CN110809275A - Micro cloud node placement method based on wireless metropolitan area network - Google Patents
Micro cloud node placement method based on wireless metropolitan area network Download PDFInfo
- Publication number
- CN110809275A CN110809275A CN201911085727.3A CN201911085727A CN110809275A CN 110809275 A CN110809275 A CN 110809275A CN 201911085727 A CN201911085727 A CN 201911085727A CN 110809275 A CN110809275 A CN 110809275A
- Authority
- CN
- China
- Prior art keywords
- particle
- value
- particles
- micro cloud
- optimal
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 26
- 239000002245 particle Substances 0.000 claims abstract description 143
- 230000004044 response Effects 0.000 claims abstract description 26
- 238000005457 optimization Methods 0.000 claims abstract description 10
- 230000006870 function Effects 0.000 claims description 13
- 230000019771 cognition Effects 0.000 claims description 12
- 230000005540 biological transmission Effects 0.000 claims description 10
- 230000001934 delay Effects 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 230000002068 genetic effect Effects 0.000 claims description 6
- 230000035772 mutation Effects 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000003247 decreasing effect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000001149 cognitive effect Effects 0.000 description 2
- 239000002131 composite material Substances 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000012804 iterative process Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 238000007667 floating Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000002028 premature Effects 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
Images
Classifications
-
- 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
-
- 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
- H04W16/22—Traffic simulation tools or models
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W64/00—Locating users or terminals or network equipment for network management purposes, e.g. mobility management
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W84/00—Network topologies
- H04W84/02—Hierarchically pre-organised networks, e.g. paging networks, cellular networks, WLAN [Wireless Local Area Network] or WLL [Wireless Local Loop]
- H04W84/10—Small scale networks; Flat hierarchical networks
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention relates to a micro cloud node placement method based on a wireless metropolitan area network. Selecting a K access point in a wireless metropolitan area network to place a micro cloud node, directly coding particles into access point serial numbers, initializing a batch of particles at random based on a particle swarm optimization algorithm, calculating the fitness value of each particle through a fitness function, recording the fitness value as the optimal particle value of the particle, selecting an individual with the optimal fitness value as an optimal population particle, recording the fitness value of the particle as the optimal population particle value, obtaining new particles through variation operation and cross operation on the excellent individual during each iteration, selecting the optimal value of each particle, entering the next generation together with the optimal population particle value, repeating the iteration to the preset threshold number of times, and completing the placement and search of the micro cloud node. According to the invention, by optimizing the placement of the micro cloud nodes, the average response time of the user task is better reduced, and high-efficiency and low-delay service is obtained.
Description
Technical Field
The invention relates to a micro cloud node placement method based on a wireless metropolitan area network.
Background
With the advent of everything interconnection, network connection objects are expanding from people to things. The intelligent devices such as mobile phones and pads have increasingly greater functions in the fields of daily life, learning, social contact, work and the like. Meanwhile, various sensors, intelligent instruments, intelligent cameras and the like are widely applied to industries such as industry and agriculture, medical treatment, education, traffic, intelligent home, environmental protection and the like. Statistically, as of 2017, mobile cloud traffic already accounts for 84% of the entire mobile traffic. According to the prediction of IDC, the total amount of global mobile data reaches 40000EB in 2020, the annual composite growth rate reaches 36%, and the mobile cloud flow accounts for 94% of the whole mobile flow; the growth rate of Chinese internet data flow is more prominent, the Chinese internet data flow reaches 8806EB in 2020, which accounts for 22% of global data yield, and the annual composite growth rate reaches 49%.
Communication flow of the mobile terminal shows explosive growth, meanwhile, requirements of users and enterprises on various aspects of terminal use fluency, user experience and the like are increasingly increased, mobile application programs become more and more computing-intensive, computing power requirements are increased, and computing power, battery life, storage capacity and the like of the mobile device are limited due to the size of the mobile device, especially the electric quantity endurance.
Edge Computing (Edge Computing) takes place at the end of the run. OEC gives a definition of the edge calculation: edge computing provides a small data center, i.e., an edge node, at a location close to the user. The micro cloud is a trusted computer or a computer cluster which is rich in resources, can be connected to a high-speed internet and can provide services for the mobile equipment, and by applying the micro cloud technology, the battery pressure of the mobile equipment can be relieved, the data storage and processing capacity can be improved, and the ultra-low time delay transmission of information can be realized. Compared with Mobile Edge Computing (MEC) and fog computing, micro-clouds are mainly used for mobile enhancement, can provide rich computing resources for mobile devices, especially for video analysis applications focusing on edges, and can extract tags and metadata of edge data and transmit the tags and metadata to the clouds to achieve efficient global search.
The user unloads the calculation-intensive tasks to the nearest cloudless, so that the time delay can be obviously shortened, and the cost is reduced. The user has the advantages of short time delay, single hop, high bandwidth, low cost and the like when accessing the micro cloud, can obtain real-time response, and is one of effective methods for reducing the system response time of programs in the mobile equipment. However, compared with the cloud computing center, the micro cloud has the following disadvantages: the cloudlets can only be accessed through Wi-Fi access points with small coverage, and compared with cloud data centers, the computing resources are obviously insufficient,
although a great deal of research is already carried out on the aspects of micro cloud computing and micro cloud offloading technologies at present, little attention is paid to the placement of micro cloud nodes, and how to further optimize the problem of the placement of the micro cloud nodes so as to improve the performance of mobile applications still needs to be deeply researched.
Disclosure of Invention
The invention aims to provide a micro cloud node placement method based on a wireless Metropolitan Area Network (MAN) aiming at the complex network environment in the MAN under the edge environment.
In order to achieve the purpose, the technical scheme of the invention is as follows: a method for placing micro cloud nodes based on a wireless metropolitan area network includes the steps of selecting K access points in the wireless metropolitan area network to place the micro cloud nodes, recording access point serial numbers in a particle set, directly coding particles into the access point serial numbers, initializing a batch of particles randomly based on a particle swarm optimization algorithm, calculating a fitness value of each particle through a fitness function, marking the fitness value as a particle optimal value of the particle, selecting an individual with the optimal fitness value as a population optimal particle, marking the fitness value of the particle as a population particle optimal value, obtaining new particles through variation operation and cross operation on the excellent individual during each iteration, selecting the self optimal value of each particle, enabling the self optimal value of each particle and the population particle optimal value to enter the next generation, repeating iteration to preset threshold times, and completing placing and searching of the micro cloud nodes.
In an embodiment of the present invention, the method is specifically implemented as follows:
step S1, particle encoding:
selecting a K access point in a wireless metropolitan area network to place a micro-cloud node, namely the coding length of a particle is K, recording the access point serial number in a particle set, directly coding the particle into the access point serial number, and increasing progressively according to the serial number, wherein the value does not exceed the maximum value of the access point serial number;
step S2, fitness function:
calculating the average response time of the particles by using the following formula to be recorded as the fitness function value of the current particles, wherein the shorter the average response time of the particles is, the more excellent the particles are;
wherein:
ti＝wij+Djk+tcloudlet(k)
tcloudlet(k)＝FuncQ(Λ(k))+1/μ
FuncQ(λ)＝C(c,λ/μ)/(cμ-λ)
W＝{wiji0 ≦ i < n,0 ≦ j < m } representing a set of wireless delays between the user and the wirelessly connected access pointAnd D ═ DjkJ is more than or equal to |0, and k is less than m, and represents a set of transmission delays among the access points; mu represents the service rate of the server in the micro cloud; funcQ(lambda) obtaining the queue time of the micro cloud under the load according to the sum lambda of the task arrival rates and the working capacity of the micro cloud; Λ (k) represents the sum of the arrival rates of tasks scheduled to cloudlet k; useriRepresenting a set of users; c (C, rho) calculation formula is a classic Erlang queue formula, C represents the number of servers, and rho represents the processing time of each server under the corresponding task;
step S3, particle iteration:
the improved PSO-GA algorithm particle is adopted to introduce the cross mutation operation of genetic operators, and the updating strategy formula is as follows:
wherein t represents the current iteration number; xi tRespectively representing the position of the ith particle at the time of the t iteration; w is an inertial weight, representing the ability of the particle to maintain the current velocity; p _ besti tAnd g _ besttRespectively representing the historical optimal value of the particle and the historical optimal value of the population after the particle is iterated for t times; c. C1And c2The learning factor can control the learning ability of the particles to the self historical optimal value and the population historical optimal value; cgThe cross operation of the particles and the historical optimal value of the population is represented; cpA crossover operation representing a particle and the historical optimum of the particle itself;
the particles can go through three processes of variation of the particle individuals, individual cognition crossing and social population cognition crossing in the updating process, so that the particles are searched towards the optimal solution; the particle individual variation formula, the individual cognition cross formula and the social population cognition cross formula are respectively as follows:
step S4, because the micro cloud node placement problem is a nonlinear problem, the inertia weight w formula is adjusted to
Is to solve for the ratio of the different values between the particle and the optimal solution in the particle length.
Compared with the prior art, the invention has the following beneficial effects: the method of the invention considers the relation between the user and the access point in the complex wireless metropolitan area network environment, finds a reasonable and proper micro cloud node placing scheme to bring higher cost performance in the research and verification of the reasonable number of the micro cloud nodes, and combines the advantages of the genetic algorithm and the particle swarm optimization algorithm to generate the micro cloud node placing result; the method provided by the invention can better reduce the average response time of the user task under the condition of a reasonable number of micro clouds by optimizing the placement of the micro cloud nodes, obtain high-efficiency and low-delay service, accord with the user density rule, and simultaneously obtain optimization on the basis.
Drawings
Fig. 1 is a simplified schematic diagram of a wireless metropolitan area network.
Fig. 2 is an example of particle encoding.
FIG. 3 shows the variation of the particles.
Fig. 4 is a particle intersection.
Fig. 5 shows the average response time of the system under different numbers of micro cloud nodes.
Fig. 6 shows three types of micro-cloud node placements when K is 8.
Detailed Description
The technical scheme of the invention is specifically explained below with reference to the accompanying drawings.
The invention provides a micro cloud node placement method based on a wireless metropolitan area network, which selects K access points in the wireless metropolitan area network to place micro cloud nodes, records the serial numbers of the access points in a particle set, directly coding the particles into access point serial numbers, randomly initializing a batch of particles based on a particle swarm optimization algorithm, calculating the fitness value of each particle through a fitness function, recording the fitness value as the optimal value of the particle, simultaneously selecting the individual with the optimal fitness value as the optimal population particle, recording the fitness value of the particle as the optimal population particle value, meanwhile, in each iteration, the particles are subjected to mutation operation and cross operation on excellent individuals to obtain new particles, the self optimal value of each particle is selected, and entering the next generation together with the optimal value of the population particles, and repeating iteration until the number of times of a preset threshold value is reached to complete the placement and search of the micro cloud nodes.
The following is a specific implementation of the present invention.
1. Design mode
The wireless metropolitan area network comprises a plurality of access point sets (AP) which can be connected with each other through the internet and a plurality of User sets (User) which can be accessed to the wireless network through the wireless access points, the association relationship between users and the access points in the wireless metropolitan area network can be represented by an undirected graph G (V, E), wherein the V (AP) ∪ User represents all the access points and the User sets, and the E comprises two types of edges, one is two Access Points (AP), and the other is two Access Points (AP)iAnd APjConnection (AP) betweeni,APj) The two access points are directly connected through wired connection and have a low-delay high-speed network transmission path; the other side is the user i and the access point APjBetween (User)i,APj) The user may connect to the network through the access point. Meanwhile, we define that the graph formed by the access points is a non-directional connected graph, which means that any access point contained in G is connected with the graphAnother access point may be accessed through the high speed internet.
Fig. 1 is a simple wireless metropolitan area network system model, all access points are connected with each other through a wired high-speed network to form a directed connection diagram, and an access point can be directly connected with all users in the coverage range of the access point, so that a user can be connected with a plurality of access points, for example, a user in fig. 1 can select an access point 4 and an access point 6. In a wireless metropolitan area network system, the task amount generated by a mobile user is floating and can not be accurately predicted, especially when a plurality of applications are operated at the same time, so that the average task amount of the user in a certain time period is defined as the task amount of the user at a certain moment, and the task is a bundle of unloadable task flows, and the task randomly enters the system by a poisson process of an arrival rate lambdai.
For W ═ WijI 0 ≦ i < n,0 ≦ j < m represents the set of wireless delays between the user and his several wirelessly connected access points. For D ═ DjkJ is less than or equal to |0, and k is less than m, and represents the set of transmission delays between access points.
In a wireless network system, delay required by task transmission is generated firstly, a user in the network transmits task requirements to the micro cloud through an access point directly connected with the user, and the task of the user is unloaded to the micro cloud for execution. Assuming that the offloaded tasks all have the same packet size and therefore the delays incurred by each of the users when transmitting between the same set of access points through the network are all equal, we define a matrix D e Rm*mWherein D isj,kIndicating that the task is at the access point APjWith access point APkThe transmission delay caused by the transmission between the two. Firstly, the user i transmits a task to the access point AP wirelessly connected with the user ijNeed for wireless delay omegaijAt this time, if the task requirement of user i needs to be scheduled to be deployed at APkTo be executed, the task needs to be executed from the APjTo APkThen, a delay D for transmitting tasks between the aps is additionally generatedjk。
And defining a function FuncQ, wherein the function obtains the queue time of the micro cloud under the load according to the sum lambda of the task arrival rates input to the micro cloud from all users, namely the workload of the current micro cloud and the working capacity of the micro cloud.
FuncQ(λ)＝C(c,λ/μ)/(cμ-λ) (1)
Mu represents the service rate of the server in the micro cloud;
equation (2) is the classical Erlang queue equation; c represents the number of servers, and rho represents the processing time of each server under the corresponding task;
as can be seen from equation (1), if the workload of a cloudlet is too heavy, the queue time becomes exceptionally long. The micro cloud load is increased, so that the waiting time of each workflow is prolonged, the average response time of the whole WMAN system is seriously increased, the application processing capacity of a user is weakened, and the poor use experience of the user is caused by that the user expresses a load set of the micro cloud scheduled to be placed at AP (k) by Λ.
Wherein UserjThe representation is scheduled for deployment at the APkA set of users of the micro cloud.
Deployed at APkQueue latency of each task on the micro cloud of
tcloudlet(k)＝FuncQ(Λ(k))+1/μ (4)
According to the formula (4), the average waiting time of the user i after unloading the task in the wireless metropolitan area network is the sum of the transmission delay and the queue waiting delay.
ti＝wij+Djk+tcloudlet(k) (5)
Wherein user i and access point APjWirelessly connected and scheduled to be deployed at the AP through the access point of the wired connectionkA micro cloud. Thus all subscribers in the wireless metropolitan area network systemThe average waiting time for offloading tasks, i.e. the response time of the system, is:
the location problem of deploying a micro cloud node in a wireless metropolitan area network is defined as (KCP): according to the given integer K is larger than or equal to 1 and system parameters (G, Lambda, W, D, mu, c), a micro-cloud node placement is given, so that the task average response time of the wireless metropolitan area network system is minimized:
2. the method of the invention
2.1 particle coding
The problem coding mode directly affects the algorithm search efficiency and performance, so a good problem code is needed to solve the problem of micro cloud deployment in the wireless metropolitan area network. Particle encoding rules: in the wireless metropolitan area network, if K micro clouds need to be deployed, the encoding length of the particles is K, if the micro clouds are deployed at the access point, the access point serial number is recorded in the particle set, the particles are directly encoded into the access point serial number, the order is increased according to the serial number, and the value does not exceed the maximum value of the access point serial number.
As shown in fig. 2 by the particle code, in the wireless metropolitan area network, K equals to 5 clouds, and when the access points with serial numbers 2,3,5,8, and 10 are selected as cloud deployment points, the particle code is as shown in fig. 2.
2.2 fitness function
The fitness of a particle is a main index for evaluating the superiority or inferiority of the particle, and it is generally specified that a particle having a small fitness function value corresponds to a better solution. In the present invention, the average response time of a particle is calculated by applying equation (7) and is referred to as the fitness function value of the current particle, and the shorter the average response time of the particle is, the more excellent the particle is.
2.3 particle iteration
Traditional PSO particle update strategy:
where t represents the current number of iterations, Vi tAnd Xi tRespectively representing the speed and the position of the ith particle at the t iteration; w is an inertial weight, representing the ability of the particle to maintain the current velocity; p _ besti tAnd g _ besttRespectively representing the self historical optimal value and the population historical optimal value r of the particle after t iterations1And r2Is two random factors, c1And c2The learning factor can control the learning ability of the particles to the self historical optimal value and the population historical optimal value. However, the conventional PSO particle update strategy is suitable for solving the continuous problem, and is prone to premature convergence and fall into local optimum in the iterative process, so that a more satisfactory result cannot be obtained. The improved PSO-GA algorithm particle introduces the cross mutation operation of a genetic operator, and the updating strategy formula is as follows:
Cgthe cross operation of the particles and the historical optimal value of the population is represented; cpA crossover operation representing a particle and the historical optimum of the particle itself;
the particles can go through three processes of variation of the particle individuals, individual cognition crossing and social population cognition crossing in the updating process, so that the particles are searched towards the optimal solution.
As shown in fig. 3, a particle has a w probability of generating an individual variation in the iterative process, where w is an inertial weight, and the variation probability of the particle is lower as the iteration progresses.
Fig. 4 is a cross-operation of the individual (social) cognitive segments. Wherein K1 and K2 respectively represent two quantiles of randomly selected coded particles in individual crossing operation and population crossing operation, and all the quantiles between the two quantiles are replaced by the values of the same quantile in p _ best and g _ best to form new particles.
Variation of individual particles:
individual cognitive crossover formula:
social population cognition cross formula:
2.4, parameter setting
Wherein the learning factors c1, c2 and the original inertia weight w of the individual and the population are as follows: as the iteration progresses progressively less, meaning that individuals prefer to retain their original particles, the optimal particle learning capacity for individual history and population decreases:
c＝(cstart-cend)*i/iteration (14)
w＝wmax-(wmax-wmin)*i/iteration (15)
however, the original inertial weight w is suitable for solving a linear problem, and the micro-cloud node placement problem of the invention is a nonlinear problem, so that the inertial weight formula is adjusted as follows:
the meaning of equation (17) is to solve for the ratio of the difference between the particle and the optimal solution in the particle length.
2.5 Algorithm flow
3. Results and analysis of the experiments
Firstly, in a wireless metropolitan area network system, calculating the average response time of a micro cloud node placement scheme with different micro cloud numbers generated by the system, and then drawing a line graph for comparison.
Relevant parameters of the PSO-GA algorithm are set as follows: maximum number of iterations 1000, initial population size 100, w _ max 0.9, w _ min 0.4, c1_ start 0.9, c1_ end 0.3, c2_ start 0.2, and c2_ end 0.9. The inertia factor w, the self-cognition factor c1 and the population cognition factor c2 of the algorithm are all set by adopting linear increasing and decreasing strategies of formulas (14) and (16).
In fig. 5, it shows how the average response time of the system changes with the change of the deployment number of the micro clouds under the different deployment situations of the micro clouds in 10 wireless metropolitan area networks with 20 access points m and 80 users n after many experiments. It can be seen that under all deployment schemes, the average response time of the system is gradually decreased along with the increase of the deployment number of the micro-clouds, but the decrease speed of the average response time of the system is decreased along with the increase of the number of the micro-clouds.
It can be seen from fig. 5 that the method of the present invention is always superior to other algorithms, and when K >8, there are enough micro clouds in the network so that all user tasks can be offloaded to the nearby micro clouds for execution, so that each user can save transmission time and waiting time, and the system response time will tend to be stable and no longer significantly reduced.
Because the method is an intelligent search algorithm, a better solution approaching to the optimal solution in a global scope can be always found under the condition of proper parameter setting. The essential meaning of the stochastic micro-cloud deployment algorithm is the average response time of the system after micro-cloud deployment under average conditions. The density-first micro-cloud deployment algorithm has performance close to that of a PSO-GA micro-cloud deployment algorithm, but the fact that the partial micro-cloud load is too heavy due to too intensive user tasks is sometimes considered, and the total average response time of the system is influenced. The load-first micro cloud deployment algorithm performs a little worse, the average response time of the system when K is 2 is longer than that of the other three deployment schemes, the micro clouds are deployed on two access points with the maximum workload, but the access point with the maximum workload is not necessarily the access point closest to other served users, meanwhile, some access points are wirelessly connected with a small number of users, the workload is close to zero, but most users in the network are not far away from the access point, the access point is a better micro cloud deployment point, so that the average response time of the system is longer than that of the other algorithms, and then when K is 4, the deployment number of the micro clouds is increased, the deployment scheme approaches the density-first deployment scheme, and the average response time of the system is greatly reduced.
Fig. 6 is a diagram illustrating a situation obtained by applying three micro-cloud deployment algorithms in a wireless metropolitan area network. Fig. 6(a) is an example of WMAN with 20 access points and 80 users, the solid black line represents the wired connection between the access points, the dotted line represents the wireless connection between the user and the access point, the red dots in the three graphs of fig. 6(b) (c) (d) represent the deployment location of the cloudlet, and the dotted line represents the access point selected by the user for direct wireless connection. By operating the algorithm, it can be seen that when K is 8, the average response times of the systems obtained by deploying the micro clouds by the three algorithms are not greatly different, and meanwhile, by observing the micro cloud deployment points obtained by the three algorithms in fig. 6, it can be seen that some micro cloud deployment position points in the three algorithms are overlapped, which also proves that the algorithm result provided by the user accords with the user density rule, and meanwhile, the optimization is obtained on the basis.
The invention provides a micro cloud node placement method based on a particle swarm optimization (PSO-GA) of a genetic algorithm operator, which considers the relation between users and access points in a complex wireless metropolitan area network environment, finds a reasonable and proper micro cloud node placement scheme to bring higher cost performance in the research and verification of reasonable micro cloud node number, and combines the advantages of the genetic algorithm and the particle swarm optimization algorithm to generate a micro cloud node placement result. According to the method, by optimizing the placement of the micro cloud nodes, under the condition of a reasonable number of micro clouds, the average response time of the user task is better reduced, high-efficiency and low-delay service is obtained, the user density rule is met, and meanwhile, the optimization is obtained on the basis.
The above are preferred embodiments of the present invention, and all changes made according to the technical scheme of the present invention that produce functional effects do not exceed the scope of the technical scheme of the present invention belong to the protection scope of the present invention.
Claims (2)
1. A micro cloud node placement method based on wireless metropolitan area network is characterized in that K access points are selected in the wireless metropolitan area network to place micro cloud nodes, the serial numbers of the access points are recorded in a particle set, directly coding the particles into access point serial numbers, randomly initializing a batch of particles based on a particle swarm optimization algorithm, calculating the fitness value of each particle through a fitness function, recording the fitness value as the optimal value of the particle, simultaneously selecting the individual with the optimal fitness value as the optimal population particle, recording the fitness value of the particle as the optimal population particle value, meanwhile, in each iteration, the particles are subjected to mutation operation and cross operation on excellent individuals to obtain new particles, the self optimal value of each particle is selected, and entering the next generation together with the optimal value of the population particles, and repeating iteration until the number of times of a preset threshold value is reached to complete the placement and search of the micro cloud nodes.
2. The method for placing the micro cloud node based on the wireless metropolitan area network according to claim 1, wherein the method is implemented as follows:
step S1, particle encoding:
selecting a K access point in a wireless metropolitan area network to place a micro-cloud node, namely the coding length of a particle is K, recording the access point serial number in a particle set, directly coding the particle into the access point serial number, and increasing progressively according to the serial number, wherein the value does not exceed the maximum value of the access point serial number;
step S2, fitness function:
calculating the average response time of the particles by using the following formula to be recorded as the fitness function value of the current particles, wherein the shorter the average response time of the particles is, the more excellent the particles are;
wherein:
ti＝wij+Djk+tcloudlet(k)
tcloudlet(k)＝FuncQ(Λ(k))+1/μ
FuncQ(λ)＝C(c,λ/μ)/(cμ-λ)
W＝{wiji 0 ≦ i < n,0 ≦ j < m } representing a set of wireless delays between the user and the wirelessly connected access point, D ≦ mjkJ is more than or equal to |0, and k is less than m, and represents a set of transmission delays among the access points; mu represents the service rate of the server in the micro cloud; funcQ(lambda) obtaining the queue time of the micro cloud under the load according to the sum lambda of the task arrival rates and the working capacity of the micro cloud; Λ (k) represents the sum of the arrival rates of tasks scheduled to cloudlet k; useriRepresenting a set of users; c (C, rho) calculation formula is a classic Erlang queue formula, C represents the number of servers, and rho represents the processing time of each server under the corresponding task;
step S3, particle iteration:
the improved PSO-GA algorithm particle is adopted to introduce the cross mutation operation of genetic operators, and the updating strategy formula is as follows:
wherein t represents the current iteration number; xi tRespectively representing the position of the ith particle at the time of the t iteration; w is an inertial weight, representing the ability of the particle to maintain the current velocity; p _ besti tAnd g _ bestt respectively represents the historical optimal value of the particle and the historical optimal value of the population after the particle is iterated for t times; c. C1And c2The learning factor can control the learning ability of the particles to the self historical optimal value and the population historical optimal value; cgThe cross operation of the particles and the historical optimal value of the population is represented; cpA crossover operation representing a particle and the historical optimum of the particle itself;
the particles can go through three processes of variation of the particle individuals, individual cognition crossing and social population cognition crossing in the updating process, so that the particles are searched towards the optimal solution; the particle individual variation formula, the individual cognition cross formula and the social population cognition cross formula are respectively as follows:
step S4, because the micro cloud node placement problem is a nonlinear problem, the inertia weight w formula is adjusted to
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911085727.3A CN110809275B (en) | 2019-11-08 | 2019-11-08 | Micro cloud node placement method based on wireless metropolitan area network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911085727.3A CN110809275B (en) | 2019-11-08 | 2019-11-08 | Micro cloud node placement method based on wireless metropolitan area network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110809275A true CN110809275A (en) | 2020-02-18 |
| CN110809275B CN110809275B (en) | 2021-09-28 |
Family
ID=69501588
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911085727.3A Expired - Fee Related CN110809275B (en) | 2019-11-08 | 2019-11-08 | Micro cloud node placement method based on wireless metropolitan area network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110809275B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111885493A (en) * | 2020-08-04 | 2020-11-03 | 西安邮电大学 | Micro-cloud deployment method based on improved cuckoo search algorithm |
| CN112118312A (en) * | 2020-09-17 | 2020-12-22 | 浙江大学 | Network burst load evacuation method facing edge server |
| CN112256415A (en) * | 2020-10-19 | 2021-01-22 | 福州大学 | Micro-cloud load balancing task scheduling method based on PSO-GA |
| CN112492032A (en) * | 2020-11-30 | 2021-03-12 | 杭州电子科技大学 | Workflow cooperative scheduling method under mobile edge environment |
| CN112785407A (en) * | 2021-02-10 | 2021-05-11 | 中国工商银行股份有限公司 | Multi-population genetic algorithm-based method and device for optimizing distribution proportion of nodes |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105812461A (en) * | 2016-03-09 | 2016-07-27 | 福州大学 | Mobile cloud environment context awareness computing migration method |
| US20170048876A1 (en) * | 2015-08-14 | 2017-02-16 | Nec Laboratories America, Inc. | Enabling high-bandwidth, responsive mobile applications in lte networks |
| CN109005211A (en) * | 2018-06-29 | 2018-12-14 | 福建师范大学 | Thin cloud deployment and scheduling user task method under a kind of wireless MAN environment |
| CN109451012A (en) * | 2018-11-01 | 2019-03-08 | 宜通世纪物联网研究院(广州)有限公司 | End cloud collaborative load balancing scheduling method, system and storage medium |
| US20190098474A1 (en) * | 2016-05-28 | 2019-03-28 | Huawei Technologies Co., Ltd. | Application migration method in a mobile edge system, related device, and system |
| CN110033076A (en) * | 2019-04-19 | 2019-07-19 | 福州大学 | Mix the Work stream data layout method below cloud environment to cost optimization |
| CN110138612A (en) * | 2019-05-15 | 2019-08-16 | 福州大学 | A kind of cloud software service resource allocation methods based on QoS model self-correcting |
| CN110418353A (en) * | 2019-07-25 | 2019-11-05 | 南京邮电大学 | A kind of edge calculations server laying method based on particle swarm algorithm |
-
2019
- 2019-11-08 CN CN201911085727.3A patent/CN110809275B/en not_active Expired - Fee Related
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170048876A1 (en) * | 2015-08-14 | 2017-02-16 | Nec Laboratories America, Inc. | Enabling high-bandwidth, responsive mobile applications in lte networks |
| CN105812461A (en) * | 2016-03-09 | 2016-07-27 | 福州大学 | Mobile cloud environment context awareness computing migration method |
| US20190098474A1 (en) * | 2016-05-28 | 2019-03-28 | Huawei Technologies Co., Ltd. | Application migration method in a mobile edge system, related device, and system |
| CN109005211A (en) * | 2018-06-29 | 2018-12-14 | 福建师范大学 | Thin cloud deployment and scheduling user task method under a kind of wireless MAN environment |
| CN109451012A (en) * | 2018-11-01 | 2019-03-08 | 宜通世纪物联网研究院(广州)有限公司 | End cloud collaborative load balancing scheduling method, system and storage medium |
| CN110033076A (en) * | 2019-04-19 | 2019-07-19 | 福州大学 | Mix the Work stream data layout method below cloud environment to cost optimization |
| CN110138612A (en) * | 2019-05-15 | 2019-08-16 | 福州大学 | A kind of cloud software service resource allocation methods based on QoS model self-correcting |
| CN110418353A (en) * | 2019-07-25 | 2019-11-05 | 南京邮电大学 | A kind of edge calculations server laying method based on particle swarm algorithm |
Non-Patent Citations (3)
| Title |
|---|
| 曾金晶: "《基于无线城域网的微云负载均衡算法》", 《计算机科学》 * |
| 林兵: "《混合云环境下面向时延优化的科学工作流数据布局策略》", 《计算机集成制造系统》 * |
| 邓茂菲: "基于移动边缘计算的任务迁移策略研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111885493A (en) * | 2020-08-04 | 2020-11-03 | 西安邮电大学 | Micro-cloud deployment method based on improved cuckoo search algorithm |
| CN111885493B (en) * | 2020-08-04 | 2022-03-29 | 西安邮电大学 | Micro-cloud deployment method based on improved cuckoo search algorithm |
| CN112118312A (en) * | 2020-09-17 | 2020-12-22 | 浙江大学 | Network burst load evacuation method facing edge server |
| CN112118312B (en) * | 2020-09-17 | 2021-08-17 | 浙江大学 | Network burst load evacuation method facing edge server |
| CN112256415A (en) * | 2020-10-19 | 2021-01-22 | 福州大学 | Micro-cloud load balancing task scheduling method based on PSO-GA |
| CN112256415B (en) * | 2020-10-19 | 2023-08-04 | 福州大学 | Micro cloud load balancing task scheduling method based on PSO-GA |
| CN112492032A (en) * | 2020-11-30 | 2021-03-12 | 杭州电子科技大学 | Workflow cooperative scheduling method under mobile edge environment |
| CN112492032B (en) * | 2020-11-30 | 2022-09-23 | 杭州电子科技大学 | Workflow cooperative scheduling method under mobile edge environment |
| CN112785407A (en) * | 2021-02-10 | 2021-05-11 | 中国工商银行股份有限公司 | Multi-population genetic algorithm-based method and device for optimizing distribution proportion of nodes |
| CN112785407B (en) * | 2021-02-10 | 2024-02-06 | 中国工商银行股份有限公司 | Lattice point distribution proportion optimizing method and device based on multiple swarm genetic algorithms |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110809275B (en) | 2021-09-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110809275B (en) | Micro cloud node placement method based on wireless metropolitan area network | |
| CN110109745B (en) | Task collaborative online scheduling method for edge computing environment | |
| CN111274035B (en) | Resource scheduling method and device in edge computing environment and computer equipment | |
| CN110717300B (en) | Edge calculation task allocation method for real-time online monitoring service of power internet of things | |
| CN112380008B (en) | Multi-user fine-grained task unloading scheduling method for mobile edge computing application | |
| CN109885397B (en) | Delay optimization load task migration algorithm in edge computing environment | |
| CN111372314A (en) | Task unloading method and task unloading device based on mobile edge computing scene | |
| US11968122B2 (en) | Joint optimization method and system for delay and spectrum occupation in cloud-edge collaborative network | |
| CN112822234A (en) | Task unloading method based on deep reinforcement learning in Internet of vehicles | |
| CN109639833B (en) | Task scheduling method based on wireless metropolitan area network micro-cloud load balancing | |
| CN111988787B (en) | Task network access and service placement position selection method and system | |
| CN112188627B (en) | Dynamic resource allocation strategy based on state prediction | |
| Fragkos et al. | Artificial intelligence enabled distributed edge computing for Internet of Things applications | |
| CN113220364A (en) | Task unloading method based on vehicle networking mobile edge computing system model | |
| CN111984419A (en) | Complex task computing and transferring method for marginal environment reliability constraint | |
| CN112040512B (en) | Mist computing task unloading method and system based on fairness | |
| CN113076177A (en) | Dynamic migration method of virtual machine in edge computing environment | |
| Selvamanju et al. | Machine learning based mobile data traffic prediction in 5g cellular networks | |
| CN113312183A (en) | Edge calculation method for deep neural network | |
| Wang et al. | Multi-objective joint optimization of communication-computation-caching resources in mobile edge computing | |
| CN116567651A (en) | Decision method and system for MEC task unloading and migration based on particle swarm | |
| CN109104455B (en) | Method for optimizing road edge micro-cloud load balance | |
| CN114615705B (en) | Single-user resource allocation strategy method based on 5G network | |
| CN111580943A (en) | Task scheduling method oriented to multi-hop unloading in low-delay edge calculation | |
| CN116367310A (en) | Maximum gain oriented channel allocation method for mobile edge calculation |
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: 20210928 |