CN106331083B - A kind of heterogeneous network selection method considering content distribution energy consumption - Google Patents
A kind of heterogeneous network selection method considering content distribution energy consumption Download PDFInfo
- Publication number
- CN106331083B CN106331083B CN201610695982.XA CN201610695982A CN106331083B CN 106331083 B CN106331083 B CN 106331083B CN 201610695982 A CN201610695982 A CN 201610695982A CN 106331083 B CN106331083 B CN 106331083B
- Authority
- CN
- China
- Prior art keywords
- content
- base station
- user
- particle
- small
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/56—Provisioning of proxy services
- H04L67/568—Storing data temporarily at an intermediate stage, e.g. caching
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/56—Provisioning of proxy services
- H04L67/568—Storing data temporarily at an intermediate stage, e.g. caching
- H04L67/5682—Policies or rules for updating, deleting or replacing the stored data
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W48/00—Access restriction; Network selection; Access point selection
- H04W48/20—Selecting an access point
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Security & Cryptography (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
A kind of heterogeneous network selection method for considering content distribution energy consumption of the present invention, belongs to mobile communication field；Specifically: firstly, establishing the communication system between macro base station, small-cell base station and user；Then, each user sends respective request content simultaneously, and communication system uses quanta particle swarm optimization, calculates when system total energy consumption minimum, the content of base station selected user and node B cache, and respective connection relationship and content caching corresponding relationship are broadcasted to each small-cell base station and user；Finally, small-cell base station whether there is in itself according to the request content judgement of each selection user, if so, user directly obtains cache contents from the small-cell base station；Otherwise, user obtains content by wireless backhaul link from content supplier.Advantage is: reducing the complexity of selection mechanism realization, reaches the optimal solution of content caching strategy and user access network selection in acceptable time range, enable the system to consumption and reach minimum.
Description
Technical field
The invention belongs to mobile communication fields, describe a kind of heterogeneous network selection method for considering content distribution energy consumption.
Background technique
It is predicted according to Cisco's white paper, the 75% of global metadata total flow will be accounted in mobile data stream in 2019, due to
Smart machine is popularized, and the growth trend of gusher formula will be presented in the following mobile data flow.Present macrocellular network capacity
Further expansion has been unable to meet growing flow demand；In order to solve this problem, cell is introduced
(Small Cell) network, under the covering of macrocellular network, largely dispose FemtoCell, PicoCell, MicroCell,
WiFi AP etc., to constitute heterogeneous network.By the dense deployment of small-cell base station, the capacity of system is greatly improved, is delayed
User has been solved to the flow demand of network.But the dense deployment of small-cell base station aggravates the serious interference of minizone, and
The operation of extra base station brings additional energy consumption；The need for being also no longer satisfied network by merely disposing small-cell base station
It asks.
Discovery when being counted by the particular content for including to data traffic in network, the content of user's request are mostly weights
Multiple, the frequent requests and output transmission of such a large amount of duplicate contents have carried out data error rate promotion, time delay increases, time delay is trembled
More violent etc. drawbacks are moved, declines the service experience of user greatly, causes network congestion and the decline of overall performance of network.
And the content distributing network (Content Delivery Network, CDN) based on caching technology is then to improve this
One situation provides an ideal scheme.Caching is installed additional by the base station into heterogeneous network, and more frequently by user's request
Numerous content is first stored in caching, can preferably promote the service experience of user, such as: time delay is lower, delay variation is more stable,
The bigger fairness between user of content selection degree is more preferable.
At the same time, the high speed development of communication industry is not concerned only with industry effectively to meet the needs of users, it is also contemplated that
The social responsibility and ecological requirements of communication industry.This demand is embodied a concentrated reflection of such as how lower energy consumption processing and in time height
Effect ground transmission information, preferably to meet future communications green, environmental protection, energy-efficient demand.
In existing content distribution mechanism, since the repetition of a large amount of contents is transmitted, lead to the bandwidth of wireless backhaul link
The very big consumption of resource significant wastage and transimission power.And content caching measure is taken in CDN, it can satisfy the following net just
The energy saving requirement of network.By storing the interested content of user into the caching of base station, it is lower that operator can use power consumption
Storage equipment rather than the higher communication equipment of power consumption meets user demand, energy consumption can also further decrease.Simultaneously by content
The caching for being issued to base station can also reduce the data transmission times of wireless backhaul link, and the access point for requesting user more collects
In, reduce extra access point operation bring extra power consumption, therefore the CDN scheme bring energy consumption based on caching is saved very
Obviously.
Although there are some research achievements for the content distribution scenario based on energy consumption in CDN, such as: in heterogeneous network
In network Selecting research, document " a kind of cell selection strategy being limited based on wireless backhaul link ", the strategy is in order to overcome backhaul
The limitation of link size and the experience for improving user joined caching for access point, shifted to an earlier date by the analysis to content popularit
The high content of deposit popularity in a base station.
But this method only loosely gives the content distribution situation based on content popularit, and there is no examine well
Consider the instant content requirements of user.The high content of the popularity of buffered in advance might not be exactly active user's request in base station
The content hit rate of content, i.e. user request may be very low, will lead to wireless backhaul link bandwidth resources expense, Yi Jiji in this way
The further expense for caching energy consumption of standing.
Further, since the energy consumption of caching is really lower, most of work can selectively ignore caching energy consumption, existing technology
The scheme overwhelming majority seldom considers system from considering that system entire throughput, user's time delay, user fairness etc. are optimization aim
Total energy consumption optimization, the requirement of future communication technologies green energy conservation cannot be met well.
Summary of the invention
The present invention generally shifts to an earlier date according to the popularity distribution situation of content in base station in existing content distribution scenario
The high content of middle caching popularity proposes a kind of consideration content there is no this problem of the instant content requirements of user is considered
Distribute the heterogeneous network selection method of energy consumption, this method carries out the net of user under the premise of considering content caching energy consumption strategy
Network access selection, to reduce the energy consumption of content distribution.
Specific step is as follows:
Step 1: being directed to the downlink of some cellular network, establish logical between macro base station, small-cell base station and user
Letter system；
It include: that a macro base station covers K small-cell base station in communication system, collection is combined into SBSk, k ∈ 1,2,3,
4......K}；I user's collection is combined into UEi, i ∈ { 1,2,3,4......I }；J properties collection is Cj, j ∈ 1,2,3,
4......J}；
Each user is respectively connected at any time on a small-cell base station, uses ai,kConnection relationship is indicated, if user
I is connected on small-cell base station k, then ai,k=1, otherwise ai,k=0；
Each content size is the same, and the same content can be cached to simultaneously on different small-cell base stations, usesIndicate slow
Relationship is deposited, if content j is cached on small-cell base station k,OtherwiseMeet
Different user can request the same content simultaneously, according to customer requirement retrieval user's request content；User and content
Incidence relation useIt indicates, if user i request content j,Otherwise
Step 2: each user sends to each small-cell base station connecting with itself respective simultaneously in some period
Request content；
Step 3: communication system uses quanta particle swarm optimization, the base station selected use when system total energy consumption minimum is calculated
The optimal solution of the content of family and node B cache；
Specific step is as follows:
Step 301, initialization algorithm parameter；
Initialization: particle position, total number of particles and maximum number of iterations；
Each particle position includes two parts information: the result of I user's access base station and J content caching are placed
As a result；
Total number of particles is set as M；Maximum number of iterations is set as S；
Step 302 is directed to each base station and each content, initializes the initial value and use of node B cache location of content respectively
The initial value of family connection base station；
For small-cell base station k, M group cached parameters vector is randomly selectedAs node B cache
The initial value of location of content；zm,kMeet node B cache length limitation and user accesses number limitation.
For user i, chooses M group and access vector am,i=[ai,1′,ai,2′,...,ai,k'] as user's connection base station
Initial value；
Step 303, the initial value for connecting base station with I user using the initial value of K node B cache location of content, setting
The initial position of each particle；
The initial position of each particle is exactly the position of each particle of the number of iterations s=0；
The initial position X of m-th of particlem(0):
Xm(0)=[am,1(0),am,2(0),...,am,I(0),zm,1(0),zm,2(0),...,zm,K(0)]
To the position in the s times iteration of m-th of particle are as follows:
Xm(s)=[am,1(s),am,2(s),...,am,I(s),zm,1(s),zm,2(s),...,zm,K(s)]
Step 304, sometime, calculate the corresponding system total energy consumption value of the particle using the position of each particle；
Each particle position is corresponding to calculate a system total energy consumption value；
pi,kIndicate small-cell base station SBSkTo user UEiTransmission power,Indicate content CjIt is buffered in small-cell base station
SBSkOn caching power, pk,j,cIndicate content CjSmall-cell base station is transferred to by wireless backhaul link from content supplier
SBSkOn transimission power.
Step 305, in subsequent time, update the position and speed of each particle；
The next position of particle is codetermined by the current location of particle and present speed:
More new formula is as follows:
Wherein, γ is the random number that value range is (0,1), and β represents shrinkage expansion coefficient, and μ is that value range is (0,1)
Random number；C (s) indicates the average desired positions in the s times iterative process,Pm(s) it indicates m-th
Desired positions of the particle in the s times iterative process；It calculates as follows:
As s=0, Pm(0)=Xm(0)；F[Pm(s-1)] indicate that m-th of particle is best in the s-1 times iterative process
The fitness function of position；F[Xm(s)] fitness function of position of m-th of particle in the s times iteration is indicated.
Vector P expression formula is as follows:
It is the random number between 0 to 1；G (s) indicates global desired positions of all particles in the s times iterative process:
As s=0, from Pm(0) select a desired positions as overall situation desired positions G (0) in.
Behind step 306, every position for updating primary particle, the fitness function F (X of each particle is calculatedm(s)) it, and records
The correspondence penalty value of the fitness function value of whole M particles and M particle
Fitness function F (XmIt (s)) include two parts: total energy consumption function and penalty；
Its expression formula are as follows:
Wherein α is penalty factor, determines specific influence of the value on fitness function of penalty；
The expression formula of penalty is as follows:
sk(k ∈ { 1,2,3......K }) indicates small-cell base station SBSkBuffer memory capacity size, meet: be cached to thereon
Content total number
lk(k ∈ { 1,2,3......K }) indicates small-cell base station SBSkLoad capacity size, meet: be linked into thereon
User total number
Step 307, the optimum position P that each particle is updated according to the fitness function of M particlem(s), and M grain is recorded
Overall situation optimum position G (s) optimal location G corresponding with true energy consumption minimum in sonr(s)；
It is as follows to update rule:
If F (Xm(s)) < F (Xm(s-1)), then Pm(s)=Xm(s), otherwise Pm(s)=Pm(s-1)。
If F (Pm(s)) < F (G (s)), then G (s)=Pm(s), otherwise G (s)=G (s-1).
If F (Pm(s)) < F (Gr(s)) andThen Gr(s)=Pm(s), otherwise Gr(s)=Gr(s-
1)。
Step 308 judges whether the number of iterations s reaches maximum number of iterations S, if so, optimal position in M particle of output
Set Gr(s) corresponding solutionEnter step 309；Otherwise, return step 305.
Step 309, according to optimal location Gr(s) optimal solution X*, obtain user and base base station selected in communication system
Stand caching content；
Optimal solution X*It include: the content optimal solution of base station selected user's optimal solution and node B cache, specifically:
Step 4: according to base station selected user and the content optimal solution of node B cache, to each small-cell base station and use
Broadcast respective connection relationship and content caching corresponding relationship in family.
Step 5: being used for selected handy family and cached some small-cell base station of content according to each selection
The request at family judges that base station itself whether there is request content, caches if it does, user directly obtains from the small-cell base station
Content；Otherwise, user obtains content by wireless backhaul link from content supplier.
The present invention has the advantages that
A kind of heterogeneous network selection method considering content distribution energy consumption, by with specific user content substitutable demand base
In the requirement forecasting of the content popularit of random distribution, realizes and be based on energy-efficient joint content caching strategy and user access network
The scheme of selection makes full use of QPSO algorithm to solve the characteristic of np problem, i.e., algorithm parameter is few and realizes simply, reduces selection
The complexity that mechanism is realized；To guarantee proposed mechanism, reaches content caching strategy in acceptable time range and user connects
The optimal solution for entering network selection enables the system to consumption and reaches minimum.
Detailed description of the invention
Fig. 1 is the communication system schematic diagram between macro base station, small-cell base station and user that the present invention establishes；
Fig. 2 is a kind of heterogeneous network selection method flow chart for considering content distribution energy consumption of the present invention；
Fig. 3 method flow diagram of base station selected user and cache contents when being computing system total energy consumption minimum of the present invention；
Fig. 4 is the relational graph of present invention number of users and system total energy consumption under three kinds of algorithms；
Fig. 5 is number of users and system total energy consumption relational graph in the case of different base station cache size of the present invention；
Fig. 6 is number of users and system total energy consumption relational graph in the case of different cache policies of the invention.
Specific embodiment
Below in conjunction with attached drawing, the present invention is described in further detail.
For in heterogeneous network selection scheme, user rate is not only limited by forward link rate the present invention, also by
The limitation of backward wireless backhaul link rate.By the way that caching technology is introduced heterogeneous network, content caching is carried out in access point, it can
To reduce redundant data transmissions, slows down wireless backhaul link load, reduce backhaul link energy consumption.Therefore, the invention proposes one
Kind considers the heterogeneous network selection method of content distribution energy consumption, considers that the real time content of user in current system requests situation, base
The limitation of loading condition and cache size and the limitation of wireless backhaul link stood, propose the content more than user's request number of times
Before be cached in base station, and make user concentrate be linked into the base station: first according to the instant content need of users all in system
It pleads condition, determines whether content needs buffered in advance into base station and cache location selection, then being needed with identical content
All users asked concentrate and are linked into the base station for having required content；To reduce the load and energy consumption of wireless backhaul link
Expense enables the system to consumption and minimizes.
Specific step is as follows:
Step 1: being directed to the downlink of some cellular network, establish logical between macro base station, small-cell base station and user
Letter system；
As shown in Figure 1, considering that a macro base station covers the downlink of the cellular network of K small-cell base station, communication system
It include I user, J content in system.
K small-cell base station, collection are combined into SBSk, k ∈ { 1,2,3,4......K }；I user's collection is combined into UEi, i ∈ 1,2,
3,4......I}；J properties collection is Cj, j ∈ { 1,2,3,4......J }；
Assuming that each user must connect at any time and at most can only connect on a small-cell base station, a is usedi,k
It indicates connection relationship, defines incidence matrix A, characterize the incidence relation of user and small-cell base station:
If user i is connected on small-cell base station k, ai,k=1, otherwise ai,k=0；
Assuming that each content size is the same, the same content can be cached to simultaneously on different small-cell base stations, be usedTable
Show caching relation, define incidence matrix Z, characterizes the incidence relation of content and small-cell base station:
If content j is cached on small-cell base station k,OtherwiseMeet
It is thought of as content CjThe one or more optimal small-cell base stations of selection are stored in advance, so that small-cell base station
The content C of upper storagejMore users are able to satisfy to content CjDemand；Under the premise of herein, then optimal access is selected for user
Point allows more users to directly acquire content C from the caching of small-cell base stationj, from reduce from content supplier to
The wireless backhaul link of small-cell base station loads and energy expense.
Different user can request the same content simultaneously, and known users demand, i.e. user are requested in some
Appearance is known；The incidence relation of user and content is usedIt indicates, define incidence matrix B, characterization user is associated with request content
Relationship:
If user i request content j,Otherwise
Step 2: each user sends to each small-cell base station connecting with itself respective simultaneously in some period
Request content；
Base station decides whether to cache content according to the request situation of all users, and it is minimum to reach system total energy consumption；
Meanwhile access base station is selected for user according to the caching situation of content, so that system total energy consumption is minimum.
User UEiTo system request content CjIf content CjIt can be in UEiThe small-cell base station SBS accessedkUpper acquisition is (small
Cell base station SBSkShift to an earlier date from content supplier CP and has obtained simultaneously cache contents Cj), then content CjDirectly from small-cell base station
SBSkIt is sent to user UEi；If content CjIt cannot be from small-cell base station SBSkUpper acquisition, then content CjFirst pass through wireless backhaul link
Small-cell base station SBS is transferred to from content supplier CPk, then from small-cell base station SBSkIt is transferred to user UEi。
Step 3: communication system uses quanta particle swarm optimization, the base station selected use when system total energy consumption minimum is calculated
The optimal solution of the content of family and node B cache；
Joint content caching position Placement Strategy based on quantum behavior particle swarm optimization algorithm is selected with user access network
The flow chart of optimization problem is selected, as shown in Figure 3, the specific steps are as follows:
Step 301, initialization algorithm parameter；
Initialization: the position X of each particlem(0), total number of particles M and maximum number of iterations S；
In order to which quantum behavior particle swarm optimization algorithm to be applied to established content caching position Placement Strategy and user
Access network selection joint optimization problem in, by the access base station selection result comprising all I users and J content
Caching places result and combines the position for being defined as a particle.
Assuming that a total of M particle, the position vector X of m-th of particlem, m=1,2..., M expression are as follows:
It can be seen that multi-C vector XmElement be made of multiple variables, element is divided into two parts.First part is from the 1st
I K dimension is tieed up, indicates the connection of I user and K base station.Second part is tieed up to I K+JK from I K+1 and is tieed up, and is indicated
The caching for being J content on K base station places situation.Namely each particle position includes two parts information: I user
The result that the result of access base station and J content caching are placed；
The position of each particle is initialized, i.e., when content caching position and user network select, it is necessary to assure each particle
It is initially in the range of feasible solution, while P is setm(0)=Xm(0)。
Step 302 is directed to each base station and each content, initializes the initial value and use of node B cache location of content respectively
The initial value of family connection base station；
For small-cell base station k, M group cached parameters vector is randomly selectedAs node B cache
The initial value of location of content；zm,kMeet node B cache length limitation and user accesses number limitation.
For user i, chooses M group and access vector am,i=[ai,1′,ai,2′,...,ai,k'] as user's connection base station
Initial value；
Step 303, the initial value for connecting base station with I user using the initial value of K node B cache location of content, setting
The initial position of each particle；
The initial position of each particle is exactly the position of each particle of the number of iterations s=0；
The initial position X of m-th of particlem(0):
Xm(0)=[am,1(0),am,2(0),...,am,I(0),zm,1(0),zm,2(0),...,zm,K(0)]
To the position in the s times iteration of m-th of particle are as follows:
Xm(s)=[am,1(s),am,2(s),...,am,I(s),zm,1(s),zm,2(s),...,zm,K(s)]
Step 304, sometime, calculate the corresponding system total energy consumption value of the particle using the position of each particle；
The total energy consumption of system is expressed as the forward link power consumption, wireless backhaul link power consumption and base station of all users in system
Cache the sum of power consumption.While guaranteeing system total energy consumption minimum, the caching capabilities for considering small-cell base station, cell are taken into account
The load capacity of base station, wireless backhaul link limitation.
Each particle position is corresponding to calculate a power consumption values；
pi,kIndicate small-cell base station SBSkTo user UEiTransmission power,Indicate content CjIt is buffered in small-cell base station
SBSkOn caching power, pk,j,cIndicate content CjBy wirelessly returning from content supplier (Content Provider, CP)
Journey link transmission is to small-cell base station SBSkOn transimission power.
In order to make system total energy consumption minimum, i.e. optimization aim are as follows:
System total energy consumption is restricted by the following conditions:
sk(k ∈ { 1,2,3......K }) indicates small-cell base station SBSkBuffer memory capacity size, be defined as being cached to thereon
Content total number
lk(k ∈ { 1,2,3......K }) indicates small-cell base station SBSkLoad capacity size, be defined as being linked into thereon
User total number
C1Indicate that each user must be linked into one and at most only have access under a small-cell base station；C2In expression
The caching of appearance is limited by small-cell base station caching capabilities, though user to it is a certain it is interior have demand, if but small-cell base station
Caching has expired the cache request for just not receiving content；C3Indicate that the access of user is limited by small-cell base station load capacity；C4
Characterize user UEiWith small-cell base station SBSkRelevance, work as ai,kUser UE is indicated when=1iIt is linked into small-cell base station SBSk
On, work as ai,kIndicate that the two is not associated with when=0.C5Characterize content CjWith small-cell base station SBSkRelevance, whenWhen table
Show content CjIt is buffered in small-cell base station SBSkOn, whenWhen indicate content CjIt is not buffered in small-cell base station SBSkOn；C6Table
Take over family UE for useiWith content CjRelevance, whenWhen indicate user UEiRequest content delays Cj, whenWhen indicate user UEi
Not request content Cj。
In view of optimization problem is Integral nonlinear program-ming problem, belong to nondeterministic polynomial difficulty (Non-
Deterministic Polynomial-hard, NP-hard) problem, although can be solved with the method for exhaustion, calculating is opened
It sells very big, it is common practice to which approximate solution is carried out using heuritic approach.In order to reduce computational complexity, it is excellent to introduce population
Change algorithm to be solved.Since quantum behavior particle swarm optimization algorithm is improved on the basis of particle swarm optimization algorithm,
Therefore it is considered as quantum behavior particle swarm optimization algorithm to solve Integral nonlinear program-ming problem, obtains suboptimal solution.
Step 305, in subsequent time, update the position and speed of each particle；
The next position of particle is codetermined by the current location of particle and present speed, by constantly updating particle position
It sets, that is, the cache location of content and the network of user select, and calculate corresponding system total energy consumption.
The number of iterations s, for each particle from 1 to M, executes following steps, until the number of iterations reaches S since 0
It is secondary:
More new formula is as follows:
Wherein, γ is the random number that value range is (0,1), and β represents shrinkage expansion coefficient, and μ is that value range is (0,1)
Random number；C (s) indicates the average desired positions in the s times iterative process,Pm(s) it indicates m-th
Desired positions of the particle in the s times iterative process；It calculates as follows:
As s=0, Pm(0)=Xm(0)；F[Pm(s-1)] indicate that m-th of particle is best in the s-1 times iterative process
The fitness function of position；F[Xm(s)] fitness function of position of m-th of particle in the s times iteration is indicated.
Vector P expression formula is as follows:
It is the random number between 0 to 1；G (s) indicates global desired positions of all particles in the s times iterative process,
It is according to fitness function F (Xm(s)), from the desired positions P of all particlesm(s) it selects to obtain G (s) in, it may be assumed that
G (s)=Pξ(s)
As s=0, from Pm(0) select a desired positions as overall situation desired positions G (0) in；I.e. global desired positions
It is the position of the smallest particle of system total energy consumption corresponding to the particle after initializing.
Behind step 306, every position for updating primary particle, the fitness function F (X of each particle is calculatedm(s)) it, and records
The correspondence penalty value of the fitness function value of whole M particles and M particle
Fitness function F (Xm(s)) include two parts, i.e. the particle connection relationship that is included and caching relation determine it is total
Energy consumption function, and the penalty to express restrictive condition；
Penalty is mainly used to judge after particle position updates either with or without restrictive condition is met, because particle is each
The secondary obtained solution of location updating is not necessarily feasible solution, and particle is limited with this and updates the solution obtained behind position in feasible solution
In range, it that is to say and meet C1-C6This six restrictive conditions.
Firstly, original constrained optimization problem to be converted to the form of unconfined condition according to penalty functional method.This
Sample obtains the fitness function comprising an objective function three classes constraint condition, expression formula are as follows:
Wherein α is penalty factor, determines specific influence of the value on fitness function of penalty；
PenaltyExpression formula it is as follows:
Penalty includes six parts, corresponds respectively to three classes constraint condition.
The first item and Section 2 of penalty are an inequality, are first class constraint condition；Max () expression takes
Biggish one in two numbers.
Section 3, Section 4 and the Section 5 of penalty are converted into an equation form, are second class constraint condition,
The Section 6 of penalty is an equation form, is third class constraint condition,
Fitness function:
Step 307, the optimum position P that each particle is updated according to the fitness function of M particlem(s), and M grain is recorded
Overall situation optimum position G (s) optimal location G corresponding with true energy consumption minimum in sonr(s)；
Specific update rule is as follows:
If F (Xm(s)) < F (Xm(s-1)), then Pm(s)=Xm(s), otherwise Pm(s)=Pm(s-1)。
If F (Pm(s)) < F (G (s)), then G (s)=Pm(s), otherwise G (s)=G (s-1).
If F (Pm(s)) < F (Gr(s)) andThen Gr(s)=Pm(s), otherwise Gr(s)=Gr(s-
1)。
Step 308 judges whether the number of iterations s reaches maximum number of iterations S, if so, optimal position in M particle of output
Set Gr(s) corresponding solutionEnter step 309；Otherwise, return step 305.
Step 309, according to optimal location Gr(s) optimal solution X*, obtain user and base base station selected in communication system
Stand caching content；
The cache location with user network for comparing content under different distribution condition select corresponding system total energy consumption, really
Determine optimal solution, i.e., optimal content caching position and user network select and minimum system power consumption values.According to fitness function, calculate
The corresponding adaptive value at global desired positions, and obtained result is exported.
Optimal solution X*It include: the content optimal solution of base station selected user's optimal solution and node B cache, specifically:
Step 4: according to base station selected user and the content optimal solution of node B cache, to each small-cell base station and use
Broadcast respective connection relationship and content caching corresponding relationship in family.
Step 5: being used for selected handy family and cached some small-cell base station of content according to each selection
The request content at family judges that base station itself whether there is, if it does, user directly obtains cache contents from small-cell base station；It is no
Then, user obtains content by wireless backhaul link from content supplier.
Simulating scenes are set as the common artificial network configuration of heterogeneous network；5 cell bases are provided in communication system
It stands, the load capacity size of each base station is 10, and cache size ability is 3；Around base station random distribution user；System
In total content number be 20, wherein the size of each content is the same.Detailed simulation parameter is as shown in table 1:
Table 1
| Parameter | Value |
| Wireless backhaul link power | 26dBm(0.398w) |
| The transimission power of base station | 26dBm(0.398w) |
| Node B cache power | 5dBm(0.003w) |
The application mainly the influence from the increase of user to system total energy consumption, different base station cache size to system total energy consumption
Influence, the different influence of cache policy (cache policy and Zipf distributed buffer strategy in institute's climbing form type) to system total energy consumption it is several
Aspect carries out simulation analysis.In addition, using classical particle colony optimization algorithm to embody the performance of the strategy of proposition
(PSO) and thoroughly method (ENUM) both algorithms are searched, is compared with QPSO algorithm of the present invention.
As shown in figure 4, method is searched by comparison QPSO algorithm, PSO algorithm and thoroughly, it is fixed not in number of base stations and content number
When change, the relationship of number of users and system total energy consumption is considered.As can be seen that with the increase of number of users, system total energy consumption exists
Rise.This is because number of users increases in system, the content for needing to transmit also is increasing, so system energy consumption is also increasing.
In addition, the system total energy consumption under QPSO algorithm is lower than the total energy consumption under PSO algorithm, reason is that PSO algorithm is easily trapped into part
Optimal, QPSO algorithm can overcome deficiency existing for PSO algorithm, available overall situation suboptimal solution.In addition, by being obtained with the poor method of searching
To optimal solution compare, the suboptimal solution that QPSO and PSO is obtained does not differ greatly, but has on algorithm complexity apparent
It improves.It can be seen that the present invention is based on the content caching positions of quantum behavior particle swarm optimization algorithm to place the access with user
Network selection, performance have a distinct increment.
As shown in figure 5, comparing user under the cache size of different base station when number of base stations and content number immobilize
The relationship of number and system total energy consumption.As can be seen that with the increase of number of users, system total energy consumption is rising.This is because
Number of users increases in system, and the content for needing to transmit also is increasing, so system energy consumption is also increasing.In addition, node B cache
Size is the curve of 0 (i.e. base station does not have caching capabilities), the curve that node B cache size is 3 curves, node B cache size is 5
Position successively reduces.Reason is the increase with node B cache size, and the content number that base station can cache also increases, so more
More users directly can obtain content from base station and be requested without being obtained from content supplier by wireless backhaul link
Content, and the caching energy consumption of base station is far below the energy consumption of wireless backhaul link.It can be seen that increasing base in a certain range
Cache size of standing has very big performance boost to system total energy consumption is reduced.
As shown in fig. 6, more different cache policies are (in institute's climbing form type when number of base stations and content number immobilize
Cache policy and Zipf distributed buffer strategy) under number of users and system total energy consumption relationship.Zipf distributed buffer strategy is as follows:
Most popular all base stations of content are bound to cache；The content that popularity rankings are the 2nd and the 3rd, each base station from the two with
Machine selects a content to carry out buffered in advance；The content that popularity rankings are 4-6, each base station is from selecting a content between three
Carry out buffered in advance；Popularity rankings 7-11, base station selects a content to carry out buffered in advance from this 5；Popularity
The content of ranking 12-20, base station select a content to carry out buffered in advance from this 9.As can be seen that with number of users
Increase, system total energy consumption is rising.This is because number of users increases in system, the content for needing to transmit also is increasing, so
System energy consumption is also increasing.In addition, carrying out content buffered in advance according to the instant content requests situation of user in institute's climbing form type
Strategy energy consumption than according to content popularit Zipf be distributed energy consumption of the content buffered in advance into base station is small.Reason is
Content of the buffered in advance into base station is distributed by Zipf and is unsatisfactory for the instant demand of user, and the content of user's current request is simultaneously
It is not necessarily the high content of popularity.Model proposed by the invention considers the instant demand of user, and cache user is current
In the case of the frequent content of request number of times.The content that but active user high for popularity does not request not buffered in advance.This
Reduce the caching energy consumption expense of base station and the energy consumption of backward wireless backhaul link to a certain extent.It can be seen that institute of the present invention
The scheme that joint content caching strategy and the user access network of proposition select has very big performance to mention to system total energy consumption is reduced
It rises.
To sum up, the present invention is pre- with the demand of specific content popularit of the user content substitutable demand based on random distribution
Survey, the base station that the base station and user that should be cached using QPSO algorithm to content in the CDN system based on caching should be accessed into
Row distribution, makes system reach energy consumption minimized with lower complexity.
By the cache policy of Joint regulation content and the access network selection scheme of user, the minimum of correspondence system is calculated
Power consumption values compare and obtain whole system optimal content caching strategy, user access network selection scheme and system minimum energy
Consumption value.
The present invention be user select heterogeneous network access when, not only ensure that each user for the quality of required content, but also
The energy consumption parameter for having taken into account backward wireless backhaul link, forward link and node B cache, effectively improves user in heterogeneous network
Service experience and reduce the whole energy consumption of system, minimize system total energy consumption.
Claims (2)
1. a kind of heterogeneous network selection method for considering content distribution energy consumption, which is characterized in that specific step is as follows:
Step 1: being directed to the downlink of some cellular network, the communication system between macro base station, small-cell base station and user is established
System；
Step 2: each user sends respective request to each small-cell base station connecting with itself simultaneously in some period
Content；
Step 3: communication system use quanta particle swarm optimization, calculate when system total energy consumption minimum, base station selected user with
And the optimal solution of the content of node B cache；
Specific step is as follows:
Step 301, initialization algorithm parameter；
Initialization: particle position, total number of particles M and maximum number of iterations S；
Each particle position includes two parts information: the knot that the result of I user's access base station and J content caching are placed
Fruit；
Step 302 is directed to each base station and each content, and the initial value and user for initializing node B cache location of content respectively connect
Connect the initial value of base station；
For small-cell base station k, M group cached parameters vector is randomly selectedAs node B cache content position
The initial value set；zm,kMeet node B cache length limitation and user accesses number limitation；zm,kIndicate particle m in small-cell base station k
Cached parameters vector；
For user i, chooses M group and access vector am,i=[ai,1′,ai,2′,...,ai,k'] as the initial of user's connection base station
Value；am,iIndicate the access vector of the particle m of user i access；
Step 303, the initial value for connecting base station with I user using the initial value of K node B cache location of content, setting are each
The initial position of particle；
The initial position X of m-th of particlem(0):
Xm(0)=[am,1(0),am,2(0),...,am,I(0),zm,1(0),zm,2(0),...,zm,K(0)]
Position of m-th of particle in the s times iteration is Xm(s) are as follows:
Xm(s)=[am,1(s),am,2(s),...,am,I(s),zm,1(s),zm,2(s),...,zm,K(s)]；
Step 304, sometime, calculate the corresponding system total energy consumption value of the particle using the position of each particle；
Each particle position is corresponding to calculate a system total energy consumption value；
pi,kIndicate small-cell base station SBSkTo user UEiTransmission power,Indicate content CjIt is buffered in small-cell base station SBSk
On caching power, pk,j,cIndicate content CjSmall-cell base station SBS is transferred to by wireless backhaul link from content supplierk
On transimission power；Indicate the incidence relation of user and content；It indicates caching relation, it is small to judge whether content j is cached to
On cell base station k, if so,Otherwise
Step 305, in subsequent time, update the position and speed of each particle；
More new formula is as follows:
Wherein, γ is the random number that value range is (0,1), and β represents shrinkage expansion coefficient, μ be value range be (0,1) with
Machine number；C (s) indicates the average desired positions in the s times iterative process,Pm(s) m-th of particle is indicated
Desired positions in the s times iterative process；It calculates as follows:
As s=0, Pm(0)=Xm(0)；F[Pm(s-1)] desired positions of m-th of particle in the s-1 times iterative process are indicated
Fitness function；F[Xm(s)] fitness function of position of m-th of particle in the s times iteration is indicated；
Vector P expression formula is as follows:
It is the random number between 0 to 1；G (s) indicates global desired positions of all particles in the s times iterative process:
Behind step 306, every position for updating primary particle, the fitness function F (X of each particle is calculatedm(s)), and whole M are recorded
The correspondence penalty value of the fitness function value of a particle and M particle
Fitness function F (XmIt (s)) include two parts: total energy consumption function and penalty；
Its expression formula are as follows:
Wherein α is penalty factor, determines specific influence of the value on fitness function of penalty；For total energy
Consume function；
The expression formula of penalty is as follows:
sk(k ∈ { 1,2,3......K }) indicates small-cell base station SBSkBuffer memory capacity size, meet: being cached to thereon interior
The total number of appearance
lk(k ∈ { 1,2,3......K }) indicates small-cell base station SBSkLoad capacity size, meet: be linked into use thereon
The total number at family
Step 307, the optimum position P that each particle is updated according to the fitness function of M particlem(s), it and records complete in M particle
Office optimum position G (s) optimal location G corresponding with true energy consumption minimumr(s)；
It is as follows to update rule:
If F (Xm(s)) < F (Xm(s-1)), then Pm(s)=Xm(s), otherwise Pm(s)=Pm(s-1)；
If F (Pm(s)) < F (G (s)), then G (s)=Pm(s), otherwise G (s)=G (s-1)；
If F (Pm(s)) < F (Gr(s)) andThen Gr(s)=Pm(s), otherwise Gr(s)=Gr(s-1)；
Step 308 judges whether the number of iterations s reaches maximum number of iterations S, if so, optimal location G in M particle of outputr
(s) corresponding solutionEnter step 309；Otherwise, return step 305；
Step 309, according to optimal location Gr(s) optimal solution X*, it is slow to obtain user and base station base station selected in communication system
The content deposited；
Optimal solution X*It include: the content optimal solution of base station selected user's optimal solution and node B cache, specifically:
Respectively base station selected user's optimal solution；The respectively content optimal solution of node B cache；
Step 4: according to base station selected user and the content optimal solution of node B cache, it is wide to each small-cell base station and user
Broadcast respective connection relationship and content caching corresponding relationship；
Step 5: for selected handy family and cached some small-cell base station of content, according to each selection user's
Request judges that base station itself whether there is request content, if it does, user directly obtains cache contents from the small-cell base station；
Otherwise, user obtains content by wireless backhaul link from content supplier.
2. a kind of heterogeneous network selection method for considering content distribution energy consumption as described in claim 1, which is characterized in that described
The step of one in, communication system include: macro base station cover K small-cell base station, collect be combined into SBSk, k ∈ 1,2,3,
4......K}；I user's collection is combined into UEi, i ∈ { 1,2,3,4......I }；And J properties collection is Cj, j ∈ 1,2,3,
4......J}；
Each user is respectively connected at any time on a small-cell base station, uses ai,kConnection relationship is indicated, if user i connects
It is connected on small-cell base station k, then ai,k=1, otherwise ai,k=0；
Each content size is the same, and the same content allows while being cached on different small-cell base stations, usesIndicate caching
Relationship, if content j is cached on small-cell base station k,OtherwiseMeet
Different user allows while requesting the same content, according to customer requirement retrieval user's request content；User and content
Incidence relation is usedIt indicates, if user i request content j,Otherwise
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610695982.XA CN106331083B (en) | 2016-08-19 | 2016-08-19 | A kind of heterogeneous network selection method considering content distribution energy consumption |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610695982.XA CN106331083B (en) | 2016-08-19 | 2016-08-19 | A kind of heterogeneous network selection method considering content distribution energy consumption |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106331083A CN106331083A (en) | 2017-01-11 |
| CN106331083B true CN106331083B (en) | 2019-07-09 |
Family
ID=57743567
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610695982.XA Active CN106331083B (en) | 2016-08-19 | 2016-08-19 | A kind of heterogeneous network selection method considering content distribution energy consumption |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106331083B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106912079B (en) * | 2017-02-20 | 2020-04-03 | 北京邮电大学 | Combined user access selection and resource allocation method in cache heterogeneous network |
| CN107182079B (en) * | 2017-06-08 | 2020-02-18 | 清华大学 | Small base station caching method |
| CN107968835B (en) * | 2017-12-05 | 2020-06-16 | 南京大学 | Wireless heterogeneous network video cache facility deployment method based on coding |
| CN108600365B (en) * | 2018-04-20 | 2020-05-22 | 西安交通大学 | Wireless heterogeneous network caching method based on sequencing learning |
| CN108668288B (en) * | 2018-04-23 | 2021-04-02 | 电子科技大学 | Method for optimizing small base station positions in wireless cache network |
| CN108566636B (en) * | 2018-04-28 | 2020-07-31 | 中国人民解放军陆军工程大学 | D2D random cache layout method oriented to different user preferences |
| CN110035415A (en) * | 2019-04-03 | 2019-07-19 | 西安交通大学 | A kind of D2D network-caching method for down loading of latency model |
| CN110138836B (en) * | 2019-04-15 | 2020-04-03 | 北京邮电大学 | Online cooperative caching method based on optimized energy efficiency |
| CN110324175B (en) * | 2019-05-27 | 2022-04-22 | 北京工业大学 | Network energy-saving method and system based on edge cache |
| CN111654873B (en) * | 2019-09-27 | 2022-08-16 | 西北大学 | Mobile CDN link selection energy consumption optimization method based on global utility cache strategy |
| CN111479312B (en) * | 2020-03-02 | 2022-05-17 | 重庆邮电大学 | Heterogeneous cellular network content caching and base station dormancy combined optimization method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102186072A (en) * | 2011-04-20 | 2011-09-14 | 上海交通大学 | Optimized transmission method of multi-rate multicast communication for scalable video stream |
| WO2014131000A2 (en) * | 2013-02-25 | 2014-08-28 | Interdigital Patent Holdings, Inc. | Centralized content enablement service for managed caching in wireless networks |
| CN104917809A (en) * | 2015-04-13 | 2015-09-16 | 南京邮电大学 | 5G wireless network virtualization system structure based on calculation and communication fusion |
| WO2016049333A1 (en) * | 2014-09-24 | 2016-03-31 | Interdigital Patent Holdings, Inc. | Method and system for creating a pre-fetching list for managed caching in small cell networks |
-
2016
- 2016-08-19 CN CN201610695982.XA patent/CN106331083B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102186072A (en) * | 2011-04-20 | 2011-09-14 | 上海交通大学 | Optimized transmission method of multi-rate multicast communication for scalable video stream |
| WO2014131000A2 (en) * | 2013-02-25 | 2014-08-28 | Interdigital Patent Holdings, Inc. | Centralized content enablement service for managed caching in wireless networks |
| WO2016049333A1 (en) * | 2014-09-24 | 2016-03-31 | Interdigital Patent Holdings, Inc. | Method and system for creating a pre-fetching list for managed caching in small cell networks |
| CN104917809A (en) * | 2015-04-13 | 2015-09-16 | 南京邮电大学 | 5G wireless network virtualization system structure based on calculation and communication fusion |
Non-Patent Citations (1)
| Title |
|---|
| LTE-Advanced系统能量节省自优化技术研究;李玥;《中国优秀硕士学位论文全文数据库》;20150415;全文 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106331083A (en) | 2017-01-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106331083B (en) | A kind of heterogeneous network selection method considering content distribution energy consumption | |
| CN112020103B (en) | Content cache deployment method in mobile edge cloud | |
| Pantisano et al. | In-network caching and content placement in cooperative small cell networks | |
| CN109121141A (en) | A kind of star based on MEC server ground two-stage edge network and its collaboration method | |
| CN107295619B (en) | Base station dormancy method based on user connection matrix in edge cache network | |
| CN110351754A (en) | Industry internet machinery equipment user data based on Q-learning calculates unloading decision-making technique | |
| CN112218337A (en) | Cache strategy decision method in mobile edge calculation | |
| CN108093435B (en) | Cellular downlink network energy efficiency optimization system and method based on cached popular content | |
| CN108900355A (en) | A kind of multistage edge network resource allocation methods in star ground | |
| CN108156596A (en) | Support the association of D2D- honeycomb heterogeneous networks federated user and content buffering method | |
| CN110138836A (en) | It is a kind of based on optimization energy efficiency line on cooperation caching method | |
| Mehrabi et al. | Energy-aware QoE and backhaul traffic optimization in green edge adaptive mobile video streaming | |
| CN105245592B (en) | Mobile network base station cache contents laying method based on adjacent cache cooperation | |
| CN108848521B (en) | Cellular heterogeneous network joint user association, content caching and resource allocation method based on base station cooperation | |
| CN108259628A (en) | Content caching and user-association combined optimization method in isomery cellular network | |
| CN108600998A (en) | Super density honeycomb and D2D isomery converged network cache optimization decision-making techniques | |
| CN106792995A (en) | The user access method of content low time delay transmission is ensured in a kind of following 5G networks | |
| Anokye et al. | A survey on machine learning based proactive caching | |
| CN108566636A (en) | D2D random cache distribution methods towards different user preference | |
| CN106304307B (en) | A kind of resource allocation methods under heterogeneous network converged | |
| CN108882269A (en) | The super-intensive network small station method of switching of binding cache technology | |
| Li et al. | Moving to green edges: A cooperative MEC framework to reduce energy demand of clouds | |
| Lei et al. | Partially collaborative edge caching based on federated deep reinforcement learning | |
| CN113709853B (en) | Network content transmission method and device oriented to cloud edge collaboration and storage medium | |
| Kollias et al. | Joint consideration of content popularity and size in device-to-device caching scenarios |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |