CN108834080B - Distributed caching and user association method based on multicast technology in heterogeneous networks - Google Patents
Distributed caching and user association method based on multicast technology in heterogeneous networks Download PDFInfo
- Publication number
- CN108834080B CN108834080B CN201810341885.XA CN201810341885A CN108834080B CN 108834080 B CN108834080 B CN 108834080B CN 201810341885 A CN201810341885 A CN 201810341885A CN 108834080 B CN108834080 B CN 108834080B
- Authority
- CN
- China
- Prior art keywords
- user
- base station
- strategy
- association
- 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
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/06—Selective distribution of broadcast services, e.g. multimedia broadcast multicast service [MBMS]; Services to user groups; One-way selective calling services
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B17/00—Monitoring; Testing
- H04B17/30—Monitoring; Testing of propagation channels
- H04B17/391—Modelling the propagation channel
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/02—Traffic management, e.g. flow control or congestion control
- H04W28/10—Flow control between communication endpoints
- H04W28/14—Flow control between communication endpoints using intermediate storage
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W72/00—Local resource management
- H04W72/50—Allocation or scheduling criteria for wireless resources
- H04W72/53—Allocation or scheduling criteria for wireless resources based on regulatory allocation policies
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Multimedia (AREA)
- Physics & Mathematics (AREA)
- Electromagnetism (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
本发明公开了一种异构网络中基于多播技术的分布式缓存和用户关联方法,属于无线通信技术领域。本发明针对移动数据流量剧增及小基站(Small Cell Base Station,SBS)密集部署带来的回程拥塞问题,结合多播传输和协作缓存两者的优势,在保证所有用户的服务质量前提下,提出了一种多播协作缓存算法对用户关联策略和缓存配置策略进行联合优化以最小化系统总回程成本。与现有技术相比,本发明具有更低的系统回程成本和功率消耗。
The invention discloses a distributed cache and user association method based on multicast technology in a heterogeneous network, and belongs to the technical field of wireless communication. Aiming at the backhaul congestion problem caused by the sharp increase in mobile data traffic and the dense deployment of small base stations (Small Cell Base Station, SBS), the present invention combines the advantages of multicast transmission and cooperative buffering, and under the premise of ensuring the service quality of all users, A multicast cooperative caching algorithm is proposed to jointly optimize the user association strategy and the cache configuration strategy to minimize the total backhaul cost of the system. Compared with the prior art, the present invention has lower system backhaul cost and power consumption.
Description
技术领域technical field
本发明涉及一种无线边缘缓存技术,特别是涉及一种异构网络中基于多播技术的分布式缓存和用户关联方法,属于无线通信技术领域。The invention relates to a wireless edge caching technology, in particular to a distributed caching and user association method based on multicast technology in a heterogeneous network, and belongs to the technical field of wireless communication.
背景技术Background technique
随着移动智能设备的不断增长,移动数据流量的急剧增加,对于新兴应用和服务的需求也开始呈现爆炸式增长。小基站密集部署是应对这种移动数据流量增长的一种关键技术但是这种方法将给回程链路带来巨大压力,同时也会使得系统能耗增加。With the continuous growth of mobile smart devices and the dramatic increase in mobile data traffic, the demand for emerging applications and services has also begun to explode. The dense deployment of small cells is a key technology to cope with this increase in mobile data traffic, but this approach will put enormous pressure on the backhaul link and will also increase system energy consumption.
多播和无线缓存被视为应对上述问题的两种很有前景的技术。缓存通过充分利用流量的周期模式将流量从高峰时段(如白天)转移到非高峰时段(如夜晚),具体做法是在非高峰时段将部分文件提前缓存在基站等设备中,而在高峰时段服务用户请求。缓存对于有足够多内容复用时的场景很有效。多播则是通过使用单点对多点而非点对点的传输方式来服务用户对相同文件的并发请求,从而减少通信流量,并且降低能耗及带宽消耗。多播在许多用户并发地产生对相同文件的请求时非常有效。这样的场景在人群聚集事件中更常见,因为有大量的共同定位的人对相同的内容感兴趣,例如,体育比赛和音乐会活动中,经常有成千上万的参与者。社交网络平台和新闻服务等多种新服务也采用了一对多的通信范例,例如,Tweeter、Facebook等的更新,预计多播技术将会被更频繁地应用。目前学术界和工业界一般都是将这两种技术分别应用于不同场景和方面。将无线缓存和多播进行结合的文献研究还较少。Multicast and wireless caching are seen as two promising technologies to address the above problems. Cache transfers traffic from peak hours (such as daytime) to off-peak hours (such as night) by making full use of the periodic pattern of traffic. The specific method is to cache some files in devices such as base stations in advance during off-peak hours, and serve them during peak hours. User request. Caching works well for scenarios where there is enough content to reuse. Multicast is to use point-to-multipoint rather than point-to-point transmission to serve users' concurrent requests for the same file, thereby reducing communication traffic, energy consumption and bandwidth consumption. Multicasting is very effective when many users are generating requests for the same file concurrently. Such scenarios are more common in crowd gathering events because there are a large number of co-located people interested in the same content, for example, sports games and concert events, often with thousands of participants. Multiple new services such as social networking platforms and news services are also adopting the one-to-many communication paradigm, for example, updates from Tweeter, Facebook, etc., and multicast technology is expected to be applied more frequently. At present, academia and industry generally apply these two technologies to different scenarios and aspects. There are few literature studies on the combination of wireless caching and multicasting.
发明内容SUMMARY OF THE INVENTION
发明目的:基于上述现有技术的不足,本发明结合无线缓存技术和多播技术两者的优势,提出一种异构网络中基于多播技术的分布式缓存和用户关联方法,旨在降低系统总回程成本,同时在一定程度上降低系统功耗。Purpose of the invention: Based on the shortcomings of the above-mentioned prior art, the present invention combines the advantages of both wireless caching technology and multicast technology, and proposes a distributed caching and user association method based on multicast technology in a heterogeneous network, aiming to reduce the system Total backhaul cost while reducing system power consumption to a certain extent.
技术方案:为实现上述发明目的,本发明采用如下技术方案:Technical scheme: In order to realize the above-mentioned purpose of the invention, the present invention adopts the following technical scheme:
一种异构网络中基于多播技术的分布式缓存和用户关联方法,包括如下步骤:A distributed cache and user association method based on multicast technology in a heterogeneous network, comprising the following steps:
(1)以(M+1)×K的矩阵表示基站(包括宏基站和小基站)和用户关联策略,以(M+1)×N的矩阵表示基站的缓存配置策略,设每个时隙内请求相同文件的用户形成一个多播组,同组用户采用多播方式进行文件传输,在保证用户服务质量,保证每个用户在每个时隙内只能由一个基站提供服务,以及保证缓存在基站的文件不会超过基站缓存容量的约束条件下,以最小化各多播组中用户的最大传输速率之和为目标,建立用户关联策略和缓存配置策略的联合优化问题;其中基站包括宏基站和小基站,M为小基站总数目,网络中还包括1个宏基站,N为一段时间内所有用户请求的文件总数目,K为网络中移动用户总数目;(1) The base station (including the macro base station and the small base station) and the user association strategy are represented by a matrix of (M+1)×K, and the cache configuration policy of the base station is represented by a matrix of (M+1)×N. Set each time slot Users who request the same file form a multicast group, and users in the same group use multicast to transmit files, which ensures user service quality, ensures that each user can only be served by one base station in each time slot, and ensures that cache Under the constraint that the file of the base station will not exceed the buffer capacity of the base station, with the goal of minimizing the sum of the maximum transmission rates of users in each multicast group, a joint optimization problem of user association strategy and cache configuration strategy is established; the base station includes Acer station and small base station, M is the total number of small base stations, the network also includes a macro base station, N is the total number of files requested by all users within a period of time, and K is the total number of mobile users in the network;
(2)对用户关联策略和缓存配置策略进行迭代优化,得到最优的用户关联策略和缓存配置策略,每轮迭代优化过程包括:(2) Iteratively optimize the user association strategy and cache configuration strategy to obtain the optimal user association strategy and cache configuration strategy. Each round of iterative optimization includes:
(2.1)基于给定缓存配置策略,采用博弈论方法求解用户关联策略;(2.1) Based on the given cache configuration strategy, the game theory method is used to solve the user association strategy;
(2.2)基于给定用户关联策略,采用拉格朗日松弛算法求解缓存配置策略。(2.2) Based on the given user association policy, the Lagrangian relaxation algorithm is used to solve the cache configuration policy.
进一步地,步骤(1)中建立的联合优化问题描述为:Further, the joint optimization problem established in step (1) is described as:
s.t.C1: stC1:
C2: C2:
C3: C3:
C4: C4:
C5: C5:
其中,tmk表示用户k是否由基站m服务,xmn表示基站m中缓存文件n的比例,Rmk表示基站m到用户k的文件传输速率,SINRmk表示关联到基站m的用户k的信干噪比值,γk表示用户k的接收信干噪比阈值,Cm表示基站m的缓存容量,cn表示文件n的大小。代表基站集合,其中0代表宏基站,{1,2,...,M}代表小基站集合,代表一段时间内所有用户请求的文件集合,代表网络中移动用户集合,代表请求文件n的用户形成的一个多播组。Among them, t mk indicates whether user k is served by base station m, x mn indicates the proportion of cached file n in base station m, R mk indicates the file transfer rate from base station m to user k, and SINR mk indicates the information of user k associated with base station m. Interference-to-noise ratio value, γ k represents the received signal-to-interference-to-noise ratio threshold of user k, C m represents the buffer capacity of base station m, and cn represents the size of file n . represents the set of base stations, where 0 represents the macro base station, {1,2,...,M} represents the set of small base stations, represents the collection of files requested by all users over a period of time, represents the set of mobile users in the network, A multicast group formed on behalf of users requesting file n.
进一步地,步骤(2.1)中采用博弈论方法求解用户关联策略,包括:Further, in step (2.1), a game theory method is used to solve the user association strategy, including:
(2.1.1)设置迭代次数j=0,每个用户各个可能的关联策略被选择的初始概率向量为其中表示用户k的所有可能的关联策略集合;(2.1.1) Set the number of iterations j=0, and the initial probability vector of each possible association strategy for each user is selected as in represents the set of all possible association policies for user k;
(2.1.2)所有用户根据概率选择第bk(j)种基站关联策略进行基站关联;(2.1.2) All users according to probability Select the b k (j)th base station association strategy for base station association;
(2.1.3)各用户更新其效用函数Uk(j)及其当前归一化效用函数然后用户更新其各自概率向量,更新规则为:(2.1.3) Each user updates its utility function U k (j) and its current normalized utility function Then users update their respective probability vectors, and the update rule is:
其中,0<β<1是学习参数,用户k的效用函数定义为可服务用户k的基站的总回程成本;Among them, 0<β<1 is the learning parameter, and the utility function of user k is defined as the total backhaul cost of the base station that can serve user k;
(2.1.4)若对于都存在大于设定阈值,算法结束,否则返回步骤(2.1.2)继续循环,直到循环结束。(2.1.4) If for both exist If it is greater than the set threshold, the algorithm ends, otherwise it returns to step (2.1.2) to continue the loop until the loop ends.
进一步地,步骤(2.1)中采用博弈论方法求解用户关联策略时,用来表示用户基站关联策略的博弈论模型,其中用户集合为该博弈论模型中的博弈者集合,Sk表示表示用户k的基站关联策略集合,Uk表示用户k的效用函数;策略集合Sk={sk},其中所有用户的基站关联策略表示为矩阵其中表示所有用户的联合策略空间,除了用户k外的其他所有用户的基站关联策略为其中用户k的效用函数定义为可服务用户k的基站的总回程成本,表示为:Further, when using the game theory method to solve the user association strategy in step (2.1), use to represent the game theory model of the user base station association strategy, where the user set is the set of players in the game theory model, Sk represents the base station association strategy set of user k, U k represents the utility function of user k; the strategy set Sk = {s k }, where The base station association policies of all users are expressed as a matrix in Represents the joint policy space of all users, and the base station association policies of all users except user k are: in The utility function of user k is defined as the total backhaul cost of base stations that can serve user k, expressed as:
其中,表示可以服务用户k的基站集合,集合中基站满足约束SINRmk≥γktmk。in, represents the set of base stations that can serve user k, and the base stations in the set satisfy the constraint SINR mk ≥ γ k t mk .
进一步地,步骤(2.2)中采用拉格朗日松弛算法求解缓存配置策略的方法为:Further, in step (2.2), the method of using the Lagrangian relaxation algorithm to solve the cache configuration policy is:
采用梯度下降法对缓存配置变量xmn和拉格朗日乘子进行迭代更新至收敛得到最优解:Use gradient descent on cache configuration variables xmn and Lagrange multipliers Perform iterative updates until convergence to get the optimal solution:
其中,为更新缓存配置变量xmn的步长,为更新乘子的步长,参数j代表第j次迭代,表示:当z小于a时取值为a,当z大于b时取值为b,否则取值本身,[z]+=max{0,z}表示:当z小于0时取值为0,否则取值本身。in, Configure the step size of the variable x mn for updating the cache, to update the multiplier The step size of , the parameter j represents the jth iteration, Representation: when z is less than a, the value is a, when z is greater than b, the value is b, otherwise the value itself, [z] + =max{0,z} means: when z is less than 0, the value is 0, Otherwise the value itself.
有益效果:本发明针对异构网络,在保证所有用户的QoS前提下,结合多播和缓存技术,以最小化系统总回程成本为目标建立缓存配置策略和用户关联策略的联合优化问题。为解决该NP-hard问题,将其划分为了两个子问题,即基于给定缓存配置策略的用户关联问题和基于给定用户关联策略优化缓存配置策略的问题。本发明方法充分利用多播技术和缓存技术两者的优势,相对于现有单播缓存算法和多播缓存算法,该方法在保证系统中各用户QoS前提下,不仅可以有效降低系统回程总成本,而且在降低能耗方面也具有明显优势。Beneficial effects: The present invention is aimed at heterogeneous networks, under the premise of ensuring the QoS of all users, combined with multicast and cache technology, and aims to minimize the total backhaul cost of the system to establish a joint optimization problem of cache configuration strategy and user association strategy. In order to solve this NP-hard problem, it is divided into two sub-problems, namely the user association problem based on a given cache configuration policy and the problem of optimizing the cache configuration policy based on a given user association policy. The method of the invention makes full use of the advantages of both the multicast technology and the cache technology. Compared with the existing unicast cache algorithm and multicast cache algorithm, the method can not only effectively reduce the total cost of the system backhaul under the premise of ensuring the QoS of each user in the system , but also has obvious advantages in reducing energy consumption.
附图说明Description of drawings
图1为异构网络的系统模型图。Figure 1 is a system model diagram of a heterogeneous network.
图2为本发明的方法的流程图。Figure 2 is a flow chart of the method of the present invention.
图3为本发明与现有技术的系统功耗随SBS存储空间大小的变化而变化的仿真曲线图。FIG. 3 is a simulation graph showing the variation of the system power consumption with the variation of the SBS storage space in the present invention and the prior art.
图4-图6为本发明与现有技术的系统总回程成本随SBS存储空间大小、Zipf分布参数α、系统中用户数目的变化而变化的仿真曲线图。FIG. 4-FIG. 6 are simulation graphs showing the change of the total backhaul cost of the system according to the present invention and the prior art with changes in the size of the SBS storage space, the Zipf distribution parameter α, and the number of users in the system.
具体实施方式Detailed ways
下面结合附图对本发明的具体实施方式做详细说明。The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
如图1所示,本发明所基于的异构缓存网络由1个宏基站(Macro Base Station,MBS),M个小基站(Small cell Base Station,SBS)和K个用户组成。代表基站集合,其中0代表MBS,{1,2,...,M}代表SBS集合,代表移动用户集合。SBS密集部署,即用户可以处在多个SBS的覆盖范围之内。用矩阵表示基站和用户关联情况,其中tmk表示传输决策,当tmk=1时表示用户k由基站m服务,tmk=0时则表示用户k不由基站m提供服务,每个用户最多只能由一个基站服务。表示一段时间内所有用户请求的文件集合,文件总数目为N,每个文件大小相同,为方便,归一化为1。宏基站和小基站均具有缓存能力,缓存空间为本发明采用MDS编码缓存方案,当用户以任意顺序接收的文件编码数据包总比特数超过阈值1时,就能恢复原文件。缓存配置策略用一个(M+1)×N的矩阵来表示,其中表示基站m中缓存文件n的比例。如果基站m缓存文件n的比例为1,则文件可以从基站m直接获取,如果基站m缓存文件n的比例在0和1之间,文件n的一部分可以从基站m直接获取,另一部分则需要通过基站从核心网获取,如果基站m未缓存文件n,文件需要全部由基站从核心网获取。As shown in FIG. 1 , the heterogeneous cache network on which the present invention is based consists of one macro base station (Macro Base Station, MBS), M small cell base stations (Small cell Base Station, SBS) and K users. represents the base station set, where 0 represents the MBS, {1, 2,..., M} represents the SBS set, Represents a collection of mobile users. SBS is densely deployed, that is, users can be within the coverage of multiple SBSs. with a matrix Represents the association between the base station and the user, where t mk represents the transmission decision, when t mk = 1, it means that user k is served by base station m, and when t mk = 0, it means that user k is not served by base station m, and each user can only be served by at most A base station serves. Represents the set of files requested by all users within a period of time. The total number of files is N, and the size of each file is the same. For convenience, it is normalized to 1. Both the macro base station and the small base station have buffer capacity, and the buffer space is The present invention adopts the MDS encoding buffer scheme, and when the total number of bits of the file encoding data packets received by the user in any order exceeds the threshold value of 1, the original file can be restored. The cache configuration strategy uses a (M+1)×N matrix to indicate that Indicates the proportion of cache file n in base station m. If the ratio of file n cached by base station m is 1, the file can be obtained directly from base station m. If the ratio of file n cached by base station m is between 0 and 1, part of file n can be directly obtained from base station m, and the other part needs to be Obtained from the core network by the base station, if the base station m does not cache the file n, all the files need to be obtained by the base station from the core network.
本发明实施例中的异构网络包括1个宏基站(Macro Base Station,MBS),M=15个小基站(Small cell Base Station,SBS)和K=100个用户组成。一段时间内所有用户请求的文件总数目为N=100,每个文件大小相同,为方便,归一化为1。宏基站和小基站均具有缓存能力,MBS存储空间大小为15,SBS的存储空间大小为8。当MBS或SBS缓存了多播组请求文件时,直接通过无线链路将所请求文件多播给该多播组,当基站未缓存多播组请求文件时,基站首先通过有限回程从核心网获取文件然后再将文件多播给用户。The heterogeneous network in the embodiment of the present invention includes one macro base station (Macro Base Station, MBS), M=15 small cell base stations (Small cell Base Station, SBS), and K=100 users. The total number of files requested by all users in a period of time is N=100, and the size of each file is the same, which is normalized to 1 for convenience. Both the macro base station and the small base station have buffer capacity, the storage space size of MBS is 15, and the storage space size of SBS is 8. When the MBS or SBS caches the multicast group request file, it directly multicasts the requested file to the multicast group through the wireless link. When the base station does not cache the multicast group request file, the base station first obtains it from the core network through a limited backhaul file and then multicast the file to users.
如图2所示,本发明实施例公开的一种异构网络中基于多播技术的分布式缓存和用户关联方法,具体包括以下步骤:As shown in FIG. 2 , a distributed cache and user association method based on multicast technology in a heterogeneous network disclosed in an embodiment of the present invention specifically includes the following steps:
步骤1:基于上述异构缓存网络场景,以最小化系统总回程成本为目标,建立用户关联策略和缓存配置策略的联合优化问题。具体过程为:Step 1: Based on the above heterogeneous cache network scenario, with the goal of minimizing the total backhaul cost of the system, a joint optimization problem of user association strategy and cache configuration strategy is established. The specific process is:
首先,制定多播传输原则。First, the principle of multicast transmission is formulated.
在每个时隙,请求相同文件的用户形成一个多播组,用表示,其中qkn表示用户k是否请求文件n,当qkn=1表示用户k请求文件n,否则用户k在该时隙内没有请求文件n。同组用户采用多播方式进行文件传输;本例中假设用户请求服从参数为α=0.5的Zipf分布。At each time slot, users requesting the same file form a multicast group using where q kn indicates whether user k requests file n, when q kn =1, it indicates that user k requests file n, otherwise user k does not request file n in the time slot. Users in the same group use multicast to transmit files; in this example, it is assumed that user requests obey the Zipf distribution with a parameter of α=0.5.
然后,计算用户从各基站接收信号的信干噪比值,并确定SINR门限值。Then, the signal-to-interference-noise ratio value of the signal received by the user from each base station is calculated, and the SINR threshold value is determined.
在传输阶段,每个基站(包括MBS)复用相同频段频谱。在某个基站覆盖范围内,请求不同内容的用户组间使用不同频段频谱进行错开,即同一基站请求不同内容的用户组间不存在干扰,而请求相同内容的用户间使用相同频段频谱,由于采用多播方式传输,这些用户间也不存在干扰。如果假设基站到用户间信道是平坦衰落信道,信道增益为hmk,基站发射功率为pm,则关联到基站m的用户k的信干噪比值可以表示为:In the transmission phase, each base station (including MBS) reuses the same frequency spectrum. Within the coverage area of a certain base station, user groups requesting different content use different frequency bands to stagger, that is, there is no interference between user groups requesting different content from the same base station, but users requesting the same content use the same frequency band spectrum. Multicast transmission, there is no interference between these users. If it is assumed that the base station to the user The intermediate channel is a flat fading channel, the channel gain is h mk , and the transmit power of the base station is p m , then the signal-to-interference-to-noise ratio value of user k associated with base station m can be expressed as:
本例中,考虑一个(1000×1000)的区域,区域内只有一个MBS,位于区域中心,即MBS坐标为(500,500),SBS和用户则相对随机地撒入该区域范围内。MBS的覆盖范围半径为500m,每个SBS的覆盖范围半径为300m。MBS和SBS的发射功率分别为33dBm和23dBm,系统总带宽为10MHz,噪声功率谱密度为-174dBm/Hz。至于大尺度衰落,MBS和SBS的路径损耗模型分别为l0k=128.1+37.6log10(d0k)和其中dmk定义为用户k与基站m间的距离(km)。阴影衰落则取为标准差为8dB的对数正态(Log-normal)阴影衰落。此外,每个用户的接收信干噪比阈值γ设为-5dB。In this example, consider a (1000×1000) area, there is only one MBS in the area, which is located in the center of the area, that is, the MBS coordinates are (500,500), and SBS and users are scattered relatively randomly into the area. The coverage radius of MBS is 500m, and the coverage radius of each SBS is 300m. The transmit power of MBS and SBS is 33dBm and 23dBm respectively, the total system bandwidth is 10MHz, and the noise power spectral density is -174dBm/Hz. As for large-scale fading, the path loss models for MBS and SBS are l 0k =128.1+37.6log10(d 0k ) and where d mk is defined as the distance (km) between user k and base station m. Shadow fading is taken as log-normal shadow fading with a standard deviation of 8dB. In addition, the received signal-to-interference-to-noise ratio threshold γ of each user is set to -5dB.
接着,建立系统总回程成本表达式。Next, an expression for the total backhaul cost of the system is established.
从用户体验角度考虑,为使得用户获取文件延时较低,基站从核心网获取某一文件并多播该文件给其调度用户组时,文件下载速率需要满足该组中所有用户的传输速率需求,即基站从核心网获取文件的数据速率需至少和文件传输速率一样大,所以可定义系统回程成本为各多播组中用户的最大传输速率之和,具体可表示为:From the point of view of user experience, in order to keep the user's file acquisition delay low, when the base station acquires a file from the core network and multicasts the file to its scheduling user group, the file download rate needs to meet the transmission rate requirements of all users in the group , that is, the data rate at which the base station obtains files from the core network must be at least as large as the file transmission rate, so the system backhaul cost can be defined as the sum of the maximum transmission rates of users in each multicast group, which can be expressed as:
其中,CB表示系统总回程成本,Rmk表示基站m到用户k的文件传输速率,Rmk的具体表示为Rmk=Bmklog(1+SINRmk),其中W为系统总带宽;Among them, CB represents the total backhaul cost of the system, R mk represents the file transfer rate from base station m to user k, and R mk is specifically expressed as R mk =B mk log(1+SINR mk ), where W is the total bandwidth of the system;
最终,以最小化系统总回程成本为目标,建立用户关联策略和缓存配置策略的联合优化问题。该优化问题可描述为:Finally, with the goal of minimizing the total backhaul cost of the system, a joint optimization problem of user association strategy and cache configuration strategy is established. The optimization problem can be described as:
s.t.C1: stC1:
C2: C2:
C3: C3:
C4: C4:
C5: C5:
其中,C1约束保证用户服务质量(Quality of Service,QoS),C2约束保证每个用户每个时隙只能由一个基站提供服务,C3约束则保证缓存在基站的文件不会超过基站缓存容量;Among them, the C1 constraint guarantees the user quality of service (QoS), the C2 constraint ensures that each user can only be served by one base station per time slot, and the C3 constraint ensures that the files cached in the base station will not exceed the base station cache capacity;
步骤2:基于多播协作缓存算法迭代地对用户关联策略和缓存配置策略进行优化,得到最优的用户关联策略和缓存配置策略。上述优化问题为一个NP-hard问题。为解决该问题,首先将其划分为了两个子问题,即基于给定缓存配置策略的用户关联问题和基于给定用户关联策略优化缓存配置策略的问题;具体包括:Step 2: Iteratively optimize the user association strategy and the cache configuration strategy based on the multicast cooperative caching algorithm, and obtain the optimal user association strategy and cache configuration strategy. The above optimization problem is an NP-hard problem. In order to solve this problem, it is first divided into two sub-problems, namely the user association problem based on a given cache configuration policy and the problem of optimizing the cache configuration policy based on a given user association policy. Specifically, it includes:
步骤2.1:基于给定缓存配置策略,采用博弈论方法求解用户关联策略。具体如下Step 2.1: Based on the given cache configuration strategy, the game theory method is used to solve the user association strategy. details as follows
首先,建立博弈论模型。First, build a game theory model.
用来表示用户基站关联策略的博弈论模型,其中用户集合为该博弈论模型中的博弈者集合,Sk表示表示用户k的基站关联策略集合,Uk表示用户k的效用函数。策略集合Sk={sk},其中这样所有用户的基站关联策略可以表示为一个矩阵其中表示所有用户的联合策略空间。为满足约束C2,每个用户的每种关联策略向量中仅有一个元素为1,其余元素均为0。还可以定义除了用户k外的其他所有用户的基站关联策略为其中此外,认为当用户k既处在基站的覆盖范围内又满足优化问题中的C1约束SINRmk≥γktmk时,用户k可以选择基站m进行关联,也即用户k可以由基站m提供服务,用表示可以服务用户k的基站集合,则用户k的效用函数可以定义为可服务用户k的基站的总回程成本,即use to represent the game theory model of the user base station association strategy, where the user set is the set of players in the game theory model, S k represents the base station association strategy set of user k, and U k represents the utility function of user k. Policy set Sk = { sk }, where In this way, the base station association policies of all users can be expressed as a matrix in Represents a federated policy space for all users. In order to satisfy the constraint C2, only one element of each association strategy vector of each user is 1, and the rest are 0. It is also possible to define the base station association policy of all users except user k as in In addition, it is considered that when user k is both at the base station When the C1 constraint SINR mk ≥ γ k t mk in the optimization problem is satisfied within the coverage area of , user k can select base station m for association, that is, user k can be served by represents the set of base stations that can serve user k, then the utility function of user k can be defined as the total backhaul cost of the base stations that can serve user k, namely
定义博弈者k的归一化效用函数为可表示为:Define the normalized utility function of player k as can be expressed as:
当用户k的邻近用户已关联基站后,用户k会根据这些用户的关联策略选择合适的基站进行关联以提高其效用函数。此外,每个用户的效用函数仅取决于其自身及其邻近用户,所以该博弈可以通过分布式进行局部信息交换;After user k's neighboring users have been associated with base stations, user k will select appropriate base stations for association according to the association policies of these users to improve its utility function. In addition, the utility function of each user depends only on itself and its neighboring users, so the game can be distributed with local information exchange;
可以证明上述用户关联问题为势博弈问题。It can be proved that the above user association problem is a potential game problem.
本例中采用一种基于随机自动学习(Stochastic Learning Algorithm,SLA)的用户关联算法,其中用户通过各自的动态反馈来调整各自行为以逐步接近纳什均衡点,从而得到用户关联问题的最优解。In this example, a user association algorithm based on Stochastic Learning Algorithm (SLA) is adopted, in which users adjust their behavior through their own dynamic feedback to gradually approach the Nash equilibrium point, so as to obtain the optimal solution to the user association problem.
在基于SLA的用户关联算法中,将扩展为一个博弈的混合策略形式,并给出以下定义。首先,令用户k的所有可能的关联策略集合为集合的大小为而所有用户可能的关联策略空间为策略空间的大小为然后,令P=(p1,p2,...,pK)表示的混合策略组合,也即用户各个可能的基站关联策略被选择的概率矩阵,其中表示用户k各关联策略选择的概率向量,表示表示用户k选择第bk种关联策略的概率。In the SLA-based user association algorithm, the Extend it to a mixed-strategy form of the game and give the following definition. First, let the set of all possible association policies for user k be gather size is And the possible associated policy space for all users is The size of the policy space is Then, let P=(p 1 ,p 2 ,...,p K ) denote The mixed strategy combination of is the probability vector representing the selection of each association strategy of user k, represents the probability that user k chooses the b-th association strategy.
具体包括以下子步骤:Specifically, it includes the following sub-steps:
步骤2.1.1:初始化:设置迭代次数j=0,每个用户各个可能的关联策略被选择的初始概率向量为 Step 2.1.1: Initialization: set the number of iterations j=0, and the initial probability vector of each possible association strategy for each user is selected as
步骤:2.1.2:为各用户分配关联基站:所有用户根据现在的概率pk(j)选择第bk(j)种基站关联策略进行基站关联;Step: 2.1.2: Assign Associated Base Stations to Each User: All Users According to the current probability p k (j), select the b k (j)th base station association strategy for base station association;
步骤2.1.3:动态更新策略:各用户更新其效用函数Uk(j)及其当前归一化效用函数然后用户更新其各自概率向量,更新规则为:Step 2.1.3: Dynamic update strategy: each user updates its utility function U k (j) and its current normalized utility function Then users update their respective probability vectors, and the update rule is:
其中0<β<1是学习参数;where 0<β<1 is the learning parameter;
步骤2.1.4:判断程序结束或返回步骤6.3.2:若对于都存在接近于1,即算法结束,否则返回步骤6.3.2继续循环,直到循环结束;Step 2.1.4: Judge the program end or return to Step 6.3.2: If for both exist close to 1, i.e. The algorithm ends, otherwise return to step 6.3.2 to continue the loop until the loop ends;
步骤2.2:基于给定用户关联策略,采用拉格朗日松弛算法求解缓存配置策略。采用梯度下降法对缓存配置变量xmn和拉格朗日乘子进行更新:Step 2.2: Based on the given user association policy, use the Lagrangian relaxation algorithm to solve the cache configuration policy. Use gradient descent on cache configuration variables xmn and Lagrange multipliers To update:
其中为更新缓存配置变量xmn的充分小且固定的步长,为更新乘子充分小且固定的步长。参数j代表第j次迭代。表示:当z小于a时取值为a,当z大于b时取值为b,否则取值本身。[z]+=max{0,z}表示:当z小于0时取值为0,否则取值本身。乘子迭代更新过程将收敛至最优解。in A sufficiently small and fixed step size for updating the cache configuration variable xmn , to update the multiplier A sufficiently small and fixed step size. The parameter j represents the jth iteration. Representation: when z is less than a, the value is a, when z is greater than b, the value is b, otherwise the value itself. [z] + =max{0,z} means: when z is less than 0, the value is 0, otherwise the value itself. The multiplier iterative update process will converge to the optimal solution.
本实施例步骤2中采用多播协作缓存算法迭代地对用户关联策略和缓存配置策略进行优化,得到最优的用户关联策略和缓存配置策略。具体为:首先初始化迭代次数j=0,初始化最大迭代次数T=500、拉格朗日乘子及缓存配置策略X0,并根据初始X0,采用博弈论方法求得用户关联策略t0;然后对于每次迭代,先更新缓存配置策略Xj及拉格朗日乘子再采用博弈论方法重新计算用户关联策略tj,直到迭代结束。In
图3给出了本发明方案与现有技术方案的系统功耗随SBS存储空间大小的变化而变化的仿真曲线图,性能评价指标为系统功耗。图4-图6则给出了本发明方法方案与现有技术方案的系统总回程成本随SBS存储空间大小、Zipf分布参数、系统中用户数目的变化而变化的仿真曲线图,性能评价指标为系统总回程成本。可见,本发明的性能都要显著优于单播协作缓存方案、单播最热门缓存方案、多播为缓存方案以及多播最热门缓存方案,其不仅可以有效降低系统总回程成本,解决由于移动数据量激增和小基站密集部署带来的回程拥塞问题,而且还可以在一定程度上降低系统功耗成本。FIG. 3 shows a simulation curve diagram of the system power consumption of the solution of the present invention and the solution of the prior art as a function of the size of the SBS storage space, and the performance evaluation index is the system power consumption. Fig. 4-Fig. 6 shows the simulation curve diagrams of the total backhaul cost of the system of the method of the present invention and the prior art solution as a function of the size of the SBS storage space, the Zipf distribution parameter, and the number of users in the system. The performance evaluation index is The total backhaul cost of the system. It can be seen that the performance of the present invention is significantly better than that of the unicast cooperative cache scheme, the unicast most popular cache scheme, the multicast as the cache scheme and the multicast most popular cache scheme, which can not only effectively reduce the total backhaul cost of the system, but also solve the problems caused by mobile The backhaul congestion caused by the surge in data volume and the dense deployment of small base stations can also reduce the cost of system power consumption to a certain extent.
以上所述,仅是本发明的较佳实施例而已,并非是对本发明作任何其他形式的限制,而依据本发明的技术实质所作的任何修改或等同变化,仍属于本发明所要求保护的范围。The above are only preferred embodiments of the present invention, and are not intended to limit the present invention in any other form, and any modifications or equivalent changes made according to the technical essence of the present invention still fall within the scope of protection of the present invention. .
Claims (4)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810341885.XA CN108834080B (en) | 2018-04-17 | 2018-04-17 | Distributed caching and user association method based on multicast technology in heterogeneous networks |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810341885.XA CN108834080B (en) | 2018-04-17 | 2018-04-17 | Distributed caching and user association method based on multicast technology in heterogeneous networks |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108834080A CN108834080A (en) | 2018-11-16 |
CN108834080B true CN108834080B (en) | 2021-03-19 |
Family
ID=64154395
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810341885.XA Expired - Fee Related CN108834080B (en) | 2018-04-17 | 2018-04-17 | Distributed caching and user association method based on multicast technology in heterogeneous networks |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108834080B (en) |
Families Citing this family (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109548047B (en) * | 2018-11-21 | 2020-12-29 | 中国科学院计算技术研究所 | A physical layer caching method based on forward backhaul link capacity |
CN109547979B (en) * | 2018-12-26 | 2020-08-11 | 北京邮电大学 | Content distribution method combining energy consumption and user fairness under D2D cache network |
CN109788508B (en) * | 2018-12-29 | 2020-07-28 | 山东省计算中心(国家超级计算济南中心) | Data caching method and storage medium |
CN110430608B (en) * | 2019-06-28 | 2021-05-14 | 北京邮电大学 | Base station knowledge multicast routing scheduling method and device |
CN110730463B (en) * | 2019-09-27 | 2022-05-24 | 西北工业大学 | An Optimal Probabilistic Cache Method for Two-layer Heterogeneous Cache Networks |
CN111556510A (en) * | 2019-12-09 | 2020-08-18 | 中国人民解放军军事科学院国防科技创新研究院 | Base station data cooperation caching model and method based on content similarity |
CN113301145B (en) * | 2020-05-21 | 2022-07-08 | 北京航空航天大学 | Mobile edge cache placement method adopting request rate and dynamic property of information source issued content |
CN113630456B (en) * | 2020-08-05 | 2022-07-05 | 北京航空航天大学 | An Inter-Domain Cooperative Caching Method Based on Lagrangian Duality Decomposition and Nash Bargaining Game |
CN112187872B (en) * | 2020-09-08 | 2021-07-30 | 重庆大学 | A content caching and user association optimization method in mobile edge computing network |
CN115334545B (en) * | 2022-08-03 | 2025-07-04 | 中国人民解放军陆军工程大学 | A method for fast joint allocation of cache and access points in wireless multicast networks |
CN115529240B (en) * | 2022-08-18 | 2024-06-21 | 北京邮电大学 | A network optimization method combining content caching and transmission routing |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1624711B1 (en) * | 2004-08-03 | 2007-06-20 | Alcatel Lucent | Method for improving access network selection in the context of discontinuous coverage access networks, corresponding access controller, and corresponding multicast system |
WO2017041631A1 (en) * | 2015-09-10 | 2017-03-16 | Huawei Technologies Co., Ltd. | Communication network, devices and methods for proximity based distributed caching of service information within said network |
CN106791887A (en) * | 2016-12-02 | 2017-05-31 | 上海大学 | The distributed caching of video and transmission optimization method in wireless network |
CN106912079A (en) * | 2017-02-20 | 2017-06-30 | 北京邮电大学 | Federated user accesses selection and resource allocation methods in one kind caching heterogeneous network |
CN107708135A (en) * | 2017-07-21 | 2018-02-16 | 上海交通大学 | A kind of resource allocation methods for being applied to mobile edge calculations scene |
-
2018
- 2018-04-17 CN CN201810341885.XA patent/CN108834080B/en not_active Expired - Fee Related
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1624711B1 (en) * | 2004-08-03 | 2007-06-20 | Alcatel Lucent | Method for improving access network selection in the context of discontinuous coverage access networks, corresponding access controller, and corresponding multicast system |
WO2017041631A1 (en) * | 2015-09-10 | 2017-03-16 | Huawei Technologies Co., Ltd. | Communication network, devices and methods for proximity based distributed caching of service information within said network |
CN106791887A (en) * | 2016-12-02 | 2017-05-31 | 上海大学 | The distributed caching of video and transmission optimization method in wireless network |
CN106912079A (en) * | 2017-02-20 | 2017-06-30 | 北京邮电大学 | Federated user accesses selection and resource allocation methods in one kind caching heterogeneous network |
CN107708135A (en) * | 2017-07-21 | 2018-02-16 | 上海交通大学 | A kind of resource allocation methods for being applied to mobile edge calculations scene |
Non-Patent Citations (2)
Title |
---|
Cooperate Caching with Multicast for Mobile Edge Computing in 5G Networks;Xiangyue Huang, Zhifeng Zhao, Honggang Zhang;《2017 IEEE 85th Vehicular Technology Conference(VTC Spring)》;20170607;全文 * |
Energy-Efficient Resource Allocation for Cache-Assisted Mobile Edge Computing;Ying Cui, Wen He, Chun Ni, Chengjun Guo, Zhi Liu;《2017 IEEE 42nd Conference on Local Computer Networks(LCN)》;20171012;全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN108834080A (en) | 2018-11-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108834080B (en) | Distributed caching and user association method based on multicast technology in heterogeneous networks | |
CN109905918B (en) | NOMA cellular Internet of vehicles dynamic resource scheduling method based on energy efficiency | |
Pantisano et al. | Cache-aware user association in backhaul-constrained small cell networks | |
CN107548102B (en) | A base station caching method for minimizing user delay in an edge caching network | |
CN106686655B (en) | A Heterogeneous Network Federated User Association and Content Caching Method | |
CN109194763B (en) | A caching method based on self-organized cooperation of small base stations in ultra-dense networks | |
CN104378772B (en) | Towards the small base station deployment method of the amorphous covering of cell in a kind of cellular network | |
CN111565419A (en) | A Latency-Optimized Cooperative Edge Cache Algorithm in Ultra-Dense Networks | |
CN110290507A (en) | A caching strategy and spectrum allocation method for a D2D communication-assisted edge caching system | |
CN109673018B (en) | A Novel Content Cache Distribution Optimization Method in Wireless Heterogeneous Networks | |
He et al. | Cache-enabled coordinated mobile edge network: Opportunities and challenges | |
Zhao et al. | Joint content caching, service placement and task offloading in UAV-enabled mobile edge computing networks | |
Tran et al. | Mobee: Mobility-aware energy-efficient coded caching in cloud radio access networks | |
Wang et al. | On the design of computation offloading in cache-aided D2D multicast networks | |
CN111479312A (en) | A joint optimization method for content caching and base station dormancy in heterogeneous cellular networks | |
CN107343268A (en) | Nonopiate multicast and unicast transmission beam shaping method and system | |
CN104581918B (en) | Satellite layer-span combined optimization power distribution method based on non-cooperative game | |
CN107949007A (en) | A kind of resource allocation algorithm based on Game Theory in wireless caching system | |
CN103781166B (en) | Mobile terminal power distribution method in heterogeneous wireless network cooperative communication system | |
CN108322352A (en) | It is a kind of based on the honeycomb isomery caching method to cooperate between group | |
CN109068356A (en) | A kind of wireless cache allocation method in cognitive radio networks | |
CN108769729A (en) | Caching arrangement system based on genetic algorithm and caching method | |
CN109831759A (en) | A kind of three-dimensional D2D matching algorithm based on software definition wireless network | |
CN112543498B (en) | A Power Adaptive Allocation Method Based on Hierarchical Game Model | |
Zhang et al. | Energy efficient resource allocation in millimeter-wave-based fog radio access networks |
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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210319 |