CN114065894A

CN114065894A - Network arranging method, device, equipment and storage medium

Info

Publication number: CN114065894A
Application number: CN202010762863.8A
Authority: CN
Inventors: 童雅芳
Original assignee: China Mobile Communications Group Co Ltd; China Mobile Suzhou Software Technology Co Ltd
Current assignee: China Mobile Communications Group Co Ltd; China Mobile Suzhou Software Technology Co Ltd
Priority date: 2020-07-31
Filing date: 2020-07-31
Publication date: 2022-02-18

Abstract

The application provides a network arranging method, a device, electronic equipment and a computer storage medium, wherein the network arranging method comprises the following steps: obtaining an initial slice particle swarm of a network to be arranged; performing iterative processing on the initial slice particle swarm; under the condition that locally optimal slice particles exist in the initial slice particle swarm subjected to iterative processing, mapping a chaos variable obtained in advance into a solution space of the initial slice particle swarm to obtain a mapped slice particle swarm; and determining globally optimal slice particles in the initial slice particle swarm based on the mapped slice particle swarm. According to the network arranging method, the slice particles are updated by mapping the chaos variable obtained in advance to the solution space of the initial slice particle swarm, so that an optimal routing scheme is obtained, and meanwhile, the complexity of an algorithm is reduced.

Description

Network arranging method, device, equipment and storage medium

Technical Field

The present application relates to network orchestration technologies, and in particular, to a network orchestration method, apparatus, device, and storage medium.

Background

Currently, Optimization of networks can be converted into orchestration of network slices based on Genetic Algorithm-Particle Swarm Optimization (GA-PSO) algorithms. Generating a plurality of initial slice groups based on slice types by using the thought of heredity hybridization and variation, then continuously optimizing the slice groups by using the thought of heredity and variation to generate new network slices, and finally solving the optimal routing scheme through iteration. However, the method for updating slice particles through inheritance and variation is relatively complex, the algorithm complexity is high, and when a plurality of network nodes and large data traffic exist, it is very difficult to output an optimal routing scheme in time.

Disclosure of Invention

The application is expected to provide a network arranging method, a device, equipment and a storage medium. ,

in a first aspect, the present application provides a network orchestration method, the method comprising:

obtaining an initial slice particle swarm of a network to be arranged, wherein the initial slice particle swarm comprises N basic slice particles obtained after hybridization processing is carried out on basic network slices;

performing iterative processing on the initial slice particle swarm;

under the condition that locally optimal slice particles exist in the initial slice particle swarm subjected to iterative processing, mapping a chaos variable obtained in advance into a solution space of the initial slice particle swarm to obtain a mapped slice particle swarm;

and determining globally optimal slice particles in the initial slice particle swarm based on the mapped slice particle swarm.

In one embodiment, after iteratively processing the initial population of sliced particles, the method further comprises:

determining basic slice particles in the initial slice particle swarm after the iterative processing;

determining the average grain distance between the iterated basic slice grains;

determining the fitness variance corresponding to the iterated basic slice particles under the condition that the average particle distance between the basic slice particles is smaller than a preset average particle distance threshold;

and under the condition that the fitness variance corresponding to the basic slice particles is smaller than a preset fitness variance threshold, determining the basic slice particles as local optimal slice particles.

and under the condition that the local optimal slice particles do not exist in the initial slice particle swarm after the iterative processing, determining globally optimal slice particles in the initial slice particle swarm based on the initial slice particle swarm.

In one embodiment, the mapping the pre-obtained chaotic variables into the solution space of the initial slice particle swarm comprises:

generating a chaotic variable through a chaotic mapping equation based on a randomly generated initial chaotic variable;

and mapping the chaotic variable to a solution space of the initial slice particle swarm through carrier transformation.

In one embodiment, the mapping the chaotic variable into a solution space of the initial slice particle swarm by carrier transformation comprises:

obtaining initial slice particles to be updated; wherein the initial slice particles to be updated are a plurality of basic slice particles determined by taking the local optimal slice particles as centers in the initial slice particle swarm;

obtaining slice particles to be updated with chaotic motion characteristics through carrier wave transformation; the chaotic motion characteristic represents the characteristic that the position, the speed and the direction of the particle are changed;

determining the slice particles to be updated with the chaotic motion characteristics as target slice particles to be updated;

updating the initial slice particle to be updated based on the target slice particle to be updated.

In one embodiment, the iterative processing of the initial population of slice particles comprises:

obtaining initialization parameters for performing the iterative process, the initialization parameters including at least one of: initializing a learning factor, a maximum inertia weight, a minimum inertia weight, a maximum iteration number, an upper speed limit, a threshold value and an iteration number;

and performing iterative processing on the initialized particle swarm of slices based on the initialization parameters.

In one embodiment, the iterative processing of the initialized particle population of slices based on the initialization parameters comprises:

updating the position and velocity of the basic slice particles in the initialized slice particle population based on the initialization parameters.

In a second aspect, the present application further provides a network orchestration device, the device comprising: a first obtaining module, an iteration module, a mapping module, and a first determining module, wherein,

the first obtaining module is used for obtaining an initial slice particle swarm of a network to be arranged, wherein the initial slice particle swarm comprises N basic slice particles obtained after hybridization processing is carried out on basic network slices;

the iteration module is used for performing iteration processing on the initial slice particle swarm;

the mapping module is used for mapping a chaos variable obtained in advance to a solution space of the initial slice particle swarm to obtain a mapped slice particle swarm under the condition that locally optimal slice particles exist in the initial slice particle swarm after iterative processing;

the first determining module is configured to determine a globally optimal slice particle in the initial slice particle swarm based on the mapped slice particle swarm.

In one embodiment, the apparatus further comprises: a second determination module and a third determination module, wherein,

the second determining module is configured to determine basic slice particles in the initial slice particle group after the iterative processing; determining the average grain distance between the iterated basic slice grains;

the third determining module is configured to determine a fitness variance corresponding to the iterated basic slice particles when the average particle distance between the basic slice particles is smaller than a preset average particle distance threshold; and under the condition that the fitness variance corresponding to the basic slice particles is smaller than a preset fitness variance threshold, determining the basic slice particles as local optimal slice particles.

In one embodiment, the apparatus further comprises: a fourth determination module, wherein,

the fourth determining module is configured to determine, based on the initial slice particle group, a globally optimal slice particle in the initial slice particle group when the locally optimal slice particle does not exist in the initial slice particle group after the iterative processing.

In one embodiment, the mapping module is configured to generate the chaotic variable through a chaotic mapping equation based on a randomly generated initial chaotic variable; and mapping the chaotic variable to a solution space of the initial slice particle swarm through carrier transformation.

In one embodiment, the mapping module is configured to obtain an initial slice particle to be updated; wherein the initial slice particles to be updated are a plurality of basic slice particles determined by taking the local optimal slice particles as centers in the initial slice particle swarm; obtaining slice particles to be updated with chaotic motion characteristics through carrier wave transformation; the chaotic motion characteristic represents the characteristic that the position, the speed and the direction of the particle are changed; determining the slice particles to be updated with the chaotic motion characteristics as target slice particles to be updated; updating the initial slice particle to be updated based on the target slice particle to be updated.

In one embodiment, the iteration module is configured to obtain initialization parameters for performing the iterative process, where the initialization parameters include at least one of: initializing a learning factor, a maximum inertia weight, a minimum inertia weight, a maximum iteration number, an upper speed limit, a threshold value and an iteration number; and performing iterative processing on the initialized particle swarm of slices based on the initialization parameters.

In one embodiment, the iterative module is configured to update the position and velocity of the basic sliced particle in the initialized sliced particle population based on the initialization parameters.

In a third aspect, the present application further provides an electronic device comprising a processor and a memory for storing a computer program operable on the processor; wherein the content of the first and second substances,

the processor is configured to execute any one of the above network orchestration methods when running the computer program.

In a fourth aspect, the present application also provides a computer storage medium having a computer program stored thereon, which when executed by a processor, implements any of the above-described network orchestration methods.

In the application, the network arranging method realizes updating of the slice particles by mapping the chaos variable obtained in advance to the solution space of the initial slice particle swarm, and the slice particles are not updated in a genetic and variation mode, so that an optimal routing scheme is obtained, and meanwhile, the complexity of an algorithm is reduced. Further, whether the basic slice particles are the optimal slice particles or not is determined through the average particle distance between the basic slice particles and the corresponding fitness variance of the basic slice particles, and whether the basic slice particles are trapped in the local optimal solution or not can be determined more accurately.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the application.

Fig. 1 is a schematic flow chart of a network scheduling method according to the present application;

FIG. 2 is a flow chart illustrating another network orchestration method according to the present application;

FIG. 3 is a flow chart illustrating another network scheduling method according to the present application;

FIG. 4 is a flow chart illustrating another network scheduling method according to the present application;

FIG. 5 is a schematic flow chart of yet another network scheduling method according to the present application;

fig. 6 is a schematic diagram illustrating a process of implementing Network slicing based on a Software Defined Network (SDN) architecture according to the present application;

FIG. 7 is a schematic diagram of a network organization apparatus according to the present application;

fig. 8 is a schematic structural diagram of an electronic device according to the present application.

Detailed Description

The present application will be described in further detail below with reference to the accompanying drawings and examples. It should be understood that the examples provided herein are merely illustrative of the present application and are not intended to limit the present application. In addition, the following examples are provided as partial examples for implementing the present application, not all examples for implementing the present application, and the technical solutions described in the examples of the present application may be implemented in any combination without conflict.

It should be noted that in the embodiments of the present application, the terms "comprises", "comprising" or any other variation thereof are intended to cover a non-exclusive inclusion, so that a method or apparatus including a series of elements includes not only the explicitly recited elements but also other elements not explicitly listed or inherent to the method or apparatus. Without further limitation, the use of the phrase "including a. -. said." does not exclude the presence of other elements (e.g., steps in a method or elements in a device, such as portions of circuitry, processors, programs, software, etc.) in the method or device in which the element is included.

The term "and/or" herein is merely an association describing an associated object, meaning that three relationships may exist, e.g., a and/or B, may mean: a exists alone, A and B exist simultaneously, and B exists alone. In addition, the term "at least one" herein means any one of a plurality or any combination of at least two of a plurality, for example, including at least one of A, B, C, and may mean including any one or more elements selected from the group consisting of A, B and C.

In the related art, commonly used network arrangement algorithms include a network arrangement algorithm based on a greedy strategy and a network arrangement algorithm based on a GA-PSO. The network arrangement algorithm based on the greedy strategy refers to that in a pre-planning stage, namely before actual flow arrives, path planning is performed in advance through a control platform of network arrangement according to existing historical flow data and whole network topology information, construction of network slices is completed, node reservation information contained in a construction result is converted into queue configuration information, and the queue configuration information is issued to underlying network equipment. This divides the entire network into a number of different virtual networks, each of which is a network slice. Each network slice contains routing information and resource load conditions on the respective nodes. And then dividing the network demand by a greedy algorithm according to the actual flow demand, and arranging the divided networks to distribute reasonable routing schemes for different users.

The GA-PSO-based particle swarm optimization network orchestration algorithm is an algorithm that uses the shortest path first, and specifically, the network orchestration algorithm may generate 3 kinds of basic network slices according to a specific application scenario of a 5G network, that is: low-delay type slices, high-bandwidth type slices, and high-reliability type slices.

Furthermore, the 3 basic network slices can be crossed pairwise according to the concept of hybridization and variation in the genetic algorithm, and N particles with the highest fitness are selected according to the fitness function to form an initial network slice group. Each network slice is an M-dimensional matrix representing a feasible solution, i.e., a routing scheme, and M is the number of nodes included in the routing scheme. The best routing scheme is selected for the user, namely, the slice with the highest fitness in the slice group is selected, namely, the best solution in the slice group is searched. And searching an optimal solution by using a particle swarm optimization algorithm, wherein the initial network slice group is an initial slice particle swarm in the particle swarm optimization algorithm, and each network slice in the initial network slice group is each particle in the initial slice particle swarm. And after each iteration of the population is finished, respectively hybridizing the subgraph represented by the current particle with the local optimal subgraph and the global optimal subgraph, and then optimizing the hybridized subgraphs. And finally, outputting an optimal solution, namely an optimal routing scheme according to the iteration termination condition.

The network arranging algorithm based on the greedy strategy divides network slices for actual flow requirements by using a greedy algorithm through pre-divided network slices. The network arrangement algorithm is suitable for optimizing network resources of a data center with a simpler network state, and is not good for complex requirements of high bandwidth, low time delay and high reliability in a 5G practical application scene; moreover, the network arrangement algorithm can only optimize for a single target such as network resource utilization rate or Quality of Service (QoS), and global consideration is lacked.

The GA-PSO-based network arrangement algorithm is used for converting optimization of a network into arrangement of network slices. Specifically, the network arranging algorithm uses the thought of hybridization and variation in heredity, generates N initial slice particle groups based on basic network slices, then uses the thought of heredity and variation to continuously optimize the slice groups, generates new network slice particles, and finally obtains the optimal routing scheme through iteration. However, the mode of updating slice particles through inheritance and variation is complex, the algorithm complexity is high, and when a plurality of network nodes and data traffic are large, it may be difficult to output the optimal routing scheme in time.

In order to solve the above technical problem, an embodiment of the present application provides a network orchestration method, and fig. 1 is a schematic flow chart of the network orchestration method according to the embodiment of the present application, as shown in fig. 1, the flow chart may include:

step 101: obtaining an initial slice particle swarm of a network to be arranged, wherein the initial slice particle swarm comprises N basic slice particles obtained after hybridization processing is carried out on basic network slices; n is an integer of 2 or more.

In one embodiment, the basic network slice may be multiple types of basic network slices generated according to a specific application scenario of the 5G network, for example, 3 types of basic network slices may be generated, specifically, a low-latency type network slice, a high-bandwidth type network slice, and a high-reliability type network slice may be generated; the initial population of slice particles may be a collection comprising a plurality of elementary particle slices. The hybridization treatment of the basic network slices can be pairwise hybridization treatment according to a neutron map hybridization and subgraph variation method in a genetic algorithm, and of course, other hybridization treatments besides pairwise hybridization treatment can also be performed.

The basic slice particle may be an M-dimensional matrix, M representing the number of network nodes included in the network topology of the network to be programmed, and the characteristics of each basic slice particle may be represented by velocity, position, direction and fitness values. Meanwhile, the quality of each basic slice particle can be evaluated through a fitness evaluation function.

For the implementation manner of obtaining the initial slice particle group of the network to be arranged, for example, two-by-two hybridization processing may be performed on the low-delay type network slice, the high-bandwidth type network slice, and the high-reliability type network slice to obtain N basic slice particles after hybridization.

Step 102: and performing iterative processing on the initial slice particle swarm.

Here, the iterative processing of the initial slice particle group, which is aimed at a globally optimal slice particle in the initial slice particle group, may be implemented by a GA-PSO algorithm, and the basic slice particle may update a position, a velocity, and a direction of the basic slice particle by the locally optimal slice particle and the globally optimal slice particle in each iteration. It is understood that, in the process of solving the optimal solution, the initial particle slice group uses the position to represent the current solution of the basic slice particle, and uses the speed to represent the direction and speed of the slice particle approaching the optimal solution. Updating the position and velocity of the slice particles is the process by which all population particles are approaching the optimal solution.

In one example, the iteration may be performed according to the following iterative formula (1) and iterative formula (2),

wherein, c₁And c₂Is a learning factor, generally takes the value of 2; the rand () represents a random number between 0 and 1, w is an inertia weight and is used for adjusting the global search capability and the local search capability, generally, the larger w is, the stronger the global search capability is, and the smaller w is, the stronger the local search capability is; p is a radical of_iRepresenting a locally optimal solution for the current particle; g_iA global optimal solution is represented by a global optimal solution,

and

respectively representing the velocity of the particle at time k and the next time;

and

respectively representing the position of the particle at time k and the next time.

Step 103: and under the condition that locally optimal slice particles exist in the initial slice particle swarm subjected to the iterative processing, mapping a chaos variable obtained in advance into a solution space of the initial slice particle swarm to obtain the mapped slice particle swarm.

Here, the locally optimal slice particle may be expressed as an optimal solution found by the slice particle itself in an iterative process.

In one embodiment, before the iterative processing, it may be determined whether a current slice particle in the initial slice particle group after the iterative processing is a locally optimal slice particle; and under the condition that the current slice particle is determined to be the locally optimal slice particle, acquiring a chaotic variable, mapping the acquired chaotic variable to a solution space of the slice particle swarm, updating the initial slice particle swarm, and obtaining the mapped slice particle swarm.

Step 104: and determining globally optimal slice particles in the initial slice particle swarm based on the mapped slice particle swarm.

Here, the globally optimal slice particle may be represented as an optimal slice particle that can be currently found for the entire initialized slice particle population.

In an embodiment, the globally optimal slice particle in the initial slice particle group is determined based on the mapped slice particle group, and may be that the updated initial slice particle group is subjected to iterative processing to determine the globally optimal slice particle in the initial slice particle group.

In practical applications, the steps 101 to 104 may be implemented by a Processor in a service cluster, where the Processor may be at least one of an Application Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a Programmable Logic Device (PLD), an FPGA, a Central Processing Unit (CPU), a controller, a microcontroller, and a microprocessor.

It can be seen that, the embodiment of the present disclosure provides a network arranging method, obtaining an initial slice particle group of a network to be arranged, where the initial slice particle group includes N basic slice particles obtained after performing hybridization processing on basic network slices; n is an integer greater than or equal to 2; performing iterative processing on the initial slice particle swarm; under the condition that locally optimal slice particles exist in the initial slice particle swarm subjected to iterative processing, mapping a chaos variable obtained in advance into a solution space of the initial slice particle swarm to obtain a mapped slice particle swarm; and determining globally optimal slice particles in the initial slice particle swarm based on the mapped slice particle swarm. In the embodiment of the application, the network arranging method realizes the updating of the slice particles by mapping the pre-obtained chaotic variables into the solution space of the initial slice particle swarm, and the slice particles are not updated in a genetic and variation mode, so that the optimal routing scheme is favorably obtained, and meanwhile, the complexity of the algorithm is also reduced.

An embodiment of the present application provides a network arrangement method, and fig. 2 is a schematic flowchart of another network arrangement method according to the embodiment of the present application, and as shown in fig. 2, the process may include:

step 201: obtaining an initial slice particle swarm of a network to be arranged, wherein the initial slice particle swarm comprises N basic slice particles obtained after hybridization processing is carried out on basic network slices; n is an integer of 2 or more.

Step 202: and performing iterative processing on the initial slice particle swarm.

Step 203: determining basic slice particles in the initial slice particle swarm after the iterative processing.

Step 204: and determining the average grain distance between the basic slice grains after iteration.

Here, the average particle distance between particles may reflect the dispersion of distribution between individuals from a spatial perspective, and the calculation formula of the average particle distance of the basic slice particles is as follows:

wherein L is the maximum diagonal length of a search space, N is the size of the particle swarm of an initial slice, d is a solution space dimension, and p_idIs the d-dimension coordinate value, p, of the ith elementary slice particle_dRepresents the mean of the d-dimension coordinate values of all the elementary slice particles.

Step 205: and under the condition that the average particle distance between the basic slice particles is smaller than a preset average particle distance threshold value, determining the fitness variance corresponding to the basic slice particles after iteration.

Here, the preset average grain distance threshold may be an average grain distance threshold preset according to a requirement; an implementation manner in which the average grain distance between the elementary slice particles is smaller than the preset average grain distance threshold may be that the average grain distance between the elementary particles is smaller than the preset average grain distance threshold.

In one embodiment, the fitness variance reflects the particle distribution in terms of the function value, and the fitness variance is calculated as follows:

wherein N is the individual number of the initial slice particle swarm, f is a normalization scaling factor, f_iIs the fitness of the ith basic slice particle, f_avgIs the average fitness of the current population of initial slice particles.

Here, the fitness of the basic slice particle may be determined by a fitness evaluation function. The fitness evaluation function may be:

Fitness(α，β，D,Β)＝-αe^D+βe^B (5)；

d is the maximum time delay path in the single sub-graph after normalization; b is the minimum bandwidth in the sub-graph link after normalization; alpha is the proportion of the low-delay slices to all the slices; beta is the proportion of high bandwidth demand class slices to all slices.

Here, the formula for performing parameter normalization may be:

wherein, V_norDenotes the parameter after normalization, v is the performance parameter, u is the mean of the performance parameter, and σ is the variance of the performance parameter.

In one embodiment, the fitness of the current basic slice particle may be determined by a fitness evaluation function according to a threshold of an actual demand, where the threshold of the actual demand may be a maximum delay, a minimum bandwidth, a proportion of low-delay class slices in the actual demand, and a proportion of high bandwidth. For example, the fitness of the target basic slice particle can be calculated by substituting the actual required threshold as an input parameter into equation (5).

Wherein, the value of f is determined by adopting the following formula:

step 206: and under the condition that the fitness variance corresponding to the basic slice particles is smaller than a preset fitness variance threshold, determining the basic slice particles as local optimal slice particles.

Here, the preset fitness variance threshold may be a fitness variance threshold that is preset according to the demand.

In one embodiment, the adaptability variance is only used to describe the diversity of the initial slice particle group, which is not perfect, because there may be a case that the adaptability variance corresponding to the basic slice particles in the initial slice particle group is small, but the average particle distance of the particles is large. And only when the fitness variance and the average grain distance corresponding to the basic slice particles are small, the basic slice is considered to be easy to fall into the local optimal solution.

Step 207: and under the condition that locally optimal slice particles exist in the initial slice particle swarm subjected to the iterative processing, mapping a chaos variable obtained in advance into a solution space of the initial slice particle swarm to obtain the mapped slice particle swarm.

Step 208: and determining globally optimal slice particles in the initial slice particle swarm based on the mapped slice particle swarm.

According to the network arranging algorithm, whether the basic slice particles are the optimal slice particles can be determined through the average particle distance among the basic slice particles and the corresponding fitness variance of the basic slice particles, and therefore whether the basic slice particles are trapped in the local optimal solution can be determined accurately.

EXAMPLE III

The present application provides a network orchestration method, and fig. 3 is a schematic flowchart of another network orchestration method according to the present application, and as shown in fig. 3, the process may include:

step 301: obtaining an initial slice particle swarm of a network to be arranged, wherein the initial slice particle swarm comprises N basic slice particles obtained after hybridization processing is carried out on basic network slices; n is an integer of 2 or more.

Step 302: and performing iterative processing on the initial slice particle swarm.

Step 303: and under the condition that locally optimal slice particles exist in the initial slice particle swarm after the iterative processing, generating a chaotic variable through a chaotic mapping equation based on a randomly generated initial chaotic variable.

Here, the initial chaotic variable may be a d-dimensional vector, which may be expressed as Z₀，Z₀Each dimension of (a) may be a value in a value range between-1 and 1, and the number of initial chaotic variables is the same as the number of basic slice particles.

In one example, the chaotic mapping equation may be:

z_n+1＝sin(5.56/z_n),-1＜z_n＜1 (8)；

wherein Z is_nRepresenting the chaotic variable, Z, of the current moment_n+1And (3) expressing chaotic variables at the next moment, wherein n is an integer which is more than or equal to 0 and less than or equal to S R, and S R d-dimensional chaotic variables can be generated through a formula (8), wherein S is the scale of the basic slice particle swarm, and R is the chaotic radius and takes the value of 0 to 1.

Step 304: obtaining initial slice particles to be updated; wherein the initial slice particles to be updated are a plurality of basic slice particles determined by taking the local optimal slice particles as centers in the initial slice particle swarm.

In one example, the initial slice particles to be updated may be all basic slice particles within a range of the initial slice particle group with the locally optimal slice particle as a center and with R as a chaotic radius. Since the value of R is between 0 and 1, the initial slice particles to be updated may be part of the basic slice particles in the basic slice particle group.

Step 305: obtaining slice particles to be updated with chaotic motion characteristics through carrier wave transformation; the chaotic motion characteristic represents a characteristic in which the position, speed, and direction of the particle are changed.

Here, the expression formula of the carrier transform may be:

x_ij＝a_j+(b_j-a_j)z_ij (9)；

where i denotes the ith elementary slice particle in the population, j denotes the dimension in the solution space, a_jAnd b_jRepresenting the value range, x, of the current basic slice particle in the solution space_ijDenotes the ith elementary slice particle, Z, after carrier conversion_ijAnd expressing the chaotic variable corresponding to the ith basic slice particle.

Step 306: and determining the slice particle to be updated with the chaotic motion characteristic as a target slice particle to be updated.

Step 307: updating the initial slice particle to be updated based on the target slice particle to be updated.

In an example, the initial slice particle to be updated is updated based on the target slice particle to be updated, which may be that the target slice particle to be updated replaces the initial slice particle to be updated, so as to obtain an updated initial slice particle group.

According to the network arrangement algorithm, when the basic slice particles fall into a local optimal solution, the chaotic motion characteristic of the initial slice particles to be updated can be changed through carrier conversion according to the obtained chaotic variable, so that the position, the speed and the direction of the initial slice particles to be updated are changed, the local optimal solution is jumped out, the global optimal solution is favorably obtained, meanwhile, the chaotic variable obtained in advance is mapped into the solution space of the initial slice particle swarm, the slice particles are updated, the slice particles are not updated in a genetic and variation mode, and the complexity of the algorithm is reduced.

An embodiment of the present application provides a network arrangement method, and fig. 4 is a schematic flowchart of another network arrangement method according to the embodiment of the present application, and as shown in fig. 4, the process may include:

step 401: and obtaining an initial slice particle swarm of the network to be arranged, wherein the initial slice particle swarm comprises N basic slice particles obtained after the basic network slices are subjected to hybridization treatment.

Step 402: obtaining initialization parameters for performing the iterative process, the initialization parameters including at least one of: initializing a learning factor, a maximum inertia weight, a minimum inertia weight, a maximum iteration number, an upper speed limit, a threshold value and an iteration number.

Here, the maximum number of iterations may be used as a termination condition for the iterations.

Step 403: updating the position and velocity of the basic slice particles in the initialized slice particle population based on the initialization parameters.

Here, the position is used to indicate the current solution of the particle, and the velocity is used to indicate the direction and velocity in which the particle approaches the optimal solution. Updating the positions and velocities of the particles is the process by which all population particles are approaching the optimal solution.

Step 404: and under the condition that locally optimal slice particles exist in the initial slice particle swarm subjected to the iterative processing, mapping a chaos variable obtained in advance into a solution space of the initial slice particle swarm to obtain the mapped slice particle swarm.

Step 405: and determining globally optimal slice particles in the initial slice particle swarm based on the mapped slice particle swarm.

Step 406: and under the condition that the local optimal slice particles do not exist in the initial slice particle swarm after the iterative processing, determining globally optimal slice particles in the initial slice particle swarm based on the initial slice particle swarm.

The network arranging method provided by the application can improve the speed and the accuracy of obtaining the global optimal solution by setting reasonable initialization parameters.

The network slice division problem of the 5G network is essentially to calculate and find a relatively optimal routing scheme for a user, and belongs to the network routing distribution problem. According to the specific application scenario of the 5G network, 3 basic network slices are generated, namely: low-delay type slices, high-bandwidth type slices, and high-reliability type slices. Taking the 3 kinds of network slices as original basic slice particles, performing pairwise hybridization according to a neutron map hybridization and subgraph variation method in a genetic algorithm, obtaining N hybridized basic slice particles, and forming an initial slice particle swarm.

The M hybridized basic slice particles are initially divided network slices, each network slice is an M-dimensional matrix, and M represents the number of network nodes included in the network topology.

When an optimal routing scheme is selected for a user, a fitness evaluation function of the slice particles is determined according to a threshold value of actual demand, wherein the threshold value of the actual demand can be maximum time delay, minimum bandwidth, proportion of low-delay type slices in the actual demand and proportion of high bandwidth, and the threshold values are brought into the fitness evaluation function to calculate the current basic slice particles. The input parameters of the evaluation function are the threshold values, and the threshold values are substituted into the fitness evaluation function to calculate the fitness of the current particle. The fitness evaluation function is as follows:

Fitness(α，β，D,Β)＝-αe^D+βe^B (5)；

The parameter normalization formula is:

The method for updating the initial slice group by using the PSO algorithm can be that before the initial slice group is updated each time, an N-dimensional subgraph represented by a current slice is respectively crossed with a locally searched optimal slice subgraph and a globally searched optimal slice subgraph, then improper slices in the population are removed according to the variation thought in the genetic algorithm, a local optimal slice and a global optimal slice are determined according to a slice fitness function, and finally an optimal routing scheme is output according to iteration conditions.

In the process of iteratively solving the optimal solution of the initial slice group, the slice particles are easy to generate the phenomenon of early ripening, fall into the local optimal solution and lose the capability of searching the global optimal solution. Based on this, slice groups are abstracted to populations in the PSO algorithm, each slice being abstracted to one elementary particle in the population. In the GA-PSO algorithm, particles are led to jump out of a local optimal solution through heredity and variation; the method uses a GA-PSO algorithm based on chaos thought, updates the particle swarm by using chaos variables, and replaces the updating mode of hybridization and variation in the GA-PSO algorithm.

In one embodiment, before each iteration of the basic particle slice group, whether the current basic slice particle has the phenomenon of "precocity" may be determined, and the determination criteria are the average particle distance between the basic slice particles and the fitness variance of a single basic slice particle. The average particle distance between the basic slice particles reflects the distribution dispersion condition between individuals from a space angle, and the mean adaptive variance reflects the particle distribution condition from the function value aspect.

The calculation formula of the average grain distance of the basic slice particles is as follows:

The fitness variance of a single slice particle is calculated as follows:

Wherein, the value of f is determined by adopting the following formula:

in the related art, it is not perfect to describe the diversity of the population only by using the fitness variance, because there may be a case where the fitness variance corresponding to the basic slice particles is small, but the average grain distance of the basic slice particles is large. Only when the fitness variance and the average grain distance are both small, the algorithm is considered to be easy to fall into a local optimal solution, and premature convergence occurs.

In one embodiment, the basic slice particles may be considered to be trapped in the local optimal solution when the fitness variance is smaller than a preset fitness variance threshold C and when the average particle distance of the basic slice particles is smaller than a preset average particle distance threshold D; c, D are respectively threshold values preset according to actual requirements.

In one example, if the basic slice particles are premature, a set of chaotic variables is generated through a chaotic mapping characteristic equation, and the number of the chaotic variables is the same as that of the slices. Illustratively, a d-dimensional vector may be randomly generated as the initial value Z₀，Z₀Each dimension of (2) can be a value in a value range from-1 to 1, then N R d-dimensional chaotic variables are generated through a chaotic mapping equation, and N R d-dimensional chaotic variables are generated, wherein N is the size of the population scale, R is the chaotic search radius, and the value is from 0 to 1.

And (3) transforming the chaotic variable into a solution space where the slice particles are located to enable all the slice particles to obtain chaotic motion characteristics.

The expression formula of the carrier transformation may be:

x_ij＝a_j+(b_j-a_j)z_ij (9)；

And then directly using the mapped slice particle swarm to iterate, and updating the local optimal slice scheme and the global optimal slice scheme.

In an implementation manner, a further embodiment of the present application provides a network orchestration method, and fig. 5 is a schematic flowchart of the network orchestration method according to the embodiment of the present application, and as shown in fig. 5, the process may include:

step 501: and 3 kinds of base class network slices, namely a low-delay class slice, a high-bandwidth class slice and a high-reliability class slice are generated by using a shortest path algorithm.

Step 502: and carrying out pairwise hybridization on the basic slice particles to obtain an initial slice particle swarm.

Step 503: and evaluating the fitness of each basic slice particle in the initial slice particle swarm through a fitness evaluation function.

Step 504: and determining the fitness value of the slice particle with the highest fitness as an initial local optimal solution and a global optimal solution.

Here, the fitness value of the slice particle with the highest fitness is determined as a local optimal solution and a global optimal solution, and the current local optimal solution is respectively denoted as p_bestThe current global optimal solution is denoted as g_best。

Step 505: setting the initialization parameters of the PSO algorithm.

Here, setting initialization parameters may includeIncludes setting an initialization learning factor c₁And c₂Maximum inertial weight W_maxMinimum inertial weight W_minMaximum number of iterations T_maxUpper speed limit V_maxAnd chaotic search iteration times T.

Step 506: and judging whether the current iteration times are less than the maximum iteration times.

Here, if the current iteration count is smaller than the maximum iteration count, go to step 507, and if not, go to step 513.

Step 507: and performing iterative processing on the initial slice particle swarm.

Here, the iteration may be performed according to the following iterative formula (1) and iterative formula (2):

and

and

Step 508: the position and velocity of the initial population of sliced particles are updated.

Here, in the process of solving the optimal solution, the position of the initial sliced particle group is used to represent the current solution of the sliced particle, and the speed is used to represent the direction and speed of the particle approaching the optimal solution. Updating the position and velocity of the slice particles is the process by which all population particles are approaching the optimal solution.

Step 509: and calculating the fitness value of each slice particle, and updating the local optimal solution and the global optimal solution.

Step 510: the mean particle distance and fitness variance of each slice particle were calculated.

Step 511: judging whether the sliced particles have the phenomenon of early maturity. If yes, go to step 512, otherwise go to step 508.

Step 512: and generating a chaotic variable, mapping the chaotic variable to a solution space to replace the original basic slice particles, and turning to step 506.

In one example, the chaotic mapping equation may be:

z_n+1＝sin(5.56/z_n),-1＜z_n＜1 (8)；

wherein Z is_nRepresenting the chaotic variable, Z, of the current moment_n+1And (3) expressing chaotic variables at the next moment, wherein n is an integer which is more than or equal to 0 and less than or equal to S R, and S R d-dimensional chaotic variables can be generated through a formula (8), wherein S is the scale of the basic slice particle swarm, and R is the chaotic search radius and takes the value of 0-1.

And then mapping the chaotic variable into a solution space range through carrier transformation, and randomly replacing the original N × R basic slice particles. The expression formula of the carrier transformation may be:

x_ij＝a_i+(b_j-a_j)z_ij (9)

Step 513: and ending the iteration process, and returning the global optimal solution to be the optimal routing scheme.

In the embodiment of the application, because the network arranging method determines whether the basic slice particles are the optimal slice particles according to the average particle distance between the basic slice particles and the fitness variance corresponding to the basic slice particles, and does not determine the optimal slice particles according to only one of the average particle distance between the basic slice particles or the fitness corresponding to the basic slice particles, whether the basic slice particles are trapped in the locally optimal solution can be determined more accurately; the network arrangement algorithm realizes the updating of the slice particles by mapping the chaos variable obtained in advance to the solution space of the initial slice particle swarm, the slice particles are not updated in a genetic and variation mode, the optimal routing scheme is favorably obtained, and meanwhile, the complexity of the algorithm is also reduced.

A prerequisite for Network slice implementation is Network Function Virtualization (VNF) and SDN.

Fig. 6 is a schematic diagram of a process for implementing network slicing based on an SDN architecture according to an embodiment of the present application, as shown in fig. 6, the process includes the following main hardware devices: network slice Access Dedicated device (Dedicated Equipment), Access network Dedicated to device ran (radio Access network): radio access network, core network dedicated device nf (network function): a network function.

The access network functions are virtualized as Edge clouds (Edge clouds) and the Core network functions are virtualized as Core clouds (Core clouds) using NFV technology. The Edge Cloud and Core Cloud are managed using SDN architecture connectivity. Network slices are the product of virtualization of physical devices, and all slices need to be generated based on real physical devices.

When a terminal is accessed, the network arranging algorithm provided by the application can be used to obtain the optimal routing scheme for accessing Edge Cloud and Core Cloud, and finally, a slice is created for the terminal service according to the routing scheme. Here, each layer of network devices, such as access network devices, includes multiple physical devices, and on these physical devices, some software needs to be installed, and the function of these software is to virtualize the function of a physical machine and generate a network slice; different networks are connected through an internal network or an external network, for example, RAN Slice 1(RAN Slice1) and CN Slice 1(CN Slice1) in an access network are connected through the internal network, so that a whole connected link can be formed.

Fig. 7 is a schematic structural diagram of a network orchestration device according to an embodiment of the present application, and as shown in fig. 7, the network orchestration device may include: a first obtaining module 701, an iteration module 702, a mapping module 703 and a first determining module 704, wherein,

the first obtaining module 701 is configured to obtain an initial slice particle group of a network to be arranged, where the initial slice particle group includes N basic slice particles obtained by performing hybridization processing on basic network slices;

the iteration module 702 is configured to perform iteration processing on the initial slice particle swarm;

the mapping module 703 is configured to map a chaos variable obtained in advance to a solution space of the initial slice particle swarm to obtain a mapped slice particle swarm under the condition that locally optimal slice particles exist in the initial slice particle swarm after the iterative processing;

the first determining module 704 is configured to determine a globally optimal slice particle in the initial slice particle group based on the mapped slice particle group.

In one embodiment, the apparatus further comprises: a second determination module 705 and a third determination module 706, wherein,

the second determining module 705 is configured to determine basic slice particles in the initial slice particle group after the iterative processing; determining the average grain distance between the iterated basic slice grains;

the third determining module 706 is configured to determine a fitness variance corresponding to the iterated basic slice particles when the average particle distance between the basic slice particles is smaller than a preset average particle distance threshold; and under the condition that the fitness variance corresponding to the basic slice particles is smaller than a preset fitness variance threshold, determining the basic slice particles as local optimal slice particles.

In one embodiment, the apparatus further comprises: a fourth determining module 707, wherein,

the fourth determining module 707 is configured to, when there is no locally optimal slice particle in the initial slice particle group after the iterative processing, determine, based on the initial slice particle group, a globally optimal slice particle in the initial slice particle group.

In one embodiment, the mapping module 703 is configured to generate the chaotic variable through a chaotic mapping equation based on a randomly generated initial chaotic variable; and mapping the chaotic variable to a solution space of the initial slice particle swarm through carrier transformation.

In one embodiment, the mapping module 703 is configured to obtain an initial slice particle to be updated; wherein the initial slice particles to be updated are a plurality of basic slice particles determined by taking the local optimal slice particles as centers in the initial slice particle swarm; obtaining slice particles to be updated with chaotic motion characteristics through carrier wave transformation; the chaotic motion characteristic represents the characteristic that the position, the speed and the direction of the particle are changed; determining the slice particles to be updated with the chaotic motion characteristics as target slice particles to be updated; updating the initial slice particle to be updated based on the target slice particle to be updated.

In an embodiment, the iteration module 702 is configured to obtain an initialization parameter for performing the iterative process, where the initialization parameter includes at least one of: initializing a learning factor, a maximum inertia weight, a minimum inertia weight, a maximum iteration number, an upper speed limit, a threshold value and an iteration number; and performing iterative processing on the initialized particle swarm of slices based on the initialization parameters.

In one embodiment, the iterative model, 702, is configured to update the position and velocity of the basic slice particles in the initialized slice particle population based on the initialization parameters.

In practical applications, the first obtaining module 701, the iteration module 702, the mapping module 703, the first determining module 704, the second determining module 705, the third determining module 706, and the fourth determining module 707 may be implemented by a processor in an electronic device, where the processor may be at least one of an ASIC, a DSP, a DSPD, a PLD, an FPGA, a CPU, a controller, a microcontroller, and a microprocessor.

In addition, each functional module in this embodiment may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware or a form of a software functional module.

Based on the understanding that the technical solution of the present embodiment essentially or a part contributing to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium, and include several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) or a processor (processor) to execute all or part of the steps of the method of the present embodiment. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.

Specifically, the computer program instructions corresponding to a network arranging method in the present embodiment may be stored on a storage medium such as an optical disc, a hard disc, a usb disk, or the like, and when the computer program instructions corresponding to a network arranging method in the storage medium are read or executed by an electronic device, any one of the network arranging methods of the foregoing embodiments is implemented.

Based on the same technical concept of the foregoing embodiment, referring to fig. 8, it shows an electronic device provided in an embodiment of the present application, which may include: a memory 801 and a processor 802; wherein the content of the first and second substances,

the memory 801 is used for storing computer programs and data;

the processor 802 is configured to execute the computer program stored in the memory to implement any one of the network orchestration methods according to the foregoing embodiments.

In practical applications, the memory 801 may be a volatile memory (RAM); or a non-volatile memory (non-volatile memory) such as a ROM, a flash memory (flash memory), a Hard Disk (Hard Disk Drive, HDD) or a Solid-State Drive (SSD); or a combination of the above types of memories and provides instructions and data to the processor 802.

The processor 802 may be at least one of an ASIC, a DSP, a DSPD, a PLD, an FPGA, a CPU, a controller, a microcontroller, and a microprocessor. It is to be understood that, for different augmented reality cloud platforms, the electronic devices for implementing the above-described processor functions may be other, and the embodiments of the present application are not particularly limited.

In some embodiments, the functions of the apparatus provided in the embodiments of the present application or the modules included in the apparatus may be used to execute the method described in the above method embodiments, and for specific implementation, reference may be made to the description of the above method embodiments, and for brevity, details are not described here again

The foregoing description of the various embodiments is intended to highlight various differences between the embodiments, and the same or similar parts may be referred to each other, which are not repeated herein for brevity

The methods disclosed in the method embodiments provided by the present application can be combined arbitrarily without conflict to obtain new method embodiments.

Features disclosed in various product embodiments provided by the application can be combined arbitrarily to obtain new product embodiments without conflict.

The features disclosed in the various method or apparatus embodiments provided herein may be combined in any combination to arrive at new method or apparatus embodiments without conflict.

Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present application may be embodied in the form of a software product, which is stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal (such as a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present application.

While the present embodiments have been described with reference to the accompanying drawings, it is to be understood that the invention is not limited to the precise embodiments described above, which are meant to be illustrative and not restrictive, and that various changes may be made therein by those skilled in the art without departing from the spirit and scope of the invention as defined by the appended claims.

Claims

1. A method of network orchestration, the method comprising:

obtaining an initial slice particle swarm of a network to be arranged, wherein the initial slice particle swarm comprises N basic slice particles obtained after hybridization processing is carried out on basic network slices; n is an integer greater than or equal to 2;

performing iterative processing on the initial slice particle swarm;

2. The method of claim 1, wherein after iteratively processing the initial population of sliced particles, the method further comprises:

determining the average grain distance between the iterated basic slice grains;

3. The method of claim 1, wherein after iteratively processing the initial population of sliced particles, the method further comprises:

4. The method of claim 1, wherein said mapping a pre-obtained chaotic variable into a solution space of the initial slice particle population comprises:

5. The method of claim 4, wherein said mapping the chaotic variables into a solution space of the initial population of sliced particles by carrier transform comprises:

6. The method of claim 1, wherein said iteratively processing said initial population of sliced particles comprises:

7. The method of claim 6, wherein the iteratively processing the initialized population of slice particles based on the initialization parameters comprises:

8. An apparatus for network orchestration, the apparatus comprising: a first obtaining module, an iteration module, a mapping module, and a first determining module, wherein,

the first obtaining module is used for obtaining an initial slice particle swarm of a network to be arranged, wherein the initial slice particle swarm comprises N basic slice particles obtained after hybridization processing is carried out on basic network slices; n is an integer greater than or equal to 2;

9. An electronic device comprising a processor and a memory for storing a computer program operable on the processor; wherein the content of the first and second substances,

the processor is configured to execute the network orchestration method according to any one of claims 1 to 7 when running the computer program.

10. A computer storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the network orchestration method according to any one of claims 1 to 7.