CN110418288B - D2D multicast communication clustering method based on user social attributes - Google Patents

D2D multicast communication clustering method based on user social attributes Download PDF

Info

Publication number
CN110418288B
CN110418288B CN201910672348.8A CN201910672348A CN110418288B CN 110418288 B CN110418288 B CN 110418288B CN 201910672348 A CN201910672348 A CN 201910672348A CN 110418288 B CN110418288 B CN 110418288B
Authority
CN
China
Prior art keywords
user
nth
cluster
mth
users
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.)
Expired - Fee Related
Application number
CN201910672348.8A
Other languages
Chinese (zh)
Other versions
CN110418288A (en
Inventor
龚文熔
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.)
Xian University of Science and Technology
Original Assignee
Xian University of Science and Technology
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 Xian University of Science and Technology filed Critical Xian University of Science and Technology
Priority to CN201910672348.8A priority Critical patent/CN110418288B/en
Publication of CN110418288A publication Critical patent/CN110418288A/en
Application granted granted Critical
Publication of CN110418288B publication Critical patent/CN110418288B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information
    • H04W4/025—Services making use of location information using location based information parameters
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/06—Selective distribution of broadcast services, e.g. multimedia broadcast multicast service [MBMS]; Services to user groups; One-way selective calling services
    • H04W4/08—User group management
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/70—Services for machine-to-machine communication [M2M] or machine type communication [MTC]
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W84/00—Network topologies
    • H04W84/18—Self-organising networks, e.g. ad-hoc networks or sensor networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Multimedia (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The invention discloses a D2D multicast clustering method based on user social attributes, which solves the problem of user performance loss caused by poor stability of the existing D2D multicast cluster. The method comprises the following steps: s101, randomly selecting K D2D users and distributing the users into K clusters; s102, calculating a mean vector m of each clusterk(ii) a S103, respectively calculating the clustering factor distance from the nth user to each cluster; s104, dividing the user n into clusters with the closest clustering factor distance; s105, calculating a new mean vector m for the kth cluster (K ═ 1.., K)k'; s106, judging the new mean value vector m of the kth clusterk' with the mean vector m of the current clusterkWhether the mean value vectors are equal or not, if not, replacing the original mean value vectors with the newly calculated mean value vectors; if equal, no change is made; s107, returning to execute S103-S106 until the mean vector of each cluster is equal to the result of the previous calculation, and then the algorithm converges; and S108, taking the result of the algorithm convergence as a final clustering result.

Description

D2D multicast communication clustering method based on user social attributes
[ technical field ] A method for producing a semiconductor device
The invention belongs to the technical field of communication, and particularly relates to a D2D multicast communication clustering method based on user social attributes.
[ background of the invention ]
The wireless communication network is increasingly heavily loaded due to huge data flow requirements and increasingly rich service types. Statistics and predictions in 2018 of Cisco show that the data volume of wireless mobile devices will increase 7 times from 2017 to 2022, and the data traffic of mobile devices in 2022 will reach 77.5Exabytes per month. Moving data traffic between 2017 and 2022 will continue to grow at a Compound Annual Growth Rate (CAGR) of 46%. In view of this, the demand for future mobile data traffic will continue to be explosive, which presents a significant challenge to the load capacity of wireless mobile networks.
As one of the key technologies of the fifth Generation (5th-Generation, 5G) mobile communication, the D2D communication technology has a wide application prospect in the aspects of improving Quality of Service (QoS) of users, expanding the coverage of cellular systems, improving the performance of systems, and the like, and becomes a research hotspot in the current industry and academia. The D2D multicast communication technology can finish the data transmission from one user to a plurality of users at the same time, and the spectrum utilization rate is greatly improved by sharing the spectrum resources by the users. With the rapid increase of multimedia services and the commercial use of 5G communication, the application of D2D multicast communication is becoming more and more extensive, and system research for D2D multicast communication is necessary.
The terminal device is indirectly endowed with social attributes due to human portability. The communication attribute of the terminal equipment is combined with the social attribute between the users to construct a direct communication network between the users, so that the resource sharing can be realized more efficiently, and the effectiveness and the reliability of transmission are further improved.
However, in the existing D2D clustering technology, social attributes and mobility of users are less considered, and the users are assumed to be in a quasi-static state, so that the stability of the obtained D2D multicast cluster is poor, and the performance of D2D communication users cannot be guaranteed. In most of the documents at present, a user social attribute factor represents whether a user has a social relationship or not only through a binary variable, and the above model is difficult to accurately measure the social relationship among users. It can be seen that the current modeling method for the social attributes of the D2D communication users is single, and deeper analysis and more reasonable modeling of the social attributes of the users are needed. Therefore, resource sharing can be realized more efficiently, and the effectiveness and the reliability of transmission are further improved.
[ summary of the invention ]
The invention aims to provide a D2D multicast clustering method based on user social attributes, so as to solve the problem of user performance loss caused by poor stability of the existing D2D multicast cluster.
The invention adopts the following technical scheme: the D2D multicast clustering method based on the social attributes of the users comprises the following steps:
s101, randomly selecting K D2D users { u }1, ... ,un, ... ,uKDivide into K clusters { C }1, ... ,Ck, ... CK}； un(n ═ 1.. times, K) denotes the nth D2D communication user, K is the total cluster number of D2D multicast communications, Ck(K1.., K) denotes a kth D2D multicast communication cluster;
s102, calculating a mean vector of each cluster:
Figure GDA0002773843600000021
wherein m iskMean vector, x, representing the kth clusternCoordinates representing the nth D2D communication user, ckIs the coordinate of the kth cluster center user, | | xn-ck||2Indicating the euclidean distance of the nth D2D communication user to the kth cluster center point,
Figure GDA0002773843600000022
the social attribute factors from the nth user to the kth cluster center user are zeta and eta are proportionality coefficients, and zeta + eta is 1, | CkL represents the number of users D2D in the kth cluster;
s103, for the nth user, respectively calculating the clustering factor distance from the nth user to each cluster:
Figure GDA0002773843600000031
dk,nrepresenting the clustering factor distance from the nth D2D communication user to the kth cluster center user, and selecting the cluster with the closest clustering factor distance as the cluster finally selected by the nth user, namely
Figure GDA0002773843600000032
Wherein, C* nCluster numbers selected for the nth user;
s104, dividing the user n into clusters C* nPerforming the following steps;
s105, for the kth cluster (K ═ 1.., K), a new mean vector m is calculatedk′；
S106, determining a new mean vector m of the kth cluster (K ═ 1.., K.)k' with the mean vector m of the current clusterkWhether they are equal; if not equal mk≠mk', (K1.., K), replacing the original mean vector with the newly calculated mean vector
Figure GDA0002773843600000033
If equal, mk＝mk', (K-1.., K), the value of the original mean vector is unchanged;
s107, returning to execute S103-S106 until the mean vector of each cluster is equal to the mean vector calculated in the previous round, and then the algorithm converges;
and S108, taking the result of the algorithm convergence as a final clustering result.
Further, after a final clustering result is obtained, a core user is selected in a kth cluster (K ═ 1.., K.), and a specific method is as follows: the core user fitness F of the nth usernAs a parameter for measuring whether the user is suitable as a core user, it is expressed as:
Figure GDA0002773843600000034
wherein, alpha, beta,
Figure GDA0002773843600000035
Tau is a weight parameter and satisfies
Figure GDA0002773843600000036
EnAvailable battery capacity for the nth user, DnSocial factor, P, for the nth usernIs the residence probability of the nth user, SnAvailable storage space for the nth user;
according to the fitness of potential core users, selecting a user with the highest fitness in the kth cluster as a core node of the kth cluster, wherein the core node is used as a transmission node for D2D multicast communication;
Figure GDA0002773843600000037
wherein n isk *Core users in the kth cluster; fnCore user fitness, Ω, for user nkRepresenting the set of all D2D users in the kth cluster.
Further, the social factor D of the nth usernCalculated by the following formula:
Figure GDA0002773843600000041
wherein the content of the first and second substances,
Figure GDA0002773843600000042
for the social attribute factor from the nth user to the mth user, m is equal to omegakAnd m ≠ n denotes that the mth user is the user in the kth cluster, and meanwhile, the mth user is different from the nth user.
Further, the residence probability P of the nth usern＝Tc,n/Ts,nWherein, Ts,nRepresents the total historical time, T, that the nth user resided in all clustersc,nIndicating the time that the nth user resides in the current cluster.
Further, the social attribute factor between the nth user and the mth user may be expressed as:
Figure GDA0002773843600000043
where p is a weighting factor,
Figure GDA0002773843600000048
for a normalized representation of the average association duration between the nth user and the mth user,
Figure GDA0002773843600000044
the transmission revenue of D2D communication for the nth user and the mth user is proportional to the total transmission revenue of the nth user and the mth user.
Further, the normalization of the average association duration between the nth user and the mth user represents:
Figure GDA0002773843600000045
wherein
Figure GDA0002773843600000046
The average duration of content sharing between the nth user and the mth user within the Δ t time is represented by the following specific calculation formula:
Figure GDA0002773843600000047
wherein z isn,m(t) represents the period between the nth user and the mth user at time tContent sharing State, zn,m(t) ═ 1 indicates that the nth user and the mth user are in a content sharing state at the time t; otherwise zn,m(t)＝0；n,m(t) represents a content sharing value between the nth user and the mth user at time t.
Further, in the time delta t, the transmission profit of the D2D communication of the nth user to the mth user accounts for the proportion of the total transmission profit between the nth user and the mth user
Figure GDA0002773843600000051
Can be calculated by the following formula:
Figure GDA0002773843600000052
wherein Cotn,m(t) represents the transmission gain of the nth user to the mth user at the time t, Cotm,n(t) represents the transmission gain of the mth user to the nth user at time t.
The invention has the beneficial effects that: according to the invention, the user social attribute factors are introduced into the D2D multicast communication clustering algorithm, so that a more stable D2D multicast communication transmission group can be established, resource sharing is realized more efficiently, and the effectiveness and reliability of D2D communication transmission are further improved.
[ description of the drawings ]
FIG. 1 is a system efficiency diagram of an embodiment of the present invention;
FIG. 2 is a communication throughput diagram of an embodiment of the present invention;
FIG. 3 is a graph of outage probability according to an embodiment of the present invention;
FIG. 4 is a graph of cluster robustness for an embodiment of the present invention;
fig. 5 is a flowchart of a method for clustering D2D multicast communication based on social attributes of users according to the present invention.
[ detailed description ] embodiments
The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
The invention provides a D2D multicast clustering method based on user social attributes.
1. The method is based on a D2D communication user social attribute model, and the model is as follows:
the social attributes between users are influenced by the frequency of user connection, the connection duration, the user intimacy and the interaction mode. Since the research object of the project is data transmission between D2D communication users, the invention mainly considers the frequency and duration of content sharing between users. The degree of social association between users is described by the number of content shares between two users within a certain duration and the duration of each content share.
Order ton,m(t) represents a content sharing value between the nth user and the mth user at time t, zn,m(t) represents a content sharing state between the nth user and the mth user at time t, zn,m(t) ═ 1 indicates that the nth user and the mth user are in a content sharing state at the time t; otherwise zn,m(t) is 0. The average duration of content sharing between the nth user and the mth user within the Δ t time is as follows:
Figure GDA0002773843600000061
according to the Gaussian similarity function, the normalized expression of the average correlation duration between the users can be obtained:
Figure GDA0002773843600000062
under the D2D communication incentive mechanism, in the time delta t, the proportion of the transmission benefit of the D2D communication of the nth user to the mth user to the total transmission benefit of the nth user to the mth user is as follows:
Figure GDA0002773843600000063
wherein Cotn,m(t) represents the transmission gain of the nth user to the mth user at the time t, Cotm,n(t) represents the transmission gain of the mth user to the nth user at time t.
Considering the association relationship and the transmission profit among the D2D users, the social attribute factor between the nth user and the mth user can be expressed as:
Figure GDA0002773843600000064
wherein ρ is a weighting factor, which can be adjusted according to the service characteristics.
2. To stabilize a robust D2D multicast group, in clustering, we consider social relationships between users.
The social K-means clustering algorithm aims to solve the following optimization problems:
Figure GDA0002773843600000071
wherein, | | xj-ci||2Refers to the euclidean distance between user j and the ith cluster center point. OmegaijTo distribute the parameters, when user j belongs to the ith cluster (x)j∈Ci)，ωij1, otherwise, ωij0. C2 indicates that each cluster is not empty; c3 indicates that each user can only exist within one cluster. The above problem is the NP-hard problem, even if K is 2. In order to solve the problems, the invention provides a social K-means clustering algorithm.
The social-based K-means clustering algorithm is to cluster N D2D users { u }1,u2, ... ,uNDivide into K clusters { C1,C2, ... CKAs shown in fig. 5, the basic steps are as follows:
s101, randomly selecting K D2D users { u }1, ... ,un, ... ,uKDivide into K clusters { C }1, ... ,Ck, ...CK}； un(n ═ 1.. times, K) denotes the nth D2D communication user, K is the total cluster number of D2D multicast communications, Ck(K1.., K) denotes the kth D2D multicast communication cluster.
S102, calculating a mean vector of each cluster:
Figure GDA0002773843600000072
wherein m iskMean vector, x, representing the kth clusternCoordinates representing the nth D2D communication user, ckIs the coordinate of the kth cluster center user, | | xn-ck||2Indicating the euclidean distance of the nth D2D communication user to the kth cluster center point,
Figure GDA0002773843600000073
the social attribute factors from the nth user to the kth cluster center user are zeta and eta are proportionality coefficients, and zeta + eta is 1, | CkAnd | represents the number of users D2D in the kth cluster.
S103, for the nth user, respectively calculating the clustering factor distance from the nth user to each cluster:
Figure GDA0002773843600000081
dk,nrepresenting the clustering factor distance from the nth D2D communication user to the kth cluster center user, and selecting the cluster with the closest clustering factor distance as the cluster finally selected by the nth user, namely
Figure GDA0002773843600000082
Wherein, C* nCluster numbers selected for the nth user;
s104, dividing the user n into clusters C* nPerforming the following steps;
s105, for the kth cluster (K ═ 1.., K), a new mean vector m is calculatedk′；
S106, determining a new mean vector m of the kth cluster (K ═ 1.., K.)k' with the mean vector m of the current clusterkWhether they are equal; if not equal mk≠mk', (K1.., K), replacing the original mean vector with the newly calculated mean vector
Figure GDA0002773843600000083
If equal, mk＝mk', (K-1.., K), the value of the original mean vector is unchanged;
s107, returning to execute S103-S106 until the mean vector of each cluster is equal to the mean vector calculated in the previous round, and then the algorithm converges;
and S108, taking the result of the algorithm convergence as a final clustering result.
3. After the D2D multicast clustering is completed, core users need to be selected from the cluster as D2D transmitting nodes for data transmission, which is as follows:
in order to meet the data transmission rate requirement, the D2D core user should be selected by taking the power, storage space and mobility characteristics of the user terminal into consideration.
The users residing in a cluster for a long time also have higher probability of residing in the transmission process, and if the users are used as core users, the reliability is higher. Probability of residence P of usernRelated to the time the user stays within the cluster. Order residence probability Pn＝Tc,n/Ts,n，Ts,nRepresents the total historical time, T, that the nth user resided in all clustersc,nIndicating the time that the nth user resides in the current cluster.
The core user fitness F of the nth usernAs a parameter for measuring whether the user is suitable as a core user, it is expressed as:
Figure GDA0002773843600000096
wherein, alpha, beta,
Figure GDA0002773843600000097
Tau is a weight parameter and satisfies
Figure GDA0002773843600000098
EnAvailable battery capacity for the nth user, DnSocial factor, P, for the nth usernIs the residence probability of the nth user, SnAvailable for the nth userA storage space;
according to the fitness of potential core users, selecting a user with the highest fitness in the kth cluster as a core node of the kth cluster, wherein the core node is used as a transmission node for D2D multicast communication;
Figure GDA0002773843600000091
wherein n isk *Core users in the kth cluster; fnCore user fitness, Ω, for user nkRepresenting the set of all D2D users in the kth cluster.
Wherein the social factor D of the nth usernCalculated by the following formula:
Figure GDA0002773843600000092
wherein the content of the first and second substances,
Figure GDA0002773843600000099
for the social attribute factor from the nth user to the mth user, m is equal to omegakAnd m ≠ n denotes that the mth user is the user in the kth cluster, and meanwhile, the mth user is different from the nth user.
Probability of residence P for nth usern＝Tc,n/Ts,nWherein, Ts,nRepresents the total historical time, T, that the nth user resided in all clustersc,nIndicating the time that the nth user resides in the current cluster.
The social attribute factor between the nth user to the mth user may be expressed as:
Figure GDA0002773843600000093
where p is a weighting factor,
Figure GDA0002773843600000094
for a normalized representation of the average association duration between the nth user and the mth user,
Figure GDA0002773843600000095
the transmission revenue of D2D communication for the nth user and the mth user is proportional to the total transmission revenue of the nth user and the mth user.
The normalized representation of the average association duration between the nth user and the mth user is:
Figure GDA0002773843600000101
wherein
Figure GDA0002773843600000102
The average duration of content sharing between the nth user and the mth user within the Δ t time is represented by the following specific calculation formula:
Figure GDA0002773843600000103
wherein z isn,m(t) represents a content sharing state between the nth user and the mth user at time t, zn,m(t) ═ 1 indicates that the nth user and the mth user are in a content sharing state at the time t; otherwise zn,m(t)＝0；n,m(t) represents a content sharing value between the nth user and the mth user at time t.
In the delta t time, the proportion of the transmission profit of the D2D communication of the nth user to the mth user to the total transmission profit between the nth user and the mth user
Figure GDA0002773843600000104
Can be calculated by the following formula:
Figure GDA0002773843600000105
wherein Cotn,m(t) represents the transmission gain of the nth user to the mth user at the time t, Cotm,n(t) represents the transmission gain of the mth user to the nth user at time t.
Secondly, analyzing a system utility function:
let Φ be the set Φ of D2D communication core users {1, 2.Ψ is a set of receiving nodes, Ψ ═ 1, 2., U }, and user U ∈ Ψ. Let Θ be { 1., } K denote the set of all available subcarriers in the system. According to the shannon formula, it can be deduced that the transmission rate of the mth core user connected with the uth receiving user on the subcarrier k is:
Figure GDA0002773843600000106
where B is the unit bandwidth and N0Noise per unit bandwidth, NmRepresenting the set of receiving users of the mth D2D multicast group,
Figure GDA0002773843600000111
representing the channel gain from the transmitting node m to the receiving node u' on the subcarrier k, further, the average transmission rate can be obtained:
Figure GDA0002773843600000112
wherein T is the transmission time.
Let alphauRepresents the benefit, β, of user u per unit data ratem,uBandwidth, gamma, consumed when connecting a sending node m for a user um,uThe bandwidth, phi, consumed by the user u when feeding back to the sending node muIndicates the benefit per unit data rate, ψ, of user u selecting D2D communicationsuThe cost of storage space for user u.
Figure GDA0002773843600000113
Representing stored data ruOccupied memory space, BeuIndicating the channel bandwidth occupied by the estimated feedback. The utility function between user u and sending node m (m ≠ 0) is:
Figure GDA0002773843600000114
wherein, am,uRepresenting content distribution parameters: a when the data requested by user u has been stored in the sending node mm,u1, otherwise, am,u＝0。
Figure GDA0002773843600000115
The allocation relationship between the sub-carriers and the links is represented as follows:
Figure GDA0002773843600000116
indicating that subcarrier k is allocated to the link of transmitting node m and user u, otherwise,
Figure GDA0002773843600000117
em,u1 means that the user u selects the sending node m for transmission, otherwise em,u＝0。
Thirdly, analyzing cluster stability:
the stability of the mth cluster can be defined as:
Figure GDA0002773843600000118
wherein, Fm0Adaptation value for core user of mth cluster, DmuIndicates the social attribute value, lambda, of the u-th user in the m-th cluster1，λ2As a weight coefficient, satisfies λ1+λ2＝1。
The cluster mean stability of the system can be expressed as:
Figure GDA0002773843600000121
fourth, example
In order to evaluate the D2D multicast clustering method (SA-KCA for short) based on the social attributes of the users, the simulation results are compared with the performance of a distance-based clustering algorithm (DBCA for short) and a traditional K-means clustering algorithm (KCA for short). The basic idea of a distance-based clustering algorithm (DBCA) is to randomly select users as core users, and then each user selects a cluster where the core user closest to the user is located to join the core user. The basic idea of the conventional K-means clustering algorithm (KCA) is the same as that of the patent, but the updating and selection of clusters are performed according to the euclidean distance between users.
The basic simulation parameters are shown in table 1:
TABLE 1 simulation parameters
Figure GDA0002773843600000122
As shown in fig. 1, which describes the relationship between the system utility and the number of clusters K, it can be seen that the SA-KCA proposed by the present invention can obtain the highest system utility. In the figure, the system effectiveness under the SA-KCA algorithm can reach 6.7 x 109Under the other three algorithms, the system effectiveness is lower than 6.5 x 109The SA-KCA algorithm not only considers the euclidean distance between users, but also takes the social attributes into consideration, wherein the social attributes include the overhead of the system, the energy remaining of the users, the battery remaining and other factors, which are important parameters of the system utility, so the system utility under the SA-KCA algorithm can be guaranteed. It can also be seen from the figure that the system effectiveness under the SA-KCA and DBCA algorithms is slightly reduced with the increase of K, and the reduction amplitude of the system effectiveness under the SA-KCA is small with the increase of K. This is because as the number of D2D multicast groups increases, the number of users in each multicast group decreases, and therefore, the multicast efficiency decreases, the number of consumed resources increases, and the system utility decreases. The algorithm provided by the invention can avoid the consumption as much as possible in the clustering process, so that the reduction amplitude is not obvious.
A graph of D2D communication throughput performance versus K is shown in fig. 2. From the figure, it can be concluded that, first, the system of SA-KCA achieves the highest D2D throughput performance, averaging up to 225Mbps, and the system D2D employing DBCA has the second highest overall throughput performance, averaging about 215 Mbps. System D2D with KCA had the lowest overall throughput, averaging about 125 Mbps. Secondly, the throughput of the system D2D under the SA-KCA and DBCA algorithms is slightly reduced along with the increase of K, and the reduction amplitude of the throughput of the system D2D under the SA-KCA along with the increase of K is small. Finally, the throughput of the system D2D under the KCA algorithm is in an increasing trend along with the increase of K, and although the throughput performance under the KCA system is not reduced along with the increase of the number of multicast groups K, the throughput performance is lower.
Fig. 3 compares the outage probability for users under different algorithms. As can be seen from the figure: the interruption probability of users under the SA-KCA and DBCA systems tends to increase along with the increase of K. The interruption probability of the users under the SA-KCA system is the lowest, and the interruption probability of the users under the DBCA system and the KCA system is higher. This also proves that although the system utility and throughput performance under the KCA system do not decrease with the increase of the number K of multicast groups, the performance of the user cannot be well guaranteed.
Fig. 4 gives a simulation curve of cluster robustness. As can be seen from the figure, the cluster robustness obtained under the SA-KCA algorithm has the best performance, the KCA robustness is the second order, and the cluster robustness obtained under the DBCA algorithm is poor. This is because the DBCA algorithm does not consider factors such as the mobility of the user and the remaining power of the core user in the clustering process, resulting in poor cluster stability.
From the simulation results, the algorithm provided by the invention not only can improve the utility and throughput performance of the system, but also provides higher cluster robustness.

Claims (4)

1. The D2D multicast clustering method based on the social attributes of the users is characterized by comprising the following steps:
s101, randomly selecting K D2D users { u }1, ...,un, ... uKDivide into K clusters { C }1, ... ,Ck, ...CK}；unDenotes the nth D2D communication subscriber, n 1.., K; k is the total cluster number of D2D multicast communication, CkDenotes the kth D2D multicast communication cluster, K1., K;
s102, calculating a mean vector of each cluster:
Figure FDA0002773843590000011
wherein m iskMean vector, x, representing the kth clusternCoordinates representing the nth D2D communication user, ckIs the coordinate of the kth cluster center user, | | xn-ck| | represents the euclidean distance of the nth D2D communication user to the kth cluster center point,
Figure FDA0002773843590000012
the social attribute factors from the nth user to the kth cluster center user are zeta and eta are proportionality coefficients, and zeta + eta is 1, | CkL represents the number of users D2D in the kth cluster;
wherein, the social attribute factor between the nth user and the mth user can be expressed as:
Figure FDA0002773843590000013
where p is a weighting factor,
Figure FDA0002773843590000014
for a normalized representation of the average association duration between the nth user and the mth user,
Figure FDA0002773843590000015
the proportion of the transmission revenue of the D2D communication of the nth user and the mth user to the total transmission revenue of the nth user and the mth user;
the normalized representation of the average association duration between the nth user and the mth user is:
Figure FDA0002773843590000016
wherein
Figure FDA0002773843590000017
The average duration of content sharing between the nth user and the mth user within the Δ t time is represented by the following specific calculation formula:
Figure FDA0002773843590000018
wherein z isn,m(t) represents a content sharing state between the nth user and the mth user at time t, zn,m(t) ═ 1 indicates that the nth user and the mth user are in a content sharing state at the time t; otherwise zn,m(t)＝0；n,m(t) represents a content sharing value between the nth user and the mth user at time t;
in the delta t time, the proportion of the transmission profit of the D2D communication of the nth user to the mth user to the total transmission profit between the nth user and the mth user
Figure FDA0002773843590000021
Can be calculated by the following formula:
Figure FDA0002773843590000022
wherein Cotn,m(t) represents the transmission gain of the nth user to the mth user at the time t, Cotm,n(t) represents the transmission benefit of the mth user to the nth user at the time t;
s103, for the nth user, respectively calculating the clustering factor distance from the nth user to each cluster:
Figure FDA0002773843590000023
dk,nrepresenting the clustering factor distance from the nth D2D communication user to the kth cluster center user, and selecting the cluster with the closest clustering factor distance as the cluster finally selected by the nth user, namely
Figure FDA0002773843590000024
Wherein, C* nCluster numbers selected for the nth user;
s104, dividing the user n into clusters C* nPerforming the following steps;
s105, for the kth cluster, k is 1,., K, calculating a new mean vector mk′；
S106, determining a new mean vector m of the kth cluster, K being 1k' with the mean vector m of the current clusterkWhether they are equal; if not equal mk≠mk', K, replacing the original mean vector with the newly calculated mean vector
Figure FDA0002773843590000025
If equal, mk＝mk', K1, K, the value of the original mean vector is not changed;
s107, returning to execute S103-S106 until the mean vector of each cluster is equal to the mean vector calculated in the previous round, and then the algorithm converges;
and S108, taking the result of the algorithm convergence as a final clustering result.
2. The D2D multicast clustering method based on the social attributes of users as claimed in claim 1, wherein after the final clustering result is obtained, the selection of the core user is performed in the K-th cluster K-1. The core user fitness F of the nth usernAs a parameter for measuring whether the user is suitable as a core user, it is expressed as:
Figure FDA0002773843590000031
wherein, alpha, beta,
Figure FDA0002773843590000032
Tau is a weight parameter and satisfies
Figure FDA0002773843590000033
EnAvailable battery capacity for the nth user, DnSocial factor, P, for the nth usernIs the residence probability of the nth user, SnAvailable storage space for the nth user;
according to the fitness of potential core users, selecting a user with the highest fitness in the kth cluster as a core node of the kth cluster, wherein the core node is used as a transmission node for D2D multicast communication;
Figure FDA0002773843590000034
wherein n isk *Core users in the kth cluster; fnCore user fitness, Ω, for user nkRepresenting the set of all D2D users in the kth cluster.
3. The user social attribute-based D2D multicast clustering method of claim 2, wherein the nth user's social factor DnCalculated by the following formula:
Figure FDA0002773843590000035
wherein the content of the first and second substances,
Figure FDA0002773843590000036
for the social attribute factor from the nth user to the mth user, m is equal to omegakAnd m ≠ n denotes that the mth user is the user in the kth cluster, and meanwhile, the mth user is different from the nth user.
4. The method for D2D multicast clustering based on user social attributes as claimed in claim 2, wherein the n-th user's residence probability Pn＝Tc,n/Ts,n，
Wherein, Ts,nRepresents the total historical time, T, that the nth user resided in all clustersc,nIndicating the time that the nth user resides in the current cluster.
CN201910672348.8A 2019-07-24 2019-07-24 D2D multicast communication clustering method based on user social attributes Expired - Fee Related CN110418288B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910672348.8A CN110418288B (en) 2019-07-24 2019-07-24 D2D multicast communication clustering method based on user social attributes

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910672348.8A CN110418288B (en) 2019-07-24 2019-07-24 D2D multicast communication clustering method based on user social attributes

Publications (2)

Publication Number Publication Date
CN110418288A CN110418288A (en) 2019-11-05
CN110418288B true CN110418288B (en) 2020-12-25

Family

ID=68362895

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910672348.8A Expired - Fee Related CN110418288B (en) 2019-07-24 2019-07-24 D2D multicast communication clustering method based on user social attributes

Country Status (1)

Country Link
CN (1) CN110418288B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111526489B (en) * 2020-04-30 2021-11-23 上海海事大学 D2D playing content distribution method based on social network relationship

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107979824A (en) * 2017-10-20 2018-05-01 西安电子科技大学 A kind of D2D multipath resource distribution methods under wireless network virtualization scene
CN109951828A (en) * 2019-04-01 2019-06-28 西安电子科技大学 It is a kind of social activity sensing network in D2D multicast video transmission method for channel allocation

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3286961A1 (en) * 2015-04-24 2018-02-28 Nokia Solutions and Networks Oy Using networking relationship in configuring radio connectivity
CN106507316B (en) * 2016-11-02 2019-03-26 西安邮电大学 User's sub-clustering and resource allocation methods under a kind of D2D multicast scene
CN107249205B (en) * 2017-06-27 2019-08-16 北京邮电大学 A kind of resource allocation methods, device and user terminal
CN107889082B (en) * 2017-11-01 2020-04-14 南京邮电大学 D2D device discovery method utilizing social relationship among users
CN108616845B (en) * 2018-03-30 2021-10-26 佛山市顺德区中山大学研究院 D2D grouping multi-target caching method based on social content and system and device thereof

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107979824A (en) * 2017-10-20 2018-05-01 西安电子科技大学 A kind of D2D multipath resource distribution methods under wireless network virtualization scene
CN109951828A (en) * 2019-04-01 2019-06-28 西安电子科技大学 It is a kind of social activity sensing network in D2D multicast video transmission method for channel allocation

Also Published As

Publication number Publication date
CN110418288A (en) 2019-11-05

Similar Documents

Publication Publication Date Title
CN111182570B (en) User association and edge computing unloading method for improving utility of operator
Wu et al. Social-aware rate based content sharing mode selection for D2D content sharing scenarios
El-Hajj et al. A stable matching game for joint uplink/downlink resource allocation in OFDMA wireless networks
CN110430613B (en) Energy-efficiency-based resource allocation method for multi-carrier non-orthogonal multiple access system
CN111447619A (en) Joint task unloading and resource allocation method in mobile edge computing network
Pantisano et al. Match to cache: Joint user association and backhaul allocation in cache-aware small cell networks
CN103796244B (en) A kind of many net access selection methods of the wireless isomer of equally loaded
CN105407004B (en) The method and device of content distribution is carried out based on edge hotspot
CN107708152B (en) Task unloading method of heterogeneous cellular network
Zhang et al. DMRA: A decentralized resource allocation scheme for multi-SP mobile edge computing
Saberali et al. Full characterization of optimal uncoded placement for the structured clique cover delivery of nonuniform demands
Kan et al. QoS-aware mobile edge computing system: Multi-server multi-user scenario
CN110677175A (en) Sub-channel scheduling and power distribution joint optimization method based on non-orthogonal multiple access system
CN104469851B (en) Balanced handling capacity and the resource allocation methods of delay in a kind of LTE downlinks
CN110418288B (en) D2D multicast communication clustering method based on user social attributes
CN112202847A (en) Server resource allocation method based on mobile edge calculation
Prabowo et al. Energy efficient resources allocations for wireless communication systems
CN106793122B (en) Method for safely allocating each bit minimized wireless resources of heterogeneous network
CN112579203A (en) Server dormancy-based collaborative computing method in heterogeneous edge network
CN112566131A (en) C-RAN network resource allocation method based on time delay constraint
CN107995034B (en) Energy and service cooperation method for dense cellular network
Ye et al. Minimum cost based clustering scheme for cooperative wireless caching network with heterogeneous file preference
CN112770343B (en) D2D-NOMA resource allocation method and system based on HAGA
Zhu et al. Energy-aware incentive mechanism for content sharing through device-to-device communications
CN105163392B (en) Ensure the multi-BS (base station) multi-user proportional fair dispatching method of service quality

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
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20201225

Termination date: 20210724

CF01 Termination of patent right due to non-payment of annual fee