CN113099520A - Base station dormancy method based on hysteresis noise chaotic neural network - Google Patents

Base station dormancy method based on hysteresis noise chaotic neural network Download PDF

Info

Publication number
CN113099520A
CN113099520A CN202110296601.1A CN202110296601A CN113099520A CN 113099520 A CN113099520 A CN 113099520A CN 202110296601 A CN202110296601 A CN 202110296601A CN 113099520 A CN113099520 A CN 113099520A
Authority
CN
China
Prior art keywords
base station
micro base
energy
neural network
constraint
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.)
Granted
Application number
CN202110296601.1A
Other languages
Chinese (zh)
Other versions
CN113099520B (en
Inventor
孙文胜
朱讯
吴启辉
赵莹莹
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.)
Hangzhou Dianzi University
Original Assignee
Hangzhou Dianzi University
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 Hangzhou Dianzi University filed Critical Hangzhou Dianzi University
Priority to CN202110296601.1A priority Critical patent/CN113099520B/en
Publication of CN113099520A publication Critical patent/CN113099520A/en
Application granted granted Critical
Publication of CN113099520B publication Critical patent/CN113099520B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W52/00—Power management, e.g. TPC [Transmission Power Control], power saving or power classes
    • H04W52/02—Power saving arrangements
    • H04W52/0203—Power saving arrangements in the radio access network or backbone network of wireless communication networks
    • H04W52/0206—Power saving arrangements in the radio access network or backbone network of wireless communication networks in access points, e.g. base stations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Molecular Biology (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The invention provides a base station dormancy method based on a hysteresis noise chaotic neural network. According to the method, a hysteresis noise chaotic neural network model is introduced, constraint conditions and a target function in a base station sleep strategy are expressed as an energy function of the hysteresis noise chaotic neural network model, a dynamic equation is obtained by derivation and is substituted into an algorithm for iteration, an optimal sleep scheme is obtained when the energy function is converged, and the obtained scheme meets the constraint conditions and can achieve optimal energy efficiency. The method can show both hysteresis power and random chaotic simulated annealing, has the capability of jumping out of local extreme values, and the obtained dormancy scheme has lower energy efficiency than the traditional method.

Description

Base station dormancy method based on hysteresis noise chaotic neural network
Technical Field
The invention relates to the field of heterogeneous dense network systems, in particular to a method for reducing network energy consumption through base station dormancy in a heterogeneous dense network.
Background
The explosive growth of User Equipments (UEs) in 5G and the introduction of new smart devices for Multimedia Broadband Services (MBS) have resulted in a great increase in traffic. As one of the 5G new radio concepts, Ultra Dense Networks (UDNs) can meet the demand for high traffic and high traffic by densely deploying micro base stations (SBS) indoors and in hot spots to meet the demand for 1000 times of data traffic in the future. In UDNs, Macro base stations (Macro-BS) provide basic coverage, while low power micro base stations (SBS) are densely deployed within the coverage of Macro-BS.
However, densification of micro base stations greatly increases the overall energy consumption of the network, thereby reducing overall network Energy Efficiency (EE). Statistics show that densely deployed micro base stations are mostly in low traffic state and will consume 60% to 80% of the total network energy consumption in low traffic or no traffic state. A base station sleep mechanism may be employed. Namely, a proper number of low-traffic micro base stations can be shut down to reduce the static power consumption of the base stations, thereby realizing large-scale energy conservation. In addition to the large scale reduction of static power consumption, EE can be increased by improving the resource allocation method in the network. Herein, resource allocation refers to power allocation of a base station on its Resource Blocks (RBs). The power of the transmitter has a significant impact on the power consumption of the amplifier, air conditioner, etc. Therefore, the network energy consumption can be greatly reduced by intensively carrying out the sleep allocation on the base stations.
Disclosure of Invention
The invention aims to provide a base station sleep method based on a hysteresis noise chaotic neural network aiming at the defects of the prior art.
The invention adopts the following technical scheme: firstly, a heterogeneous network system model composed of a macro base station and a micro base station is established, then a hysteresis noise chaotic neural network model is constructed, an energy function of the hysteresis noise chaotic neural network model is constructed through an energy efficiency function of the heterogeneous network system model and a constraint condition, the energy function is iterated until the function is converged, and an optimal solution, namely a micro base station dormancy allocation matrix, is obtained.
In order to achieve the purpose, the invention specifically comprises the following steps:
(1) building heterogeneous cellular network system model
The heterogeneous cellular network system model is composed of a macro base station and a micro base station, the macro base station is defaulted to work in an active state all the time, and a user is preferentially connected with the micro base station, so that the energy efficiency of the heterogeneous cellular network system model is only related to the micro base station. Defining the energy efficiency function of all micro base stations in the heterogeneous cellular network system model as follows:
Figure BDA0002984589870000021
the number of the micro base stations is N, and the number of the users is M; ri,mRepresenting user umFrom micro base station BiAcquiring the data rate of each resource block; RB (radio B)i,mRepresenting user umAnd micro base station BiThe number of resource blocks required for association; x is the number ofi,mRepresents umWhether or not to react with BiAn associated binary indicator variable; p is a radical ofiA scaling factor representing the dynamic power consumption of the micro base station and the number of resource blocks allocated to its users by the micro base station; RB (radio B)iIs represented by a micro base station BiThe total number of allocated resource blocks; delta denotes the micro base station BiThe ratio of its quiescent power to its maximum operating power in the active state; pMRepresents a micro base station BiRated maximum operating power of; phi denotes the micro base station BiInfluence of whether sleep has on power consumption;
the constraint conditions of the energy efficiency functions of all the micro base stations in the heterogeneous cellular network system model are as follows:
Figure BDA0002984589870000022
constraint C1 indicates that the total actual transmit power of each micro base station on its allocated resource blocks cannot exceed an upper limit, where PmaxRepresenting the nominal transmit power of the base station.
Constraint C2 ensures that traffic of the micro base station does not reach saturation, where RBmaxRepresents the nominal transmission power of the micro base station, beta represents the proportion of resource blocks that can be allocated by the micro base station;
constraint C3 is to ensure received signal power strength for each user, where PRmRepresenting the received signal power strength of each user;
constraint C4 ensures that the interference received by each user is within the allowed range;
constraint C5 indicates that a user can only be associated with one micro base station at a time.
(2) And building a hysteresis noise chaotic neural network model and initializing parameter setting.
The parameters include: slope parameter lambda of neuron activation function, neural membrane damping factor k, coupling factor alpha between neurons, self-feedback connection weight z and positive parameter I0The noise amplitude A is uniformly distributed in [ -A, A [)]Simulated annealing rates β for random noise n (t), and z and A within the range1And beta2。
The mathematical expression of the hysteresis noise chaotic neural network model is as follows:
Figure BDA0002984589870000031
Figure BDA0002984589870000032
z(t+1)＝(1-β1)z(t) (5)
A[n(t+1)]＝(1-β2)A[n(t)] (6)
Figure BDA0002984589870000033
defining the energy function of the hysteresis noise chaotic neural network model as
Figure BDA0002984589870000034
Figure BDA0002984589870000035
Wherein x isi(t) and yi(t) represents the output and input of neuron i at time t, wijRepresenting the connection weight of neuron I to neuron j, IiRepresenting the external input bias of neuron i.
The energy function is monotonously reduced, and the energy has the characteristic of being bounded, so the hysteresis noise chaotic neural network model has convergence and stability.
(3) Substituting the energy efficiency function of the micro base station in the heterogeneous network system model and the constraint condition of the energy efficiency function into the energy function of the hysteresis noise chaotic neural network model, wherein the specific energy function is as follows:
constraint C1 defines energy E1：
Figure BDA0002984589870000041
Constraint C2 defines energy E2：
Figure BDA0002984589870000042
Constraint C3 defines energy E3：
E3＝(Puser-min-PRm)2；
Constraint C4 defines energy E4：
E4＝(γmin-SNRi,m)2；
Constraint C5 defines energy E5：
Figure BDA0002984589870000043
Target constraint representation of the energy efficiency function:
Figure BDA0002984589870000044
by combining the above analysis, the total energy function of the hysteresis noise chaotic neural network model is
E＝Q·(E1+E2+E3+E4+E5)+D·Et (10)
The parameter Q, D in the formula (10) is a undetermined parameter, and the function of the undetermined parameter is to adjust the proportion of each energy to the total energy. The power equation of the hysteresis noise chaotic neural network model is expressed as
Figure BDA0002984589870000045
And when the total energy function E of the hysteresis noise chaotic neural network model converges, the micro base station distribution meets the constraint condition, and meanwhile, the energy efficiency of the heterogeneous cellular network system model is minimized.
(4) And after the initialization parameters are calculated, delaying the output of the noise chaotic neural network model. When t is equal to 0, random noise n (0) is generated, an input signal y (0) of the hysteresis noise chaotic neural network model is randomly generated, xi (0) is equal to 0, y (0) and xi (0) are substituted into formula (5) to obtain x (0), and a power equation of the hysteresis noise chaotic neural network model is obtained through formula (11)
Figure BDA0002984589870000046
Then the result of the dynamic equation, y (0), x (0) and xi (0) are substituted into the formula (4) to obtain y (1), and finally the noise amplitude A and the self-feedback weight z are updated by the formula (5) and the formula (6);
(5) when t is 1, random noise n (1) and ξ (1) are generated, y (1) and ξ (1) are substituted for expression (3) to obtain x (1), and equation of motion (11) is used to obtain equation of motion
Figure BDA0002984589870000051
Then the dynamic equation result and y (1), x (1) and xi (1) are substituted into formula (4) to obtain y (2), and finally the noise amplitude A and the self-feedback weight z are updated by formula (5) and formula (6);
(6) when t is t +1, random noise n (t +1) is generated, y (t), y (t-1), and n (t) are substituted for expression (7) to obtain new center parameter ξ (t +1), x (t +1) is obtained by substituting expression (3), and expression (11) is used to obtain x (t +1)
Figure BDA0002984589870000052
Updating the energy value E, substituting the result into the formula (4) to obtain y (t +2), and finally updating the noise amplitude A and the self-feedback weight z by the formulas (5) and (6);
(7) if the energy function in step (6) does not converge, repeating step (6); and if the convergence is achieved, outputting a result to finish the distribution of the dormant micro base station.
The invention has the beneficial effects that: the invention utilizes the hysteresis noise chaotic neural network model to construct the energy function of the neural network according to the energy efficiency function of the base station in the heterogeneous network system model and the constraint condition of each base station, and then completes the sleep distribution to the base station according to the optimal solution of the energy function, namely the sleep distribution matrix of the base station, thereby achieving the purpose of saving the energy consumption of the heterogeneous network. Compared with the prior art, the optimal solution can be obtained through a small amount of iteration, the dormant allocation overhead of the base station is effectively reduced, and a foundation is laid for realizing lower system energy efficiency.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention. It is to be noted that the features in the following embodiments and examples may be combined with each other without conflict.
The method has the main functions that the hysteresis noise chaotic neural network model is utilized to construct the energy function of the neural network according to the energy efficiency function of the base stations in the heterogeneous network system model and the constraint condition of each base station, and then the dormancy allocation is completed for the base stations according to the optimal solution of the energy function, namely the dormancy allocation matrix of the base stations, so that the purpose of saving the energy consumption of the heterogeneous network is achieved.
See fig. 1 for details: the embodiment provides a base station dormancy method based on a hysteresis noise chaotic neural network, which comprises the following steps:
step 1: the heterogeneous cellular network system model is composed of a macro base station and a micro base station, the macro base station is defaulted to work in an active state all the time, and a user is preferentially connected with the micro base station, so that the energy efficiency of the heterogeneous cellular network system model is only related to the micro base station. Energy efficiency function eta of all micro base stations in heterogeneous cellular network system modelEEIs defined as:
Figure BDA0002984589870000061
the number of the micro base stations is N, and the number of the users is M; ri,mRepresenting user umFrom micro base station BiAcquiring the data rate of each resource block; RB (radio B)i,mRepresenting user umAnd micro base station BiThe number of resource blocks required for association; x is the number ofi,mIs a description of umWhether or not to react with BiAssociated binary indicating variables, e.g. xi,m1 represents umAnd BiAre associated with, and xi,m0 is not relevant; p is a radical ofiA scaling factor representing the dynamic power consumption of the micro base station and the number of resource blocks allocated to its users by the micro base station; RB (radio B)iIs represented by a micro base station BiThe total number of allocated resource blocks; delta denotes the micro base station BiAnd its RB in active stateiThe ratio of the maximum operating power of; pMRepresents a micro base station BiRated maximum operating power of; phi is a description of the micro base station BiThe impact of whether sleep has on power consumption. For example, when the micro base station is active, phi is 1, and when the micro base station is dormant
Figure BDA0002984589870000062
Where epsilon represents the proportion of energy consumed by a sleeping micro base station to maintain its basic management functions.
The constraint condition of the energy efficiency function of all the micro base stations in the heterogeneous cellular network system model is
Figure BDA0002984589870000063
Constraint C1 indicates that the total actual transmit power of each micro base station on its allocated resource blocks cannot exceed an upper limit, where PmaxRepresenting the nominal transmit power of the base station. Constraint C2 ensures that traffic of the micro base station does not reach saturation, where RBmaxDenotes a nominal transmission power of the micro base station, and β denotes a proportion of resource blocks that can be allocated by the micro base station. C3 is to ensure the received signal power strength of each user. PRmThe strength of the signal power received by each user is described and may be denoted as PRm＝piRBi,m-PLi,mWherein PLi,mRepresenting user umAnd micro base station BiThe path loss therebetween. C4 ensures that the interference received by each user is within the allowed range. Constraint C5 indicates that a user can only be associated with one micro base station at a time.
Step 2: and building a hysteresis noise chaotic neural network model and initializing parameter setting. The parameters include: slope parameter lambda of neuron activation function, neural membrane damping factor k, coupling factor alpha between neurons, self-feedback connection weight z of neural network, and positive parameter I0The noise amplitude A is uniformly distributed in [ -A, A [)]Simulated annealing rates β for random noise n (t), and z and A within the range1And beta2。
Defining the energy function of the hysteresis noise chaotic neural network model as
Figure BDA0002984589870000071
Figure BDA0002984589870000072
Wherein x isi(t) and yi(t) represents the output and input of neuron i at time t, wijRepresenting the connection weight of neuron I to neuron j, IiRepresenting the external input bias of neuron i.
The energy function is monotonously reduced, and the energy has the characteristic of being bounded, so the network has convergence and stability. The mathematical expression for this network is as follows:
Figure BDA0002984589870000073
Figure BDA0002984589870000074
z(t+1)＝(1-β1)z(t) (7)
A[n(t+1)]＝(1-β2)A[n(t)] (8)
Figure BDA0002984589870000081
and step 3: substituting the energy efficiency function of the base station in the heterogeneous network system model and the constraint conditions of the energy efficiency function into the energy function of the neural network, wherein the specific energy function is as follows:
constraint C1 may define energy E1Is composed of
Figure BDA0002984589870000082
Constraint C2 may define energy E2Is composed of
Figure BDA0002984589870000083
Constraint C3 may define energy E3Is E3＝(Puser-min-PRm)2(ii) a Constraint C4 may define energy E4Is E4＝(γmin-SNRi,m)2(ii) a Constraint C5 may define energy E5Is composed of
Figure BDA0002984589870000084
The target constraint of the energy efficiency function can be expressed as
Figure BDA0002984589870000085
By combining the above analysis, the total energy function of the hysteresis noise chaotic neural network model is
E＝Q·(E1+E2+E3+E4+E5)+D·Et (10)
The parameter Q, D in the formula (10) is a undetermined parameter, and the function of the undetermined parameter is to adjust the proportion of each energy to the total energy.
The power equation of the network can be expressed as
Figure BDA0002984589870000086
And when the total energy function E of the hysteresis noise chaotic neural network model converges, the micro base station distribution meets the constraint condition, and meanwhile, the energy efficiency of the heterogeneous cellular network system model is minimized.
And 4, step 4: and after the initialization parameters are calculated, delaying the output of the noise chaotic neural network model. When t is 0, a random noise n (0) is generated, an input signal y (0) of the neural network is randomly generated, ξ (0) is 0, y (0) and ξ (0) are substituted for expression (5) to obtain x (0), and a network dynamic equation is obtained by expression (11)
Figure BDA0002984589870000087
Then substituting the network dynamic equation result and y (0), x (0) and xi (0) into formula (6) to obtain y (1), and finally updating the noise amplitude A and the self-feedback weight z by formula (7) and formula (8);
and 5: when t is 1, random noise n (1) and ξ (1) are generated, y (1) and ξ (1) are substituted for expression (5) to obtain x (1), and a network dynamic equation is obtained by expression (11)
Figure BDA0002984589870000091
And then substituting the network dynamic equation result and y (1), x (1) and xi (1) into an equation (6) to obtain y (2), and finally updating the noise amplitude A and the self-feedback weight z by an equation (7) and an equation (8).
Step 6: when t is t +1, random noise n (t +1) is generated, and y (t), y (t-1), and n (t) are substituted for formula (9) to obtain a new noiseThe center parameter xi (t +1) is substituted for the formula (5) to obtain x (t +1), and the formula (11) is used to obtain x (t +1)
Figure BDA0002984589870000092
Updating the energy value E, substituting the result into the formula (6) to obtain y (t +2), and finally updating the noise amplitude A and the self-feedback weight z by the formulas (7) and (8).
And 7: if the energy function in step 6 does not converge, repeating step 6; and if the convergence is achieved, outputting a result to finish the distribution of the dormant micro base station.
In conclusion, the hysteresis noise chaotic neural network model is utilized to construct the energy function of the neural network according to the energy efficiency function of the base stations in the heterogeneous network system model and the constraint condition of each base station. Because the energy function contains the constraint conditions of the base station, the optimal solution of the energy function, namely the base station dormancy matrix, can minimize the energy consumption of the base station in the heterogeneous network under the condition of ensuring that the constraint conditions are met. The hysteresis noise chaotic neural network model is used, so that the phenomenon that the hysteresis noise chaotic neural network model is trapped in a local optimal solution can be avoided, the iteration times of the neural network are reduced, and the algorithm complexity is reduced.
It should be understood that the above description of the preferred embodiments is given for clearness of understanding and no unnecessary limitations are to be understood therefrom, and all changes and modifications that come within the spirit of the invention may be made by those skilled in the art without departing from the scope of the invention as defined by the appended claims.

Claims (1)

1. A base station dormancy method based on a hysteresis noise chaotic neural network is characterized by comprising the following steps:
(1) building heterogeneous cellular network system model
The heterogeneous cellular network system model is composed of a macro base station and a micro base station, the default macro base station always works in an active state, and a user is preferentially connected with the micro base station, so that the energy efficiency of the heterogeneous cellular network system model is only related to the micro base station; defining the energy efficiency function of all micro base stations in the heterogeneous cellular network system model as follows:
Figure FDA0002984589860000011
the number of the micro base stations is N, and the number of the users is M; ri,mRepresenting user umFrom micro base station BiAcquiring the data rate of each resource block; RB (radio B)i,mRepresenting user umAnd micro base station BiThe number of resource blocks required for association; x is the number ofi,mRepresents umWhether or not to react with BiAn associated binary indicator variable; p is a radical ofiA scaling factor representing the dynamic power consumption of the micro base station and the number of resource blocks allocated to its users by the micro base station; RB (radio B)iIs represented by a micro base station BiThe total number of allocated resource blocks; delta denotes the micro base station BiThe ratio of its quiescent power to its maximum operating power in the active state; pMRepresents a micro base station BiRated maximum operating power of; phi denotes the micro base station BiInfluence of whether sleep has on power consumption;
the constraint conditions of the energy efficiency functions of all the micro base stations in the heterogeneous cellular network system model are as follows:
Figure FDA0002984589860000012
constraint C1 indicates that the total actual transmit power of each micro base station on its allocated resource blocks cannot exceed an upper limit, where PmaxRepresents the nominal transmit power of the base station;
constraint C2 ensures that traffic of the micro base station does not reach saturation, where RBmaxRepresents the nominal transmission power of the micro base station, beta represents the proportion of resource blocks that can be allocated by the micro base station;
constraint C3 is to ensure received signal power strength for each user, where PRmRepresenting the received signal power strength of each user;
constraint C4 ensures that the interference received by each user is within the allowed range;
constraint C5 indicates that a user can only be associated with one micro base station at a time;
(2) building a hysteresis noise chaotic neural network model and initializing parameter setting;
the parameters include: slope parameter lambda of neuron activation function, neural membrane damping factor k, coupling factor alpha between neurons, self-feedback connection weight z and positive parameter I0The noise amplitude A is uniformly distributed in [ -A, A [)]Simulated annealing rates β for random noise n (t), and z and A within the range1And beta2；
The mathematical expression of the hysteresis noise chaotic neural network model is as follows:
Figure FDA0002984589860000021
Figure FDA0002984589860000022
z(t+1)＝(1-β1)z(t) (5)
A[n(t+1)]＝(1-β2)A[n(t)] (6)
Figure FDA0002984589860000023
defining the energy function of the hysteresis noise chaotic neural network model as
Figure FDA0002984589860000024
Figure FDA0002984589860000025
Wherein x isi(t) and yi(t) respectively represents the output and input of neuron i at time t,wijrepresenting the connection weight of neuron I to neuron j, IiRepresents the external input bias of neuron i;
the energy function is monotonously reduced, and the energy has the characteristic of being bounded, so the hysteresis noise chaotic neural network model has convergence and stability;
(3) substituting the energy efficiency function of the micro base station in the heterogeneous network system model and the constraint condition of the energy efficiency function into the energy function of the hysteresis noise chaotic neural network model, wherein the specific energy function is as follows:
constraint C1 defines energy E1：
Figure FDA0002984589860000031
Constraint C2 defines energy E2：
Figure FDA0002984589860000032
Constraint C3 defines energy E3：
E3＝(Puser-min-PRm)2；
Constraint C4 defines energy E4：
E4＝(γmin-SNRi,m)2；
Constraint C5 defines energy E5：
Figure FDA0002984589860000033
Target constraint representation of the energy efficiency function:
Figure FDA0002984589860000034
by combining the above analysis, the total energy function of the hysteresis noise chaotic neural network model is
E＝Q·(E1+E2+E3+E4+E5)+D·Et (10)
The parameter Q, D in the formula (10) is a undetermined parameter, and the undetermined parameter has the function of adjusting the proportion of each energy to the total energy; the power equation of the hysteresis noise chaotic neural network model is expressed as
Figure FDA0002984589860000035
When the total energy function E of the hysteresis noise chaotic neural network model converges, the micro base station distribution meets the constraint condition, and meanwhile, the energy efficiency of the heterogeneous cellular network system model is minimized;
(4) after the initialization parameters are calculated, delaying the output of the noise chaotic neural network model; when t is equal to 0, random noise n (0) is generated, an input signal y (0) of the hysteresis noise chaotic neural network model is randomly generated, xi (0) is equal to 0, y (0) and xi (0) are substituted into formula (5) to obtain x (0), and a power equation of the hysteresis noise chaotic neural network model is obtained through formula (11)
Figure FDA0002984589860000041
Then the result of the dynamic equation, y (0), x (0) and xi (0) are substituted into the formula (4) to obtain y (1), and finally the noise amplitude A and the self-feedback weight z are updated by the formula (5) and the formula (6);
(5) when t is 1, random noise n (1) and ξ (1) are generated, y (1) and ξ (1) are substituted for expression (3) to obtain x (1), and equation of motion (11) is used to obtain equation of motion
Figure FDA0002984589860000042
Then substituting the dynamic equation result and y (1), x (1) and x (1) into an equation (4) to obtain y (2), and finally updating the noise amplitude A and the self-feedback weight z by an equation (5) and an equation (6);
(6) when t is t +1, random noise n (t +1) is generated, new central parameter x (t +1) is obtained by substituting y (t), y (t-1) and n (t) into formula (7), x (t +1) is obtained by substituting formula (3), and formula (11) is used to obtain
Figure FDA0002984589860000043
Updating the energy value E, substituting the result into the formula (4) to obtain y (t +2), and finally updating the noise amplitude A and the self-feedback weight z by the formulas (5) and (6);
(7) if the energy function in step (6) does not converge, repeating step (6); and if the convergence is achieved, outputting a result to finish the distribution of the dormant micro base station.
CN202110296601.1A 2021-03-19 2021-03-19 Base station dormancy method based on hysteresis noise chaotic neural network Active CN113099520B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110296601.1A CN113099520B (en) 2021-03-19 2021-03-19 Base station dormancy method based on hysteresis noise chaotic neural network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110296601.1A CN113099520B (en) 2021-03-19 2021-03-19 Base station dormancy method based on hysteresis noise chaotic neural network

Publications (2)

Publication Number Publication Date
CN113099520A true CN113099520A (en) 2021-07-09
CN113099520B CN113099520B (en) 2022-10-11

Family

ID=76668455

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110296601.1A Active CN113099520B (en) 2021-03-19 2021-03-19 Base station dormancy method based on hysteresis noise chaotic neural network

Country Status (1)

Country Link
CN (1) CN113099520B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114867090A (en) * 2022-05-06 2022-08-05 东南大学 UDN base station dormancy and power distribution joint optimization method based on LSTM unsupervised learning

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5613042A (en) * 1994-02-14 1997-03-18 Gold Star Electron Co., Ltd. Chaotic recurrent neural network and learning method therefor
CN105050170A (en) * 2015-06-30 2015-11-11 东南大学 Base station dormancy method based on flow prediction in heterogeneous network
CN108574653A (en) * 2018-04-19 2018-09-25 南京邮电大学 Signal blind checking method based on double Sigmoid sluggishnesses noise chaotic neural networks
CN111246552A (en) * 2020-01-14 2020-06-05 南京工业大学 Base station dormancy method based on mobile network flow prediction
CN111343704A (en) * 2020-02-29 2020-06-26 重庆邮电大学 Combined dormancy and power control method for femto base station in heterogeneous cellular network

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5613042A (en) * 1994-02-14 1997-03-18 Gold Star Electron Co., Ltd. Chaotic recurrent neural network and learning method therefor
CN105050170A (en) * 2015-06-30 2015-11-11 东南大学 Base station dormancy method based on flow prediction in heterogeneous network
CN108574653A (en) * 2018-04-19 2018-09-25 南京邮电大学 Signal blind checking method based on double Sigmoid sluggishnesses noise chaotic neural networks
CN111246552A (en) * 2020-01-14 2020-06-05 南京工业大学 Base station dormancy method based on mobile network flow prediction
CN111343704A (en) * 2020-02-29 2020-06-26 重庆邮电大学 Combined dormancy and power control method for femto base station in heterogeneous cellular network

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
孙明等: "基于迟滞噪声混沌神经网络的TDMA广播调度", 《哈尔滨商业大学学报(自然科学版)》 *
崔冬等: "多通道神经群模型建模及分析", 《中国科学:信息科学》 *
邵凯等: "基于迟滞噪声混沌神经网络的导频分配", 《电子与信息学报》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114867090A (en) * 2022-05-06 2022-08-05 东南大学 UDN base station dormancy and power distribution joint optimization method based on LSTM unsupervised learning
CN114867090B (en) * 2022-05-06 2024-04-02 东南大学 UDN base station dormancy and power distribution combined optimization method based on LSTM unsupervised learning

Also Published As

Publication number Publication date
CN113099520B (en) 2022-10-11

Similar Documents

Publication Publication Date Title
CN109729528B (en) D2D resource allocation method based on multi-agent deep reinforcement learning
CN108112082B (en) Wireless network distributed autonomous resource allocation method based on stateless Q learning
Ram et al. Distributed non-autonomous power control through distributed convex optimization
CN110233755B (en) Computing resource and frequency spectrum resource allocation method for fog computing in Internet of things
CN111343704B (en) Combined dormancy and power control method for femto base station in heterogeneous cellular network
US10080200B2 (en) Intelligent deployment cascade control device based on an FDD-OFDMA indoor small cell in multi-user and interference environments
CN110519849B (en) Communication and computing resource joint allocation method for mobile edge computing
CN111446992B (en) Method for allocating resources with maximized minimum energy efficiency in wireless power supply large-scale MIMO network
Singh et al. An energy efficient approach for clustering in WSN using fuzzy logic
Li et al. User association and power allocation based on Q-learning in ultra dense heterogeneous networks
WO2021120962A1 (en) Hierarchical multiple access method for environmental backscatter network
CN113795049B (en) Femtocell heterogeneous network power self-adaptive optimization method based on deep reinforcement learning
Kim Adaptive online power control scheme based on the evolutionary game theory
CN113099520B (en) Base station dormancy method based on hysteresis noise chaotic neural network
Zhao et al. Robust power control for cognitive radio in spectrum underlay networks
Li et al. Parallel deep reinforcement learning based online user association optimization in heterogeneous networks
CN114126021B (en) Power distribution method of green cognitive radio based on deep reinforcement learning
CN102665219A (en) Dynamic frequency spectrum allocation method of home base station system based on OFDMA
CN105490794A (en) Packet-based resource distribution method for orthogonal frequency division multiple access (OFDMA) femtocell double-layer network
CN107948109B (en) Subcarrier bit joint optimization method for compromising energy efficiency and spectral efficiency in cognitive radio
CN106973440A (en) Time towards wireless power network distributes optimization method
CN106604381B (en) Millimeter wave network performance analysis method based on three-level transmission power mechanism
CN113301637A (en) D2D communication power control algorithm based on Q learning and neural network
CN107071876B (en) Wireless sensor network power control method adopting two-stage fuzzy controller
CN114529207B (en) Energy storage battery distributed economic dispatching method based on differential privacy mechanism

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