base station2；
Setting the neural network structure of the central processing unit, initializing the network weight theta and controlling parameters alpha and XfInitializing an auxiliary processing unit memory upper limit D and an auxiliary control parameter Xb。
S2, collecting and processing the service state information of the channel and the base station in the network by using the network control unit, and outputting the processed information to the central processing unit;
in this embodiment, the present invention utilizes the network control unit to collect and process the current channel and the last-time base station service status information in the network, and outputs the processed information to the central processing unit, which is represented as:
st＝φ(Ht,αt-1)
wherein s istThe current environment information after being processed; ht,αt-1Respectively, current channel information in the network and the last time base station service state information, phi (-) is a mapping function in the network control unit.
S3, analyzing by using the central processing unit according to the received current environment information, and returning the base station activation and closing scheme to the network control unit;
in this embodiment, the base station activation and deactivation scheme referred to in the present invention is the base station status indication vector αtBase station status indication vector αtIs 0 or 1, indicating that the corresponding base station is activated or deactivated.
S4, the network control unit is used for turning on or off the appointed base station according to the received scheme, then the shaped beam is calculated and configured by combining the channel information, and the power consumption indication quantity in the network is fed back;
in the embodiment, the invention utilizes the network control unit to turn on or off the appointed base station according to the received scheme, and then combines the channel information HtCalculating and configuring shaped beam WtWhile feeding back the power consumption indicator r in the networktThe method further comprises the following steps:
s4-1, setting auxiliary variables and intermediate parameters;
s4-2, initializing iteration control parameters;
s4-3, updating the current objective function value, wherein the updating formula is as follows:
Figure BDA0002364565660000081
wherein the content of the first and second substances,
Figure BDA0002364565660000082
in order to update the value of the objective function,
Figure BDA0002364565660000083
the value is the next generation objective function value, and n is the iteration number indicating quantity;
s4-4, calculating the next generation beamforming matrix according to the auxiliary variable and the intermediate parameter, and updating the auxiliary variable and the intermediate parameter, including the following steps:
s4-4-1, calculating the next generation beam forming matrix of the inner layer according to the current auxiliary variable and the current intermediate parameter, wherein the calculation formula is as follows:
Figure BDA0002364565660000091
wherein, WkBeamforming matrices, C, for base stations K to users in the networkM×IThe complex matrix of M × I is represented by the formula
Figure BDA0002364565660000092
Wherein, UkAnd ΨkIs an intermediate parameter, δ is a parameter satisfying a constraint condition, and ρ is an iterative convergence rate control quantity.
If it is
Figure BDA0002364565660000093
δ is 0;
otherwise, search delta by dichotomy>0 satisfies
Figure BDA0002364565660000094
S4-4-2, updating an auxiliary variable F, wherein the updating formula is as follows:
definition of
Figure BDA0002364565660000095
Figure BDA0002364565660000096
Wherein f isiIs the i-th row of the F,
Figure BDA0002364565660000101
is fiThe (i) th element of (a),
Figure BDA0002364565660000102
is fiRemoving
Figure BDA0002364565660000103
Vector of remaining elements, yiIs the i-th row of the Y,
Figure BDA0002364565660000104
is yiThe (i) th element of (a),
Figure BDA0002364565660000105
is yiRemoving
Figure BDA0002364565660000106
Vectors of the remaining elements, beta being a parameter satisfying a constraint condition [ ·]+Taking a non-negative value, wherein 1 is a column vector with all elements being 1;
s4-4-3, updating the auxiliary variable U, wherein the updating formula is as follows:
Figure BDA0002364565660000107
wherein, FuAnd phiuAre respectively Fn+1And phinRemoving the left sub-matrix of the last column;
s4-4-4, updating the intermediate parameters { Ψ, Φ }, wherein the updating formula is as follows:
Figure BDA0002364565660000108
s4-5, calculating the next generation objective function value, updating the inner layer iteration parameter, and judging whether the next generation objective function value, the current objective function value and the inner layer iteration frequency indication quantity meet the requirements, wherein the judgment formula is as follows:
Figure BDA0002364565660000109
wherein the content of the first and second substances,
Figure BDA00023645656600001010
in order to update the value of the objective function,
Figure BDA00023645656600001011
and the value is an objective function of the next generation, N is an iteration number indicating quantity, N is the maximum iteration number, and err is the difference value of the objective function before and after iteration.
If yes, outputting a beamforming matrix and entering the step S4-6, otherwise, entering the step S4-3;
s4-6, according to the calculated beam forming matrix WtLast moment base station clothesTraffic status indicator quantity alphat-1And current base station service state indication quantity alphatAn indication of the power consumption r in the output networktThe calculation formula is as follows:
Figure BDA00023645656600001012
Δ＝max(0,γ-γmm)
Pall(t)＝Pt(t)+λ1Pm(t)+λ2Ps(t)
wherein, C1As a feedback quantity amplitude control factor, C2Penalty factor, P, for violating user link quality of service requirementst(t),Pm(t),Ps(t) Total Transmission Power, maintenance base station operating Power and base station State switching Power, λ, of base stations in the slotted t network, respectively1,λ2Respectively maintaining the control quantity of the base station operation overhead and the control quantity of the base station working state switching, wherein gamma is a preset lower limit of user service quality, and gamma ismmThe upper limit of the user service quality in the network when the decision a is taken for the time slot t can be calculated by the dichotomy and is expressed as
Pm(t)＝pm||at||1
Ps(t)＝ps||at-at-1||1
Figure BDA0002364565660000111
Wherein p ismAnd psRespectively, the power consumed for maintaining the operation of the base station and the power consumed for switching the service state of the base station are set to be a proper normal number.
S5, utilizing the auxiliary processing unit to receive and save the history data of the system operation, including environment information, base station configuration scheme and power consumption indication amount;
in this embodiment, the present invention provides the environment information s each timetBase station configuration scheme at-1Power, powerConsumption indicating quantity rtAnd the next time environmental information st+1Recorded as a tuple in the auxiliary processing unit memory.
S6, assisting the central processing unit to update the weight of the neural network by using the data stored by the auxiliary processing unit, and assisting the normal operation of the system;
in this embodiment, the data stored in the auxiliary processing unit is used to assist the central processing unit to update the neural network weight, and the normal operation of the auxiliary system is assisted, which specifically includes the following steps:
s6-1, uniformly sampling X from the memory by using the auxiliary processing unitbA set of experience bars (s, a, r, s '), where s, a, r, s' respectively denote s for any tt、at-1、rt、st+1And output to the central processing unit;
s6-2, preparing a label, which is expressed as:
Figure BDA0002364565660000121
wherein z represents a label and a' is represented at st+1A temporal base station configuration scheme;
s6-3, defining a loss function loss | | | z-B (S | θ) | ceiling2Updating the weight of the neural network B (s | theta) to be estimated according to the loss function, and expressing as follows:
Figure BDA0002364565660000122
s6-4, every interval X of central processing unitfStep updating step, target neural network B-(s|θ-) The weight is updated to be the weight theta of the neural network B (s | theta) to be estimated-←θ。
Wherein v ═ B (s | θ) is the network to be estimated; v. of-＝B-(s|θ-) Is a target network; v. ofaB (s, a | θ) represents a value v corresponding to the selected base station configuration scheme aa(ii) a Where theta and theta-Representing a neural network weight parameter;
Figure BDA0002364565660000123
is shown in (2)KValue vectors corresponding to the base station configuration schemes; theta and target theta-Representing the network weight parameter.
And S7, judging whether the network service is finished, if not, returning to the step S2, otherwise, finishing the process.
In the embodiment of the invention, a cell with 5 4-antenna base stations BS and 5 single-antenna users is considered, K is 5, I is 5, the users are randomly distributed in a hexagonal cell with the radius of 1km, and the base station is positioned in the center of the cell. Suppose a transmission process is divided into 10 transmission time slots, and T is 10; in the specific implementation process, a standard complex Gaussian is used to generate a channel matrix H, the variance of the channel matrix H follows a probability distribution model of a cell and a user, and lambda is set250, user signal-to-noise ratio minimum threshold gamma 1, and transmitting power ps1, variance of noise σ2Base station operation maintenance power p 1mBase station service state switching power p 1s1, feedback quantity amplitude control factor C1＝103Penalty factor C for violating user link QoS requirement2＝106The CPU control parameter α is 0.3, Xf5, the APU memory upper limit D is 10, and the auxiliary control parameter Xb＝10；
Calculating the beam forming matrix W of each base station in each time slotkThe iteration number indicating amount n is 0, the iteration convergence rate control amount ρ is 10, and the minimum difference err between the target values before and after the iteration is 10-2The maximum number of iterations N is 100. The simulation performance of the invention and the traditional scheme is changed when the control quantity lambda is changed1The comparison is carried out under the value, and the simulation results of the two algorithms are shown in figure 3.
The long-term combined optimization method for base station activation control and beam forming, which is simple and convenient to operate, reduces operation and maintenance cost and saves capital investment, improves the practicability and solves the problems of low environmental adaptability, complex operation, low practicability, high operation and maintenance cost and large capital investment in the prior art.
It will be appreciated by those of ordinary skill in the art that the embodiments described herein are intended to assist the reader in understanding the principles of the invention and are to be construed as being without limitation to such specifically recited embodiments and examples. Those skilled in the art can make various other specific changes and combinations based on the teachings of the present invention without departing from the spirit of the invention, and these changes and combinations are within the scope of the invention.

Claims (3)

1. A long-term joint optimization method for base station activation control and beamforming is characterized by comprising the following steps:
s1, initializing parameters and configurations of the hub processing unit, the network control unit and the auxiliary processing unit in the MISO network, including:
setting the number K of base stations, the number M of antennas at a transmitting end, the number I of downlink users and the number T of division time slots in a network;
setting the maximum operating power P of the base stationmaxNoise power σ2Lowest signal-to-noise ratio gamma of user link and control quantity lambda for maintaining operation overhead of base station1Control quantity lambda for switching working state of base station2；
Setting the neural network structure of the central processing unit, initializing the network weight theta and controlling parameters alpha and XfInitializing an auxiliary processing unit memory upper limit D and an auxiliary control parameter Xb；
S2, using the network control unit to collect and process the channel in the network and the service state information of the base station, outputting the processed information to the central processing unit, and showing as:
st＝φ(Ht,αt-1)
wherein s istThe current environment information after being processed; ht,αt-1Respectively the current channel information and the last time base station service state information in the network, phi (-) is a mapping function in the network control unit;
s3, analyzing by using the central processing unit according to the received current environment information, and returning the base station activation and closing scheme to the network control unit; wherein the base station activation and deactivation scheme is a base station state indication vector alphatBase station status indication vector αtThe element of (1) is 0 or 1, which indicates that the corresponding base station is activated or closed;
s4, the network control unit is used to turn on or off the appointed base station according to the received scheme, then the shaped beam is calculated and configured by combining the channel information, and the power consumption indication quantity in the network is fed back, the method comprises the following steps:
s4-1, setting auxiliary variables and intermediate parameters;
s4-2, initializing iteration control parameters;
s4-3, updating the current objective function value, wherein the updating formula is as follows:
Figure FDA0002841030640000021
wherein the content of the first and second substances,
Figure FDA0002841030640000022
in order to update the value of the objective function,
Figure FDA0002841030640000023
the value is the next generation objective function value, and n is the iteration number indicating quantity;
s4-4, calculating the next generation beamforming matrix according to the auxiliary variable and the intermediate parameter, and updating the auxiliary variable and the intermediate parameter, including the following steps:
s4-4-1, calculating the next generation beam forming matrix of the inner layer according to the current auxiliary variable and the current intermediate parameter, wherein the calculation formula is as follows:
Figure FDA0002841030640000024
wherein, WkA beamforming matrix for a base station K to users in the network;
s4-4-2, updating an auxiliary variable F, wherein the updating formula is as follows:
definition of
Figure FDA0002841030640000025
Figure FDA0002841030640000026
Wherein f isiI row of F, Fi iIs fiThe (i) th element of (a),
Figure FDA0002841030640000027
is fiRemoving fi iVector of remaining elements, yiIs the i-th row of the Y,
Figure FDA0002841030640000028
is yiThe (i) th element of (a),
Figure FDA0002841030640000029
is yiRemoving
Figure FDA00028410306400000210
Vectors of the remaining elements, beta being a parameter satisfying a constraint condition [ ·]+Taking a non-negative value, wherein 1 is a column vector with all elements being 1;
s4-4-3, updating the auxiliary variable U, wherein the updating formula is as follows:
Figure FDA00028410306400000211
wherein, FuAnd phiuAre respectively Fn+1And phinRemoving the left sub-matrix of the last column;
s4-4-4, updating the intermediate parameters { Ψ, Φ }, wherein the updating formula is as follows:
Figure FDA0002841030640000031
s4-5, calculating a next generation objective function value, updating an inner layer iteration parameter, judging whether the next generation objective function value, the current objective function value and an inner layer iteration number indicating quantity meet requirements, if so, outputting a beam forming matrix and entering the step S4-6, otherwise, entering the step S4-3;
s4-6, according to the calculated beam forming matrix WtLast time base station service state indication quantity alphat-1And current base station service state indication quantity alphatAn indication of the power consumption r in the output networkt；
S5, receiving and storing historical data of system operation, including environment information, base station configuration scheme, and power consumption indication amount, by using the auxiliary processing unit, specifically:
recording the environment information, the base station configuration scheme, the power consumption indication quantity and the environment information at the next moment as a tuple in the memory of the auxiliary processing unit each time;
s6, assisting the central processing unit to update the weight of the neural network by using the data stored by the auxiliary processing unit, and assisting the normal operation of the system, specifically comprising the following steps:
s6-1, uniformly sampling X from the memory by using the auxiliary processing unitbThe experience bar group (s, a, r, s') is output to the central processing unit; wherein s, a, r, s' each represent s of any tt、at-1、rt、st+1；
S6-2, preparing a label, which is expressed as:
Figure FDA0002841030640000032
s6-3, defining a loss function loss | | | z-B (S | θ) | ceiling2Updating the weight of the neural network B (s | theta) to be estimated according to the loss function, and expressing as follows:
Figure FDA0002841030640000033
s6-4, pivot processingUnit per interval XfStep updating step, target neural network B-(s|θ-) The weight is updated to be the weight theta of the neural network B (s | theta) to be estimated-←θ；
And S7, judging whether the network service is finished, if not, returning to the step S2, otherwise, finishing the process.
2. The long-term joint optimization method for base station activation control and beamforming as claimed in claim 1, wherein in step S4-5, the formula for determining whether the requirement is met according to the next-generation objective function value, the current objective function value and the indication quantity of the number of inner-layer iterations is:
Figure FDA0002841030640000041
wherein the content of the first and second substances,
Figure FDA0002841030640000042
in order to update the value of the objective function,
Figure FDA0002841030640000043
and the value is an objective function of the next generation, N is an iteration number indicating quantity, N is the maximum iteration number, and err is the difference value of the objective function before and after iteration.
3. The long-term joint optimization method for base station active control and beamforming as claimed in claim 1, wherein in step S4-6, based on the calculated beamforming matrix WtLast time base station service state indication quantity alphat-1And current base station service state indication quantity alphatAn indication of the power consumption r in the output networktThe calculation formula of (2) is as follows:
Figure FDA0002841030640000044
Δ＝max(0,γ-γmm)
Pall(t)＝Pt(t)+λ1Pm(t)+λ2Ps(t)
wherein, C1As a feedback quantity amplitude control factor, C2Penalty factor, P, for violating user link quality of service requirementst(t),Pm(t),Ps(t) Total Transmission Power, maintenance base station operating Power and base station State switching Power, λ, of base stations in the slotted t network, respectively1,λ2Respectively maintaining the control quantity of the base station operation overhead and the control quantity of the base station working state switching, wherein gamma is a preset lower limit of user service quality, and gamma ismmAnd taking the upper limit value of the user service quality in the network when the decision a is taken for the time slot t.
CN202010031744.5A 2020-01-13 2020-01-13 Long-term joint optimization method for base station activation control and beam forming Active CN111182571B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010031744.5A CN111182571B (en) 2020-01-13 2020-01-13 Long-term joint optimization method for base station activation control and beam forming

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010031744.5A CN111182571B (en) 2020-01-13 2020-01-13 Long-term joint optimization method for base station activation control and beam forming

Publications (2)

Publication Number Publication Date
CN111182571A CN111182571A (en) 2020-05-19
CN111182571B true CN111182571B (en) 2021-02-02

Family

ID=70656432

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010031744.5A Active CN111182571B (en) 2020-01-13 2020-01-13 Long-term joint optimization method for base station activation control and beam forming

Country Status (1)

Country Link
CN (1) CN111182571B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112996090B (en) * 2021-01-21 2022-08-23 西藏先锋绿能环保科技股份有限公司 Energy-saving management system and method
CN113489521B (en) * 2021-05-26 2023-05-09 电子科技大学 Reflective surface assisted cell-free large-scale MIMO network combined beam forming method

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104581905A (en) * 2014-12-31 2015-04-29 华中科技大学 Method for collaborating activation of small base stations in HetNet
CN104581780A (en) * 2014-12-18 2015-04-29 哈尔滨工业大学 Preprocessing-based branch dividing and cutting combined network optimizing and wave beam forming method
CN106211199A (en) * 2016-07-20 2016-12-07 电子科技大学 Base station activation for MISO wireless network accesses combination control method with user
CN107087279A (en) * 2017-06-13 2017-08-22 电子科技大学 It is a kind of based on the base station activation and beam-forming method of stablizing access relation
EP3413616A1 (en) * 2016-02-04 2018-12-12 Ntt Docomo, Inc. User device, base station, channel identification method, and identifier transmission method

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20180227772A1 (en) * 2017-02-06 2018-08-09 Mediatek Inc. Mechanism for Beam Reciprocity Determination and Uplink Beam Management
WO2019161784A1 (en) * 2018-02-23 2019-08-29 Mediatek Inc. Default beam for uplink transmission after connection reestablishment
CN109743268B (en) * 2018-12-06 2022-02-15 东南大学 Millimeter wave channel estimation and compression method based on deep neural network

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104581780A (en) * 2014-12-18 2015-04-29 哈尔滨工业大学 Preprocessing-based branch dividing and cutting combined network optimizing and wave beam forming method
CN104581905A (en) * 2014-12-31 2015-04-29 华中科技大学 Method for collaborating activation of small base stations in HetNet
EP3413616A1 (en) * 2016-02-04 2018-12-12 Ntt Docomo, Inc. User device, base station, channel identification method, and identifier transmission method
CN106211199A (en) * 2016-07-20 2016-12-07 电子科技大学 Base station activation for MISO wireless network accesses combination control method with user
CN107087279A (en) * 2017-06-13 2017-08-22 电子科技大学 It is a kind of based on the base station activation and beam-forming method of stablizing access relation

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
Joint Long-Term Admission Control and Beamforming in Downlink MISO Networks;Jingran Lin 等;《2018 26th European Signal Processing Conference(EUSIPCO)》;20180907;全文 *

Also Published As

Publication number Publication date
CN111182571A (en) 2020-05-19

Similar Documents

Publication Publication Date Title
CN109729528B (en) D2D resource allocation method based on multi-agent deep reinforcement learning
Lei et al. A deep learning approach for optimizing content delivering in cache-enabled HetNet
Chu et al. Power control in energy harvesting multiple access system with reinforcement learning
CN111182571B (en) Long-term joint optimization method for base station activation control and beam forming
Cui et al. Delay-aware BS discontinuous transmission control and user scheduling for energy harvesting downlink coordinated MIMO systems
Wang et al. Joint interference alignment and power control for dense networks via deep reinforcement learning
CN109618399A (en) Distributed energy management solutions optimization method in the mobile edge calculations system of multi-user
Van Huynh et al. Digital twin empowered ultra-reliable and low-latency communications-based edge networks in industrial IoT environment
Balevi et al. A clustering algorithm that maximizes throughput in 5G heterogeneous F-RAN networks
CN110300417B (en) Energy efficiency optimization method and device for unmanned aerial vehicle communication network
Zhao et al. Context-aware TDD configuration and resource allocation for mobile edge computing
Tham et al. Deep reinforcement learning for resource allocation in 5G communications
Iqbal et al. Double deep Q-network for power allocation in cloud radio access network
Fourati et al. An efficient energy-saving scheme using genetic algorithm for 5G heterogeneous networks
Li et al. Throughput maximization by deep reinforcement learning with energy cooperation for renewable ultradense IoT networks
Masoudi et al. Digital twin assisted risk-aware sleep mode management using deep Q-networks
Xiao et al. Towards energy efficient resource allocation: When green mobile edge computing meets multi-agent deep reinforcement learning
Rahmani et al. Deep reinforcement learning-based power allocation in uplink cell-free massive MIMO
Azizi et al. MIX-MAB: Reinforcement learning-based resource allocation algorithm for LoRaWAN
CN109089307B (en) Energy-collecting wireless relay network throughput maximization method based on asynchronous dominant actor critic algorithm
Lu et al. Joint beamforming and power control for MIMO-NOMA with deep reinforcement learning
Wang et al. Sequential offloading for distributed dnn computation in multiuser mec systems
Sun et al. Autonomous cell activation for energy saving in cloud-RANs based on dueling deep Q-network
CN116074974A (en) Multi-unmanned aerial vehicle group channel access control method under layered architecture
Chen et al. Adaptive multi-objective reinforcement learning for pareto frontier approximation: A case study of resource allocation network in massive mimo

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