CN113347255A - Edge server site selection deployment model and solving method thereof - Google Patents

Edge server site selection deployment model and solving method thereof Download PDF

Info

Publication number
CN113347255A
CN113347255A CN202110618740.1A CN202110618740A CN113347255A CN 113347255 A CN113347255 A CN 113347255A CN 202110618740 A CN202110618740 A CN 202110618740A CN 113347255 A CN113347255 A CN 113347255A
Authority
CN
China
Prior art keywords
edge server
base station
individuals
individual
site selection
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110618740.1A
Other languages
Chinese (zh)
Other versions
CN113347255B (en
Inventor
鲁蔚锋
马鸿宇
徐佳
徐力杰
蒋凌云
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.)
Nanjing University of Posts and Telecommunications
Original Assignee
Nanjing University of Posts and Telecommunications
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 Nanjing University of Posts and Telecommunications filed Critical Nanjing University of Posts and Telecommunications
Priority to CN202110618740.1A priority Critical patent/CN113347255B/en
Publication of CN113347255A publication Critical patent/CN113347255A/en
Application granted granted Critical
Publication of CN113347255B publication Critical patent/CN113347255B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

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
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
    • H04W16/18—Network planning tools
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

本发明提供了一种边缘服务器选址部署模型及其求解方法，所述边缘服务器选址部署模型包括基站和边缘服务器，定义i表示所述边缘服务器，j表示所述基站，每个所述边缘服务器与一个或多个所述基站连接；建立所述边缘服务器选址部署模型，以使所述边缘服务器与所述基站之间的平均延时与费用支出之和最小。本发明不但可以有效减少移动边缘服务器放置成本，而且还可以通过移动边缘服务器所放置的地理位置，从而减少边缘服务器与用户之间通信的延迟，加快服务响应。该模型为np难问题，难以在常数时间内求解，而采用遗传算法可以在常数时间内求解出最优或接近最优的解。

Figure 202110618740

The present invention provides an edge server site selection and deployment model and a solution method thereof. The edge server site selection and deployment model includes a base station and an edge server. Definition i represents the edge server, j represents the base station, and each edge server A server is connected to one or more of the base stations; a site selection and deployment model of the edge server is established, so as to minimize the sum of the average delay and the cost between the edge server and the base station. The invention can not only effectively reduce the placement cost of the mobile edge server, but also can reduce the communication delay between the edge server and the user and speed up the service response through the geographical location where the mobile edge server is placed. This model is an np-hard problem, which is difficult to solve in constant time, but the optimal or near-optimal solution can be solved in constant time by using genetic algorithm.

Figure 202110618740

Description

Edge server site selection deployment model and solving method thereof
Technical Field
The invention relates to an edge server site selection deployment model and a solving method thereof, belonging to the field of edge computing communication.
Background
With the advent of the 5G era, the target of mobile network services is no longer a pure mobile phone, but various types of devices, such as VR, tablet, automobile, and the like. The scenes of services are also increasingly diversified, such as mobile broadband, large-scale machine type communication, industrial internet, and the like. Therefore, mobile networks must meet higher requirements in terms of mobility, security, time-ductility, and reliability.
In order to meet the requirements of high bandwidth and low delay required by the high-speed development of Mobile networks and reduce the network load, Mobile Edge Computing (MEC) has come into force.
The key technologies related to mobile edge computing are mainly divided into four technologies of edge cloud placement, computing unloading, service migration and crowd sourcing cooperation. We mainly investigate the edge server placement problem in a mobile edge computing environment. Migration of computational tasks from the core network to the network edge, reduction of core network congestion and data propagation delay are the main goals of mobile edge computing. However, there is no explicit specification of where the edge cloud, i.e., the edge server, should be placed, thus creating an edge server placement problem. The problem of placing the edge server is that constraint limits of users, resources and the like are considered in a certain geographical position range, and on the premise of meeting user requirements and targets, a proper geographical position is selected for the edge server according to a certain strategy to achieve the purposes of high resource utilization rate and small network delay.
Placing edge servers in a mobile edge computing environment is a challenge. The location of the edge server is critical to the access delay of mobile users and the resource utilization of the edge server, especially in smart cities, which include hundreds or thousands of base stations through which mobile users access the edge server. Because of the large size of these networks, inefficient placement of edge servers results in lengthy access delays between edge servers and a severe imbalance in workload, some of which will be overloaded, while others are underutilized or even idle. Thus, the policy placement of the edge server will significantly improve the performance of various mobile applications, such as the access latency of the edge server.
In view of the above, it is necessary to provide an edge server addressing deployment model and a solution method thereof to solve the above problems.
Disclosure of Invention
The invention aims to provide an edge server addressing deployment model, which not only can reduce the time delay between a base station and an edge server, but also can reduce the expenditure for deploying the edge server.
In order to achieve the above object, the present invention provides an edge server site selection deployment model, where the edge server site selection deployment model includes base stations and edge servers, where i denotes the edge server, j denotes the base station, and each edge server is connected to one or more base stations; establishing an edge server site selection deployment model to minimize the sum of the average delay and the expense between the edge server and the base station, wherein the edge server site selection deployment model is as follows:
Minimize:∑i，j∈N(t*cij*xij+(1-t)*costi) (1)
Subject to∑i∈Nxij＝1 for each j∈N， (2)
xij≤yi for each i,j∈N， (3)
∑i∈N yi＝k， (4)
xij∈{0,1},for each i,j∈N， (5)
yi∈{0,1}for each i,j∈N， (6)
wherein, yi(i e N) indicates that the edge server alternate address is selected, xi,j(i, j ∈ N) indicates that the edge server is connected to the base station, cijRepresents the distance, cost, between the edge server and the base stationiRepresenting the arrangement cost of each edge server, and t represents a weight coefficient;
constraint (2) indicates that each base station can only be connected with one edge server;
constraint (3) indicates that if the edge server i is connected to the base station j, the edge server i has been selected;
the constraint condition (4) represents that the number of the alternative points of the selected edge server is k;
the constraints (5) and (6) respectively represent xij,yiIs a variable from 0 to 1.
As a further improvement of the present invention, c isijThe calculation formula of (2) is as follows:
Figure BDA0003098831370000031
wherein l(s)i) Position coordinates representing the edge server, l (b)j) Representing the location coordinates of the base station.
As a further improvement of the present invention, the level node of the edge server is undirected graph G ═ V, E; wherein V represents a set of alternate points of the edge server and nodes of the base station, and E represents a link between the base station and the edge server.
As a further refinement of the present invention, the V ═ bhtwos, where B denotes a set of nodes of the base station and S denotes a set of alternate points of the edge server; bj∈B(j∈N+) Is a node of the base station, si∈S(i∈N+) Is an alternative point of the edge server.
The invention also aims to provide a solution method for edge server addressing deployment, which can solve the edge server addressing position of the optimal solution or the approximate optimal solution.
In order to achieve the above object, the present invention provides a solution method for edge server site selection and deployment, where the solution method is applied to the edge server site selection and deployment model, and the solution method includes the following steps:
step 1, the position coordinates of the base station and the position coordinates of the edge server are calculated according to a calculation formula:
Figure BDA0003098831370000032
converting the distance into the distance between the base station and the alternative point of the edge server; wherein l(s)i) Position coordinates representing the edge server, l (b)j) Representing the position coordinates of the base station, bjE.b (j ═ 1, 2.. times, m) is the node of the base station, siE, S (i is 1,2, n) is a candidate point of the edge server;
step 2, setting genetic algorithm parameters; the genetic algorithm parameters comprise an initialized population scale, a base factor of each individual, a gene crossing probability, a gene variation probability and a maximum iteration number;
step 3, encoding the alternative points of the edge server;
step 4, initializing a population, and selecting k genes from the edge server alternative points as genes of one individual;
step 5, performing iterative evolution for specified times;
step 6, calculating the average time delay and the cost of each individual;
and 7, calculating the fitness: taking the average time delay and the reciprocal of the cost; the fitness is a scale for measuring the quality of the individual, and the number of individual reproduction or whether death is caused is determined according to the fitness;
step 8, selecting a survivor individual by adopting a roulette mode;
step 9, random cross mutation;
step 10, random mutation operation: randomly selecting variant individuals, and randomly selecting chromosome variants;
and step 11, turning to individual evaluation, and starting a new cycle: turning to step 6 to start circulation;
and step 12, finishing the circulation, and selecting the optimal individual as a solution.
As a further improvement of the present invention, in step 2, the population is a collection of individuals; the individual is a collection of genes; the gene cross probability is the partial codes of two individuals which are randomly exchanged; the genetic variation randomly changes one or more codes of one individual.
As a further improvement of the invention, the step 8 comprises the following specific steps:
step 8.1, calculate the fitness ratio, i.e. the selection probability of each individual
Figure BDA0003098831370000041
Step 8.2, calculating the cumulative probability of each individual, wherein the cumulative probability is equivalent to the 'span' on the turntable, and the larger the span is, the easier the selection is
Figure BDA0003098831370000042
Step 8.3, randomly generating r epsilon [0,1 ∈]If q isi>r, then select individual xi；
And 8.4, selecting the individuals selected by the roulette algorithm as survivors, eliminating the worst T individuals, and replacing the survivors with the optimal T individuals.
As a further improvement of the invention, the step 9 comprises the following specific steps:
step 9.1, randomly dividing individuals in the population into two groups, and randomly selecting two individuals for cross operation;
step 9.2, firstly, randomly generating two cross positions min and max (min < max) in the range of [1, sz ] (sz is the number of cities), namely in the length of the chromosome, and interchanging the [ min, max ] areas of two individuals;
and 9.3, processing the repeated part of the chromosome, and randomly mutating the repeated chromosome.
As a further improvement of the invention, a step is also arranged between the step 1 and the step 2: for the original data cijAnd costiNormalization processing is performed to map the result value to [0-1]To (c) to (d); the calculation formula of the normalization processing is as follows:
Figure BDA0003098831370000051
as a further improvement of the invention, the coding mode in step 3 is decimal coding.
The invention has the beneficial effects that: the invention not only can effectively reduce the placement cost of the mobile edge server, but also can reduce the communication delay between the edge server and the user and accelerate the service response through the geographical position of the mobile edge server. The model is an np-hard problem and is difficult to solve in constant time, and an optimal or near-optimal solution can be solved in constant time by adopting a genetic algorithm.
Drawings
FIG. 1 is a schematic structural diagram of an edge server addressing deployment model according to the present invention.
FIG. 2 is a flow chart of the genetic algorithm of the present invention.
Figure 3 is a flow chart of the roulette algorithm of the present invention.
FIG. 4 is a graphical representation of the comparison of the results of the delay and cost per edge server for the genetic algorithm of the present invention with other algorithms.
Fig. 5 is a schematic diagram of the sum of the average delay and cost of the genetic algorithm and other algorithms at the base station and the edge server in the present invention.
FIG. 6 is a diagram illustrating the results of different iterations of the genetic algorithm of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in detail with reference to the accompanying drawings and specific embodiments.
The invention discloses an edge server site selection deployment model which comprises a base station and an edge server. Defining i to represent edge servers and j to represent base stations, wherein each edge server is connected with one or more base stations; establishing an edge server site selection deployment model to minimize the sum of the average delay and the expense between an edge server and a base station, wherein the edge server site selection deployment model comprises the following steps:
Minimize:∑i,j∈N(t*cij*xij+(1-t)*costi) (1)
Subject to ∑i∈Nxij＝1 for each j∈N， (2)
xij≤yi for each i,j∈N， (3)
∑i∈N yi＝k， (4)
xij∈{0,1}，for each i，j∈N， (5)
yi∈{0，1}for each i，j∈N， (6)
wherein, yi(i e N) indicates that the edge server alternate address is selected, xi，j(i, j ∈ N) indicates that the edge server is connected to the base station, cijRepresents the distance, cost, between the edge server and the base stationiRepresents the placement cost per edge server, and t represents a weight coefficient. Constraint (2) indicates that each base station can only be connected with one edge server; constraint (3) indicates that if the edge server i is connected to the base station j, the edge server i has been selected; the constraint condition (4) represents that the number of the alternative points of the selected edge server is k; the constraints (5) and (6) respectively represent xij,yiIs a variable from 0 to 1.
The distance between the edge server and the base station is calculated according to the following formula:
Figure BDA0003098831370000061
wherein l(s)i) Position coordinates representing edge servers, l (b)j) Indicating the location coordinates of the base station. The edge server's level node can be viewed as an undirected graph G ═ V, E. Where V ═ bcu S denotes a set of alternate points of the edge server and nodes of the base station, where B denotes a set of nodes of the base station and S denotes a set of alternate points of the edge server. Definition bj∈B(j∈N+) Being a node of a base station, si∈S(i∈N+) Is an alternative point for the edge server. E denotes a link between the base station and the edge server. Each link not only represents a physical connection between a base station and a mobile edge server, but also describes a logical connection between a base station and a mobile edge server.
First a set of concepts is defined:
genotype: internal representation of the trait chromosome.
Evolution: the population is gradually adapted to the living environment, and the quality is continuously improved. The evolution of organisms has been in the form of populations.
Fitness is as follows: measure the adaptation of a species to the living environment. Selecting: several individuals are selected from the population with a certain probability. Generally, the selection process is a process based on the eligibility of fitness.
Copying: during cell division, the genetic material DNA is transferred by replication to newly produced cells, and the new cells inherit the genes of the old cells.
And (3) crossing: the DNA of the same position of the two chromosomes is cut off, and the two front and rear strings are respectively combined in a crossed manner to form two new chromosomes. Also known as gene recombination or hybridization; the gene crossover probability encodes the random exchange of portions of two individuals.
Mutation: some replication errors may occur (with little probability) during replication, and the mutation generates a new chromosome and shows a new trait. Genetic variation is the random alteration of one or several codes of an individual.
And (3) encoding: the genetic information in DNA is arranged in a pattern over a long strand. Genetic coding can be viewed as a mapping from phenotype to genotype.
Individual: refers to an entity with characteristics of chromosome, namely, a gene set.
Population: a collection of individuals, the number of individuals in the collection being referred to as a population.
The solution method for edge server addressing deployment will be described in detail below.
The concrete solving method comprises the following steps: step 1, the position coordinates of the base station and the position coordinates of the edge server are calculated according to a calculation formula:
Figure BDA0003098831370000071
converting the distance into the distance between the base station and the alternative point of the edge server; wherein l(s)i) Position coordinates representing edge servers, l (b)j) Bits representing base stationsSet coordinates, bjE.b ( j 1, 2.. multidata., m) is a node of the base station, siE, S (i is 1,2, n) is an alternative point of the edge server;
step 2, setting genetic algorithm parameters; the genetic algorithm parameters comprise an initialized population scale, a base factor of each individual, a gene crossing probability, a gene variation probability and a maximum iteration number;
step 3, encoding the alternative points of the edge server;
step 4, initializing a population, and selecting k genes from the edge server alternative points as genes of one individual;
step 5, performing iterative evolution for specified times;
step 6, calculating the average time delay and the cost of each individual;
and 7, calculating the fitness: taking the average time delay and the reciprocal of the cost; the fitness is a scale for measuring the quality of the individual, and the number of individual reproduction or whether death is caused is determined according to the fitness;
and 8, selecting the surviving individuals by adopting a roulette mode:
(8.1) calculating the fitness ratio, i.e., the selection probability of each individual
Figure BDA0003098831370000081
(8.2) calculating the cumulative probability of each individual, wherein the larger the span is, the easier the selection is
Figure BDA0003098831370000082
(8.3) randomly generating r ∈ [0,1 ]]If q isi>r, then select individual xi；
(8.4) the individuals selected by the roulette algorithm were surviving, the worst T individuals were eliminated, and the best T individuals were substituted.
Step 9, random cross mutation:
(9.1) randomly dividing individuals in the population into two groups, and randomly selecting two individuals to carry out cross operation; (9.2) first randomly generating two crossover sites min and max (min < max) within the range of [1, sz ] (sz is the number of cities), i.e., within the length of the chromosome, and interchanging the [ min, max ] regions of the two individuals;
(9.3) processing the repeated part of the chromosome, and randomly mutating the repeated chromosome.
Step 10, random mutation operation: randomly selecting variant individuals, and randomly selecting chromosome variants;
and step 11, turning to individual evaluation, and starting a new cycle: turning to step 6 to start circulation;
and step 12, finishing the circulation, and selecting the optimal individual as a solution.
Specifically, as shown in fig. 1, initially, the candidate points of the mobile edge server and the topology of the base station network have a total of 7 base stations and 5 deployment candidate points of the mobile edge server. Suppose we need to select 3 points as actual deployment points from the deployment alternatives of 5 mobile edge servers. Each base station and mobile edge server candidate point have a coordinate (x, y), and each mobile edge server candidate point has a deployment costi. The Euclidean distance c between each base station and each mobile edge server is calculated by using formula (7)ij。
If there is a normalization requirement on the data, then cijAnd costiAnd (6) carrying out normalization processing. As shown in fig. 2, the genetic algorithm parameters are determined according to step 2. Namely, determining the maximum iteration number m, the initial population size num, each individual gene factor k, the gene cross probability cross and the gene mutation probability mutate. As shown in fig. 6, the maximum number of iterations may affect the solution result.
The decimal encoding of gene in step 3, for example [0,1,2] represents that the individual has selected edge server alternate points 0,1, 2. And creating num x k two-dimensional arrays as an initialization population array in the step 4. And randomly initializing each individual gene, and randomly selecting k points from alternative points of the edge server as the genes of one individual. And then, in steps 5 to 7, traversing the sum of the shortest time delay and the cost from each base station to the edge server, and then solving the sum of the average time delay and the cost of the individual. And then sorting the individuals from large to small according to the sum of the average time delay and the cost.
The surviving individuals are then calculated by the roulette algorithm at step 8, as shown in figure 3. The worst n individuals are eliminated and replaced by the best n individuals.
Randomly swapping genes of any individual in step 9, for example: the exchange of [0,1,2] with [1,3,4] yields [0,2,3] and [1,1,4 ]. If a chromosome duplication occurs, the chromosome is mutated, for example: [1,1,4] to [1,2,4 ]. Step 10 random chromosomal variation of random individuals is performed. A certain gene segment of a certain individual is randomly selected and then randomly altered, for example: the [1,2,4] variation is [1,2,3 ]. And continuing to calculate the fitness and starting the next iteration until reaching the specified iteration times. And outputting the optimal individuals. I.e., edge server deployment scenarios.
As shown in fig. 4 and fig. 5, through comparing different algorithms by experiments, it is shown that the present invention not only can solve the optimal solution or the near optimal solution of the np problem within a constant time, but also can reduce the average delay and cost between the edge server and the base station at the edge server deployment stage. Meanwhile, the problem that the problem is difficult to solve in constant time in the aspect of time complexity is avoided.
In summary, the invention can effectively reduce the placement cost of the mobile edge server by setting an edge server addressing deployment model and a solving method thereof, and can reduce the communication delay between the edge server and the user and accelerate the service response by moving the geographical position where the edge server is placed. The model is an np-hard problem and is difficult to solve in constant time, and an optimal or near-optimal solution can be solved in constant time by adopting a genetic algorithm.
Although the present invention has been described in detail with reference to the preferred embodiments, it will be understood by those skilled in the art that various changes may be made and equivalents may be substituted for elements thereof without departing from the spirit and scope of the present invention.

Claims (10)

1.一种边缘服务器选址部署模型，其特征在于：所述边缘服务器选址部署模型包括基站和边缘服务器，定义i表示所述边缘服务器，j表示所述基站，每个所述边缘服务器与一个或多个所述基站连接；建立所述边缘服务器选址部署模型，以使所述边缘服务器与所述基站之间的平均延时与费用支出之和最小，所述边缘服务器选址部署模型如下：1. An edge server site selection and deployment model, characterized in that: the edge server site selection and deployment model includes a base station and an edge server, and definition i represents the edge server, j represents the base station, and each edge server is associated with the edge server. One or more of the base stations are connected; the edge server site selection and deployment model is established to minimize the sum of the average delay and the cost between the edge server and the base station, and the edge server site selection and deployment model as follows: Minimize:∑i，j∈N(t*cij*xij+(1-t)*costi) (1)Minimize:∑ i, j∈N (t*c ij *x ij +(1-t)*cost i ) (1) Subject to ∑i∈Nxij＝1 for each j∈N， (2)Subject to ∑ i∈N x ij =1 for each j∈N, (2) xij≤yi for each i,j∈N， (3)x ij ≤y i for each i,j∈N, (3) ∑i∈Nyi＝k， (4)∑ i∈N y i =k, (4) xij∈{0,1}，for each i，j∈N， (5)x ij ∈ {0,1}, for each i, j ∈ N, (5) yi∈{0,1}for each i,j∈N， (6)y i ∈{0,1}for each i,j∈N, (6) 其中，yi(i∈N)表示所述边缘服务器备选地址被选中，xi，j(i，j∈N)表示所述边缘服务器与所述基站相连，cij表示所述边缘服务器与所述基站之间的距离，costi表示每个所述边缘服务器布置费用，t表示权重系数；Wherein, y i (i∈N) indicates that the alternative address of the edge server is selected, x i,j (i, j∈N ) indicates that the edge server is connected to the base station, and cij indicates that the edge server is connected to the base station. The distance between the base stations, cost i represents the cost of each edge server arrangement, t represents the weight coefficient; 约束条件(2)表示每个基站只能与一个边缘服务器相连；Constraint (2) means that each base station can only be connected to one edge server; 约束条件(3)表示若边缘服务器i与基站j相连,则边缘服务器i已被选中；Constraint (3) means that if edge server i is connected to base station j, then edge server i has been selected; 约束条件(4)表示选中的边缘服务器的备选点数量为k；Constraint (4) indicates that the number of candidate points of the selected edge server is k; 约束条件(5)、(6)分别表示xij,yi为0-1变量。Constraints (5) and (6) respectively indicate that x ij , y i are 0-1 variables. 2.根据权利要求1所述的边缘服务器选址部署模型，其特征在于：所述cij的计算公式为：2. The edge server site selection and deployment model according to claim 1, wherein: the calculation formula of the c ij is:
Figure FDA0003098831360000011
Figure FDA0003098831360000011
 其中，l(si)表示所述边缘服务器的位置坐标，l(bj)表示所述基站的位置坐标。Wherein, l(s i ) represents the location coordinates of the edge server, and l(b j ) represents the location coordinates of the base station. 3.根据权利要求2所述的边缘服务器选址部署模型，其特征在于：所述边缘服务器的级别节点为无向图G＝(V，E)；其中，V表示所述边缘服务器的备选点和所述基站的节点的集合，E表示所述基站与边缘服务器之间的链路。3. The edge server site selection and deployment model according to claim 2, wherein: the level node of the edge server is an undirected graph G=(V, E); wherein, V represents the alternative of the edge server A set of nodes and nodes of the base station, E represents the link between the base station and the edge server. 4.根据权利要求3所述的边缘服务器选址部署模型，其特征在于：所述V＝B∪S，其中，B表示所述基站的节点的集合，S表示所述边缘服务器的备选点的集合；bj∈B(j∈N+)为所述基站的节点，si∈S(i∈N+)为所述边缘服务器的备选点。4. The edge server site selection and deployment model according to claim 3, wherein: the V=B∪S, wherein B represents the set of nodes of the base station, and S represents the candidate point of the edge server The set of ; b j ∈B(j∈N + ) is the node of the base station, and s i ∈S(i∈N + ) is the candidate point of the edge server. 5.一种边缘服务器选址部署的求解方法，其特征在于：所述求解方法应用于权利要求1-4中任一项所述的边缘服务器选址部署模型，所述求解方法包括如下步骤：5. A solution method for edge server site selection and deployment, characterized in that: the solution method is applied to the edge server site selection and deployment model described in any one of claims 1-4, and the solution method comprises the steps: 步骤1，将基站的位置坐标与边缘服务器的位置坐标通过计算公式：
Figure FDA0003098831360000021
转化为基站与边缘服务器的备选点之间的距离；其中，l(si)表示所述边缘服务器的位置坐标，l(bj)表示所述基站的位置坐标，bj∈B(j＝1,2,...,m)为所述基站的节点，si∈S(i＝1,2,...,n)为所述边缘服务器的备选点；Step 1. Calculate the location coordinates of the base station and the edge server through the calculation formula:
Figure FDA0003098831360000021
It is converted into the distance between the base station and the candidate point of the edge server; wherein, l(s i ) represents the location coordinates of the edge server, l(b j ) represents the location coordinates of the base station, and b j ∈ B(j =1,2,...,m) are the nodes of the base station, s i ∈S (i=1,2,...,n) are the candidate points of the edge server; 步骤2，设置遗传算法参数；所述遗传算法参数包括初始化种群规模、每个个体的基因数、基因交叉概率、基因变异概率以及最大迭代次数；Step 2, setting genetic algorithm parameters; the genetic algorithm parameters include initialized population size, number of genes per individual, gene crossover probability, gene mutation probability and the maximum number of iterations; 步骤3，对边缘服务器的备选点进行编码；Step 3, encoding the candidate points of the edge server; 步骤4，初始化种群，从边缘服务器备选点中选取k个作为其中一个个体的基因；Step 4, initialize the population, and select k genes from the edge server candidate points as one of the individual genes; 步骤5，进行指定次数迭代进化；Step 5, perform iterative evolution for a specified number of times; 步骤6，计算每个个体的平均时延与费用成本；Step 6: Calculate the average delay and cost of each individual; 步骤7，计算适应度：取平均时延和费用成本的倒数；所述适应度为衡量个体优劣的尺度，根据所述适应度决定个体繁殖的数量、或者决定是否消亡；Step 7: Calculate the fitness: take the inverse of the average delay and the cost; the fitness is a scale for measuring the pros and cons of an individual, and the number of individual reproduction or whether to die is determined according to the fitness; 步骤8，采用轮盘赌方式选择幸存个体；Step 8, select surviving individuals by roulette; 步骤9，随机交叉变异；Step 9, random crossover mutation; 步骤10，随机变异运算：随机选取变异个体，随机选择染色体变异；Step 10, random mutation operation: randomly select mutant individuals, and randomly select chromosomal mutations; 步骤11，转向个体评价，开始新一轮循环：转向步骤6开始循环；Step 11, turn to individual evaluation, and start a new cycle: turn to step 6 to start the cycle; 步骤12，循环结束，选出最优个体作为解。Step 12, the cycle ends, and the optimal individual is selected as the solution. 6.根据权利要求5所述的边缘服务器选址部署的求解方法，其特征在于：在步骤2中，所述种群为个体的集合；所述个体为基因的集合；所述基因交叉概率为随机交换两个个体的部分编码；所述基因变异为随机改变某个体的某个或几个编码。6. The solution method for edge server location and deployment according to claim 5, characterized in that: in step 2, the population is a set of individuals; the individuals are a set of genes; the gene crossover probability is random Swap parts of the codes of two individuals; the genetic variation is the random change of one or several codes of an individual. 7.根据权利要求5所述的边缘服务器选址部署的求解方法，其特征在于，步骤8包括如下具体步骤：7. The solution method of edge server site selection and deployment according to claim 5, is characterized in that, step 8 comprises the following concrete steps: 步骤8.1，计算适应度比例，即每个个体的选择概率
Figure FDA0003098831360000031
Step 8.1, calculate the fitness ratio, that is, the selection probability of each individual
Figure FDA0003098831360000031
 步骤8.2，计算每个个体的累积概率，相当于转盘上的“跨度”，“跨度”越大，越容易选到
Figure FDA0003098831360000032
Step 8.2, calculate the cumulative probability of each individual, which is equivalent to the "span" on the turntable. The larger the "span", the easier it is to choose.
Figure FDA0003098831360000032
 步骤8.3，随机生成r∈[0,1]，若qi>r，则选择个体xi；Step 8.3, randomly generate r∈[0,1], if qi >r , select individual xi ; 步骤8.4，由轮盘赌算法选择下来的个体为幸存个体，淘汰最差的T个个体，由最优的T个个体替代。In step 8.4, the individuals selected by the roulette algorithm are surviving individuals, and the worst T individuals are eliminated and replaced by the optimal T individuals. 8.根据权利要求5所述的边缘服务器选址部署的求解方法，其特征在于，步骤9包括如下具体步骤：8. the solution method of edge server site selection and deployment according to claim 5, is characterized in that, step 9 comprises the following concrete steps: 步骤9.1，将种群中个体随机分成两组，随机选择两个个体进行交叉操作；Step 9.1: Randomly divide the individuals in the population into two groups, and randomly select two individuals for crossover operation; 步骤9.2，首先在[1，sz](sz为城市数数量)的范围内，即在染色体长度内，随机产生两个交叉位min和max(min<max)，将两个个体的[min，max]区域互换；Step 9.2, first in the range of [1, sz] (sz is the number of cities), that is, within the length of the chromosome, randomly generate two crossover bits min and max (min<max), and combine the [min, max] area swap; 步骤9.3，处理染色体重复部分，随机变异重复的染色体。Step 9.3, dealing with repetitive parts of chromosomes, randomly mutating repetitive chromosomes. 9.根据权利要求5所述的边缘服务器选址部署的求解方法，其特征在于，步骤1和步骤2之间还设有步骤：对原始数据cij和costi进行归一化处理，使结果值映射到[0-1]之间；所述归一化处理的计算公式为：9. the solution method of edge server location deployment according to claim 5, is characterized in that, between step 1 and step 2 is also provided with step: carry out normalization processing to original data c ij and cost i , make the result The value is mapped between [0-1]; the calculation formula of the normalization process is:
Figure FDA0003098831360000041
Figure FDA0003098831360000041
 10.根据权利要求5所述的边缘服务器选址部署的求解方法，其特征在于：步骤3中所述编码的方式为十进制编码。10 . The solution method for location selection and deployment of edge servers according to claim 5 , wherein the encoding method in step 3 is decimal encoding. 11 .
CN202110618740.1A 2021-06-03 2021-06-03 Edge server location deployment model and its solution method Active CN113347255B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110618740.1A CN113347255B (en) 2021-06-03 2021-06-03 Edge server location deployment model and its solution method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110618740.1A CN113347255B (en) 2021-06-03 2021-06-03 Edge server location deployment model and its solution method

Publications (2)

Publication Number Publication Date
CN113347255A true CN113347255A (en) 2021-09-03
CN113347255B CN113347255B (en) 2023-07-11

Family

ID=77473504

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110618740.1A Active CN113347255B (en) 2021-06-03 2021-06-03 Edge server location deployment model and its solution method

Country Status (1)

Country Link
CN (1) CN113347255B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114363962A (en) * 2021-12-07 2022-04-15 重庆邮电大学 Collaborative edge server deployment and resource scheduling method, storage medium and system
CN114423023A (en) * 2021-12-08 2022-04-29 重庆邮电大学 5G network edge server deployment method facing mobile users

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107318058A (en) * 2017-08-07 2017-11-03 广东电网有限责任公司电力调度控制中心 ONU dispositions methods in power distribution communication net based on Optimum cost and load balancing
CN111601327A (en) * 2020-05-14 2020-08-28 北京赛宝工业技术研究院有限公司 Service quality optimization method and device, readable medium and electronic equipment
CN112073452A (en) * 2020-05-27 2020-12-11 河南工程学院 An energy-efficient and resource-constrained task allocation method for mobile edge computing

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107318058A (en) * 2017-08-07 2017-11-03 广东电网有限责任公司电力调度控制中心 ONU dispositions methods in power distribution communication net based on Optimum cost and load balancing
CN111601327A (en) * 2020-05-14 2020-08-28 北京赛宝工业技术研究院有限公司 Service quality optimization method and device, readable medium and electronic equipment
CN112073452A (en) * 2020-05-27 2020-12-11 河南工程学院 An energy-efficient and resource-constrained task allocation method for mobile edge computing

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114363962A (en) * 2021-12-07 2022-04-15 重庆邮电大学 Collaborative edge server deployment and resource scheduling method, storage medium and system
CN114423023A (en) * 2021-12-08 2022-04-29 重庆邮电大学 5G network edge server deployment method facing mobile users
CN114423023B (en) * 2021-12-08 2024-02-02 广州大鱼创福科技有限公司 Mobile user-oriented 5G network edge server deployment method

Also Published As

Publication number Publication date
CN113347255B (en) 2023-07-11

Similar Documents

Publication Publication Date Title
CN104995870B (en) Multiple target server arrangement determines method and apparatus
JP3449923B2 (en) Network topology design apparatus, network topology design method, and recording medium recording network topology design program
CN107679750B (en) Cloud manufacturing service resource matching method based on adaptive coefficient genetic algorithm
CN106227599B (en) Method and system for resource scheduling in a cloud computing system
CN111694664B (en) A computing offload distribution method for edge servers
CN111984419B (en) A Computing Migration Method for Complex Tasks with Reliability Constraints in Edge Environments
CN108665089B (en) A Robust Optimization Model Solving Method for Location Problems
CN113347255A (en) Edge server site selection deployment model and solving method thereof
CN109298930B (en) A cloud workflow scheduling method and device based on multi-objective optimization
CN110008023B (en) Budget Constrained Random Task Scheduling Method for Cloud Computing System Based on Genetic Algorithm
Zhao et al. Optimize the placement of edge server between workload balancing and system delay in smart city
CN108989122A (en) Virtual network requests mapping method, device and realization device
CN113285832A (en) NSGA-II-based power multi-mode network resource optimization allocation method
CN112882723B (en) Edge service deployment method for parallel micro-service combination
Ramadan et al. A memetic optimization algorithm for multi-constrained multicast routing in ad hoc networks
CN108289115B (en) Information processing method and system
CN116170442A (en) A Computing Offloading Method Based on Improved Coral Reef Algorithm in Cloud-Edge Collaborative Computing
CN109889573B (en) NGSA multi-target-based copy placement method in hybrid cloud
CN110910261A (en) An adversarial enhancement method for network community detection based on multi-objective optimization
CN115016889A (en) A virtual machine optimization scheduling method for cloud computing
CN111465057B (en) Edge caching method and device based on reinforcement learning and electronic equipment
CN118018492A (en) Adaptive traffic distribution method and system based on clustering and quantum computing
CN114640643B (en) Information cross-community propagation maximization method and system based on group intelligence
CN110175287A (en) A kind of matrix decomposition implicit feedback recommended method and system based on Flink
CN116094862A (en) Scale-free Internet of things robust networking method based on quantum multi-universe co-evolution

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