CN115242295B

CN115242295B - Satellite network SDN multi-controller deployment method and system

Info

Publication number: CN115242295B
Application number: CN202210859825.3A
Authority: CN
Inventors: 钱克昌; 万颖; 熊达鹏; 王宇; 邢鹏; 刘文文; 张云帆; 苏英豪; 徐作兵; 朱沁雨; 高天昊
Original assignee: Peoples Liberation Army Strategic Support Force Aerospace Engineering University
Current assignee: Peoples Liberation Army Strategic Support Force Aerospace Engineering University
Priority date: 2022-07-21
Filing date: 2022-07-21
Publication date: 2023-04-21
Anticipated expiration: 2042-07-21
Also published as: CN115242295A

Abstract

The invention relates to the technical field of satellite networks, and particularly discloses a deployment method and a deployment system for a satellite network SDN multi-controller, wherein the deployment method comprises the steps of erecting a satellite network architecture based on SDN; constructing a multi-objective optimization model based on a satellite network; solving the multi-objective optimization model based on an improved NSGA-III algorithm, so as to obtain an optimal SDN multi-controller deployment scheme; deploying a satellite network based on the optimal SDN multi-controller deployment scheme; the method is used for constructing a multi-objective optimization model by taking the control link delay of the LEO satellite network, improving the reliability of inter-satellite links and maintaining the load balance of the LEO satellite network as research targets, solving the multi-objective optimization model by using an improved NSGA-III algorithm, determining the management and control relation between a controller and a switch, and realizing the effective deployment of controller nodes under the normal task of the LEO satellite network.

Description

Satellite network SDN multi-controller deployment method and system

Technical Field

The invention relates to the technical field of satellite networks, in particular to a method and a system for deploying a plurality of controllers of a satellite network SDN.

Background

The software defined network (Software Define Network, SDN) is composed of a data plane, a control plane and an application plane, and a typical SDN architecture is shown in fig. 1, which is different from a traditional network architecture, and has the biggest characteristics that the data plane is separated from the control plane, the control plane is centralized in logic, and the control of the network is realized through a unified and open southbound interface; the application plane of the SDN realizes the corresponding network function application, and interacts with the SDN control layer through the northbound interface; the northbound interface is an open interface between the SDN control plane and the application plane, and provides a universal open programming interface for SDN applications; the control plane consists of a plurality of controllers, is a core part of the whole framework, not only provides different layers of programmable capability for an upper application layer through a north interface, but also realizes unified control and management of a lower layer data plane through a south interface; the data plane is composed of hardware devices such as a simplified switch, the data plane device receives instructions from an upper control plane through a southbound interface, performs corresponding data forwarding and processing according to a flow table issued by the control plane, and provides current link state, statistical information and other data for the southbound interface to forward.

The appearance of the software-defined network opens up a new development space for the contemporary network, the idea of centralized control changes the idea of distributed control of the traditional network, and the idea of separation of a data plane and a control plane breaks the constraint of integration of the traditional network; SDN has software programmable flexibility, agility of centralized management control and convenience of decoupling a data plane and a control plane; the software-defined network overturns the architecture of the traditional network and provides a new idea for future network development.

The explosive growth of user data and the surge of network demands impact a satellite network at any time, and the satellite network has the advantages of large coverage range and long span distance, but is easy to cause conditions of high transmission delay, network congestion and the like of the satellite network due to the restriction factors of network topology time variation, weak link stability, unbalanced user distribution, difficult protocol updating and the like; the constraint of satellite network resources and network configuration limits the dynamic sharing of resources among heterogeneous networks, so that the utilization rate of the resources is reduced, and the network service supply is insufficient; in view of the above-mentioned ideas of SDN being programmable and reconfigurable, an SDN is introduced to drive the development of a satellite network, so as to separate a data layer and a control layer of the satellite network, and realize centralized control of the satellite network through an SDN controller; SDN is applied to a satellite network, so that flexible allocation of space network resources is effectively realized, the control capability of the satellite network is enhanced, and management convenience, network control flexibility and network configuration simplicity of a traditional satellite network can be promoted; because the number of LEO satellites is relatively large and the network topology is complex, the problem of controller node deployment of an LEO satellite control plane is mainly studied at present, namely, a proper controller deployment scheme is obtained based on an optimization algorithm by constructing a target optimization model.

The Non-dominant ranking genetic algorithm-III (Non-dominated Sorting Genetic Algorithm III, NSGA-III) in the optimization algorithm is an algorithm (the target number is more than or equal to 3) for solving the high-dimensional optimization problem, which is proposed by Deb et al, and as shown in FIG. 2, the concept of the NSGA-III algorithm mainly comprises: firstly, setting a reference point; subsequently, initializing a population, and performing cross mutation according to the initialized population to generate a new offspring population; then merging the child generation population and the initialized parent population, and carrying out rapid non-dominant sorting; finally, selecting individuals of each level according to the population number n, and selecting sparse-distributed individuals to enter a new population through a mechanism based on a reference point until the population number is reached if the individuals cannot be selected based on non-dominant levels; the multi-objective optimized solution is typically a set of optimal solutions, the set of solutions for the NSGA-iii algorithm is called Pareto optimal solution set; different from NSGA-II algorithm, the algorithm adopts a reference point-based mode to select individuals from Pareto solution (non-dominant solution) set, and has the advantages of being beneficial to solving the multi-objective problem and enabling the population to have good distribution; however, the searching capability of the population individuals is still defective, the possibility of sinking into local optimum is easy to exist, and the optimizing capability of an algorithm needs to be further enhanced.

Disclosure of Invention

Aiming at the problems, an object of the invention is to provide a satellite network SDN multi-controller deployment method, which mainly takes an LEO satellite network as a research object, analyzes the controller node deployment problem of an LEO satellite control plane, builds a multi-target optimization model for research targets by reducing the LEO satellite network control link delay, improving the reliability of inter-satellite links and maintaining the load balance of the LEO satellite network, solves the multi-target optimization model by using an improved NSGA-III algorithm, determines the management and control relation of a controller and a switch, and realizes the effective deployment of controller nodes under the normal task of the LEO satellite network.

A second object of the present invention is to provide a satellite network SDN multi-controller deployment system.

The first technical scheme adopted by the invention is as follows: a deployment method of a satellite network SDN multi-controller comprises the following steps:

s100: erecting a satellite network architecture based on SDN;

s200: constructing a multi-objective optimization model based on a satellite network;

s300: solving the multi-objective optimization model based on an improved NSGA-III algorithm, so as to obtain an optimal SDN multi-controller deployment scheme;

s400: deploying a satellite network based on the optimal SDN multi-controller deployment scheme;

Wherein, the step S300 includes the following substeps:

s310: coding the LEO satellite controller node and the LEO satellite exchanger node simultaneously, and initializing a population to generate an initialized population;

s320: calculating the fitness value of each individual in the initialized population, and performing rapid non-dominant sorting based on the fitness value of each individual to divide the initialized population into different levels; the fitness value of each individual is an objective function value of the reliability of an inter-satellite link of a controller deployment scheme, an objective function value of LEO satellite network load balancing and an objective function value of LEO satellite network control link time delay;

s330: selecting elite individuals in the rapidly non-dominated ordered population;

s340: selecting individuals participating in crossing from the elite individuals, and performing crossing operation on the individuals participating in crossing in a multi-point mixed crossing mode; randomly generating a random number, and executing mutation operation on the individuals after the cross operation when the random number is smaller than the self-adaptive mutation probability; when the random number is greater than or equal to the self-adaptive mutation probability, the mutation operation is not carried out on the individuals after the cross operation; constraint restoration is carried out on the offspring population after the crossover or mutation operation;

S350: performing elite selection based on an elite selection strategy to obtain an elite selected individual;

s360: the selected elite individuals and the current elite individuals are subjected to dominance judgment one by one, if the selected elite individuals can dominate the current elite individuals, the current elite individuals are further updated, otherwise, the current elite individuals are not updated;

s370: and repeating the steps S340-S360 to continuously optimizing and updating the population, outputting a non-dominant optimal solution set and decoding, thereby obtaining an optimal SDN multi-controller deployment scheme.

Preferably, the constructing the multi-objective optimization model in step S200 includes:

and constructing a reliability model of the inter-satellite link, an LEO satellite network load balancing model and an LEO satellite network control link time delay model.

Preferably, the reliability model of the inter-satellite link is expressed by the following formula:

wherein FAP is the average failure probability of the LEO satellite network; n is the number of LEO satellite exchanger nodes; i is LEO satellite controller node C _i Is the number of (2); j is LEO satellite switch node S _j Is the number of (2); c is a LEO satellite controller node set in the LEO satellite network; s is a set of all LEO satellite switch nodes in the LEO satellite network; AP (Access Point) _ij For LEO satellite controller node C _i And switch node S _j Probability of normal communication between; h _ij As matrix elements, the relation between LEO satellite controller nodes and LEO satellite switch nodes uses M multiplied by N matrix H= [ H ] _ij ] _M×N And (3) representing.

Preferably, the LEO satellite network load balancing model is represented by the following formula:

wherein B is _a Load balancing coefficients for the LEO satellite network; m is the number of LEO satellite controller nodes; i is LEO satellite controller node C _i Is the number of (2); c is a LEO satellite controller node set in the LEO satellite network; mu (mu) _i For LEO satellite controller node C _i Is a load of (2);

is the average load of the LEO satellite controller nodes.

Preferably, the LEO satellite network control link delay model is expressed by the following formula:

wherein T is LEO satellite network control link time delay; i is LEO satellite controller node C _i Is the number of (2); c is a LEO satellite controller node set in the LEO satellite network; tcl (Tcl) _i Processing time delay for LEO satellite controller node; tcb _i The propagation delay between the LEO satellite switch node and the LEO satellite controller node is set; j is LEO satellite switch node S _j Is the number of (2); s is a set of all LEO satellite switch nodes in the LEO satellite network; d, d _ij For LEO satellite controller node C _i And switch node S _j Shortest distance of inter-links; vc is the speed of light in free space; lambda (lambda) _j Request rate for LEO satellite switch data stream; f (f) _i For LEO satellite controller node C _i Maximum processing capacity of (2); h _ij As matrix elements, the relation between LEO satellite controller nodes and LEO satellite switch nodes uses M multiplied by N matrix H= [ H ] _ij ] _M×N And (3) representing.

Preferably, the adaptive mutation probability in step S340 is calculated by the following formula:

wherein Ap _m Is the adaptive mutation probability; gen is the current iteration number; maxgen is the maximum number of iterations.

Preferably, the mutation operation is performed in the step S340 by inserting mutation.

Preferably, the elite selection strategy in step S350 includes:

calculating the fitness value of each individual in the offspring population after the crossover or mutation operation, and carrying out population combination on the offspring population and the initialized population; deleting duplicate individuals from the pooled populations; and (3) carrying out rapid non-dominant sorting on the combined populations based on the fitness value, and selecting according to the limitations of non-dominant levels and population scale, thereby obtaining the elite selected individuals.

The second technical scheme adopted by the invention is as follows: a satellite network SDN multi-controller deployment system comprises a satellite network erection module, a multi-objective optimization model construction module, a calculation module and a deployment module;

The satellite network erection module is used for erecting a satellite network architecture based on SDN;

the multi-objective optimization model construction module is used for constructing a multi-objective optimization model based on a satellite network;

the computing module is used for solving the multi-objective optimization model based on an improved NSGA-III algorithm, so that an optimal SDN multi-controller deployment scheme is obtained;

the deployment module is used for deploying the satellite network based on the optimal SDN multi-controller deployment scheme;

wherein the computing module performs the following operations:

The beneficial effects of the technical scheme are that:

(1) The invention discloses a satellite network SDN multi-controller deployment method, which comprises the steps of erecting a satellite network architecture based on SDN; constructing a multi-objective optimization model (a reliability model of an inter-satellite link, an LEO satellite network load balancing model and an LEO satellite network control link time delay model) based on a satellite network; solving the multi-objective optimization model based on an improved NSGA-III algorithm to obtain a Pareto optimal solution set, and obtaining an optimal SDN multi-controller deployment scheme under a space network normal task according to the optimal solution set; deploying the satellite network based on an optimal SDN multi-controller deployment scheme; according to the method, an LEO satellite network is mainly used as a research object, the problem of deployment of controller nodes of an LEO satellite control plane is analyzed, the delay of an LEO satellite network control link is reduced, the reliability of inter-satellite links is improved, the load balance of the LEO satellite network is maintained as a research object, a multi-objective optimization model is built, an improved NSGA-III algorithm is used for solving the multi-objective optimization model, the management and control relation between the controller and a switch is determined, and the effective deployment of the controller nodes under a satellite network normal task is realized.

(2) The invention aims to solve the problem of deployment of reliability of a satellite network multi-controller under a space normal task, and provides a satellite network SDN multi-controller deployment method based on an improved NSGA-III algorithm; according to the invention, by analyzing the performance characteristics of the LEO satellite network multi-controller deployment problem, integrating the LEO satellite network control link time delay, the inter-satellite link reliability and the LEO satellite network load balance multiple targets, a multi-target optimization model for satellite network SDN controller deployment is established, and is a typical multi-target optimization problem, and as the multiple targets have a certain constraint relationship, the optimization performance of other targets can be influenced while optimizing one target, so that the multiple targets are difficult to achieve the optimization simultaneously; the NSGA-III algorithm is suitable for solving the problem of more than two targets, and the diversity of the population can be enriched through the distributed and balanced reference points, so the invention considers the use of the algorithm and performs optimization and improvement, and provides an improved multi-controller deployment method of the NSGA-III algorithm, which is suitable for the deployment problem of SDN controllers, determines the management and control relationship between the controllers and the switches, and realizes the effective deployment of the controller nodes under the normal task of the satellite network.

(3) The invention discloses a satellite network SDN multi-controller deployment method, which is to carry out discretization improvement on an NSGA-III algorithm on the basis of constructing a reliability model of an inter-satellite link, an LEO satellite network load balancing model and an LEO satellite network control link delay model based on a satellite network architecture based on SDN, so that the improved NSGA-III algorithm is prevented from being trapped into local optimum.

(4) According to the invention, the NSGA-III algorithm is improved in a discretization manner, elite individuals are introduced in the aspect of improvement of a crossover operator, and a three-point mixed crossover mode is adopted, so that the self-adaptive mutation probability and the insertion mutation method are added for further optimizing a mutation operator; the simulation result effectively verifies that the INSSGA-III algorithm (improved NSGA-III algorithm) has superiority in the multi-controller deployment method optimizing aspect compared with the traditional NSGA-III and INSSGA-II algorithms, whether in the aspect of single target optimizing performance or the aspect of overall performance optimizing.

Drawings

FIG. 1 is a schematic diagram of a typical SDN architecture;

FIG. 2 is a schematic flow chart of NSGA-III algorithm;

fig. 3 is a flow chart of a deployment method of multiple controllers of a satellite network SDN according to an embodiment of the present invention;

fig. 4 is a schematic diagram of an SDN based satellite network architecture according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of a coding sequence of a multi-controller deployment scenario provided by one embodiment of the present invention;

FIG. 6 is a schematic diagram of an allocation processing sequence of a multi-controller deployment scenario provided by one embodiment of the present invention;

FIG. 7 is a schematic diagram of a randomly initialized population Y according to one embodiment of the present invention;

FIG. 8 is a schematic diagram of an individual selected elite provided in accordance with one embodiment of the present invention;

FIG. 9 is a schematic diagram of a three-point hybrid crossover process provided by one embodiment of the present invention;

FIG. 10 is a graph showing the variation of the adaptive variation probability with the number of iterations according to an embodiment of the present invention;

FIG. 11 is a schematic diagram illustrating an insertion mutation operation procedure according to an embodiment of the present invention;

FIG. 12 is a Pareto front-end diagram of three multi-objective genetic algorithms with a controller number of 10 in the simulation experiment of the present invention;

FIG. 13 is a graph of delay versus load balancing analysis in a simulation experiment of the present invention;

FIG. 14 is a graph of load balancing and reliability coefficient analysis in a simulation experiment of the present invention;

FIG. 15 is a graph showing reliability coefficients and time delay analysis in the simulation test of the present invention

FIG. 16 is a Pareto front-end diagram of 12 controllers in a simulation experiment according to the present invention;

FIG. 17 is a Pareto front-end diagram of 8 controllers in a simulation experiment according to the present invention;

Fig. 18 is a schematic structural diagram of a satellite network SDN multi-controller deployment system according to an embodiment of the present invention.

Detailed Description

Embodiments of the present invention are described in further detail below with reference to the accompanying drawings and examples. The following detailed description of the embodiments and the accompanying drawings are provided to illustrate the principles of the invention and are not intended to limit the scope of the invention, i.e. the invention is not limited to the preferred embodiments described, which is defined by the claims.

In the description of the present invention, it is to be noted that, unless otherwise indicated, the meaning of "plurality" means two or more; the terms "first," "second," and the like are used for descriptive purposes only and are not to be construed as indicating or implying relative importance; the specific meaning of the above terms in the present invention can be understood as appropriate by those of ordinary skill in the art.

Non-dominant solution: any two solutions Q1 and Q2 in the problem, if all targets of Q1 are better than Q2, then Q1 is said to dominate Q2; if Q1 is not dominated by other solutions, then Q1 is referred to as a non-dominated solution (i.e., a Pareto solution).

Pareto front: all Pareto solutions in space constitute planes.

Example 1

Fig. 3 is a method for deploying multiple controllers in a satellite network according to an embodiment of the present invention, including the following steps:

s100: erecting a satellite network architecture based on SDN;

as shown in fig. 4, the control plane of the SDN-based satellite network architecture is used as a core part of the satellite network, and a hierarchical control architecture is formed by GEO satellites, LEO satellites and ground stations; the master controller is arranged on the ground station, dynamically manages the whole satellite network in a whole network view angle, and sets GEO satellites as regional controllers, wherein LEO satellites have low link loss and low communication time delay, so that part of LEO satellites are set as slave controllers, and the rest LEO satellites are set as switches; the slave controllers may collect status information of the switches within the respective control domains, and when the slave controllers are unable to process the intra-domain switch service requests, the slave controllers may send requests to the zone controllers, which may establish communications with the master controller, which then makes decisions to be distributed to the slave controllers via the zone controllers.

The invention refers to Iridium constellation data to construct LEO satellite network environment, and obtains satellite orbit related information through STK, the satellite orbit data of related Iridium constellation is as follows: the orbit height of LEO satellites is 780km, the number of LEO satellites is 66, the orbit number is 11, the number of links in orbit and the number of links between orbits of each LEO satellite are 2, the orbit inclination angle is 86.4 degrees, and the orbit period is 6061.1s; the invention mainly divides the operation period of the LEO satellite network into a group of time slices, the network topology and the data flow change are not considered in a single time slice, the deployment problem of the SDN controller is studied in the state, 66 LEO satellites are all set as LEO satellite switch nodes according to the Iridium constellation characteristics, and proper nodes are selected from the LEO satellite switch nodes in an in-band mode to deploy the LEO satellite controller nodes, so that individual nodes can take roles of the switch and the controller at the same time, the propagation delay generated between two points is considered to be 0, and in addition, the data link and the control link are mutually independent and are not interfered with each other.

S200: constructing a multi-target optimization model based on a satellite network, wherein the multi-target optimization model comprises a reliability model of an inter-satellite link, an LEO satellite network load balancing model and an LEO satellite network control link time delay model;

(1) Defining parameters;

the current LEO satellite network topology is represented by an undirected graph g= (S, E, C), where s= { S ₁ ,S ₂ ,...,S _N E= { E } is a set of all LEO satellite switch nodes in the LEO satellite network ₁ ,e ₂ ,...,e _l And is a set of links between LEO satellite switch nodes, c= { C ₁ ,C ₂ ,...,C _M A set of LEO satellite controller nodes in the LEO satellite network; because 66 LEO satellites are set as LEO satellite switch nodes by referring to an Iridium constellation, the number of the LEO satellite switch nodes is N=66, and the number of the LEO satellite controller nodes is M; lambda (lambda) _j Request rate, f, for LEO satellite switch data stream _i For LEO satellite controller node C _i Maximum processing capacity of (2); the relationship between LEO satellite controller node and LEO satellite switch node uses M x N dimensional matrix H= [ H ] _ij ] _M×N Representation, wherein the matrix element H _ij If the LEO satellite switch node is in the control domain of the LEO satellite controller node in the controller deployment scheme, the element H is the binary variable _ij And the value of (2) is 1, otherwise 0.

(2) Constraint relation;

in order to form an effective deployment scheme of the LEO satellite controller, the current situation of the LEO satellite network and the SDN principle are combined, and the constraint relation existing in the deployment problem of the controller is comprehensively analyzed, wherein the main constraint relation comprises: management constraints between the LEO satellite switch node and the LEO satellite controller node, processing power constraints of the LEO satellite controller node.

1) The management constraint between the LEO satellite switch nodes and the LEO satellite controller nodes mainly constrains that each LEO satellite switch node can only be associated with a unique LEO satellite controller node, and one LEO satellite controller node can be simultaneously associated with a plurality of LEO satellite switch nodes, and the constraint relation is expressed by the following formula:

s is a set of all LEO satellite exchanger nodes in the LEO satellite network; c is a LEO satellite controller node set in the LEO satellite network; n is the number of LEO satellite exchanger nodes; h _ij As matrix elements, the relation between LEO satellite controller nodes and LEO satellite switch nodes uses M multiplied by N matrix H= [ H ] _ij ] _M×N Representation, wherein the matrix element H _ij Is a binary variable; i is LEO satellite controller node C _i Is the number of (2); j is LEO satellite switch node S _j Is a number of (3).

2) The processing power constraints of the LEO satellite controller nodes essentially constrain that controller deployment must be performed within the capabilities of the LEO satellite controller nodes, the constraint relationship of which is expressed by the following equation:

wherein mu is _i For LEO satellite controller node C _i Is the LEO satellite controller node C _i The sum of the flow request rates of all the managed LEO satellite switches; i is LEO satellite controller node C _i Is the number of (2); j is LEO satellite switch node S _j Is the number of (2); h _ij As matrix elements, the relation between LEO satellite controller nodes and LEO satellite switch nodes uses M multiplied by N matrix H= [ H ] _ij ] _M×N Representation, wherein the matrix element H _ij Is a binary variable; s is a set of all LEO satellite switch nodes in the LEO satellite network; c is a LEO satellite controller node set in the LEO satellite network; lambda (lambda) _j Request rate for LEO satellite switch data stream; f (f) _i For LEO satellite controller node C _i Maximum processing capacity of (a) is provided.

(3) Constructing a multi-target optimization model, namely constructing a reliability model of an inter-satellite link, an LEO satellite network load balancing model and an LEO satellite network control link time delay model;

1) Constructing a reliability model of an inter-satellite link;

in a satellite network, the reliability of an inter-satellite link is the probability that a control link in the satellite network can be normally communicated; because of the existence of a plurality of connecting links among LEO satellites, if the connecting relation between the LEO satellite switch node and the LEO satellite controller node is determined by taking time delay or load balancing as an element, the influence of the inter-satellite link failure fault condition is ignored, and the problem that the data transmission rate of part of LEO satellite controller nodes is low is caused; meanwhile, under the influence of the polar region, the inter-satellite links between the tracks need to be closed, and communication can be carried out only by the inter-satellite links in the tracks, and at the moment, the failure rate of the inter-satellite links in the same track is lower than that of the inter-satellite links between the tracks; the inter-satellite link fault can disconnect the control plane from the data plane, so that the LEO satellite switch node can not transmit a message request to the LEO satellite controller node, and can not receive an instruction from the LEO satellite controller node, and the problems of data packet loss, serious network performance reduction and the like can exist; therefore, the invention mainly considers the influence of the reliability of the inter-satellite links of the satellite network from the two aspects of the physical failure condition of the control links and the congestion condition of the inter-satellite links.

(1) Physical failure conditions of the control link;

the probability of failure of the control link is mainly related to the length of the control link, LEO satellite controller node C _i To the switch node S _j Probability P of failure of inter-control link _ij Calculated by the following formula:

/>

wherein P is _ij For LEO satellite controller node C _i To the switch node S _j Probability of failure of the inter-control link; p (P) _b Failure rate for control link per unit length; d, d _ij For LEO satellite controller node C _i And switch node S _j Shortest distance of inter-links; i is LEO satellite controller node C _i Is the number of (2); j is LEO satellite switch node S _j Is a number of (3).

(2) Congestion conditions of inter-satellite links;

the probability of inter-satellite link congestion is related to the amount of traffic in the link, and within the threshold of link traffic, the network congestion is negligible and exceedsWhen the link threshold value is in a range, the larger the flow is, the larger the probability of congestion of the inter-satellite link is; LEO satellite controller node C _i And switch node S _j Congestion probability P of inter-satellite link _con Calculated by the following formula:

wherein P is _con For LEO satellite controller node C _i And switch node S _j Congestion probability of inter-satellite links; lambda (lambda) _j Request rate for LEO satellite switch data stream; r is a threshold value of congestion condition of a link; h _ij Is a matrix element; w (w) _ij For LEO satellite controller node C _i And switch node S _j Inter-link bandwidth; i is LEO satellite controller node C _i Is the number of (2); j is LEO satellite switch node S _j Is a number of (3).

Node C combined with LEO satellite controller _i To the switch node S _j Probability P of failure of inter-control link _ij And congestion probability P of inter-satellite link _con Obtaining LEO satellite controller node C _i And switch node S _j Probability of inter-normal communication AP _ij Is expressed by the following formula:

AP _ij ＝(1-P _con )P _ij

in AP _ij For LEO satellite controller node C _i And switch node S _j Probability of normal communication between; p (P) _con For LEO satellite controller node C _i And switch node S _j Congestion probability of inter-satellite links; p (P) _ij For LEO satellite controller node C _i To the switch node S _j Probability of failure of the inter-control link; i is LEO satellite controller node C _i Is the number of (2); j is LEO satellite switch node S _j Is a number of (3).

The invention evaluates the reliability of the inter-satellite links of the LEO satellite network through the average failure probability FAP of the LEO satellite network representing the reliability coefficient, wherein the lower the value of the FAP is, the higher the reliability of the inter-satellite links is, the higher the reliability of the whole satellite network is, and the average failure probability FAP of the LEO satellite network (namely the reliability model of the inter-satellite links) is represented by the following formula:

Wherein FAP is the average failure probability of the LEO satellite network; n is the number of LEO satellite exchanger nodes; i is LEO satellite controller node C _i Is the number of (2); j is LEO satellite switch node S _j Is the number of (2); c is a LEO satellite controller node set in the LEO satellite network; s is a set of all LEO satellite switch nodes in the LEO satellite network; AP (Access Point) _ij For LEO satellite controller node C _i And switch node S _j Probability of normal communication between; h _ij As matrix elements, the relation between LEO satellite controller nodes and LEO satellite switch nodes uses M multiplied by N matrix H= [ H ] _ij ] _M×N Representation, wherein the matrix element H _ij Is a binary variable.

2) Constructing an LEO satellite network load balancing model;

the load balancing coefficient of the LEO satellite network can evaluate the load difference degree of each controller node of the LEO satellite network; calculating the difference between the load of each LEO satellite controller node (namely the sum of the flow request rates of all LEO satellite switches controlled by the LEO satellite controller node) and the average load of the whole LEO satellite network controller node through a variance method, wherein the closer the result is to 0, the more balanced the loads of each LEO satellite controller node are indicated; average load of LEO satellite controller nodes

Is expressed by the following formula:

Load balancing coefficient B of LEO satellite network _a (i.e. LEO)Satellite network load balancing model) is represented by the following formula:

wherein B is _a Load balancing coefficients for the LEO satellite network; m is the number of LEO satellite controller nodes; i is LEO satellite controller node C _i Is the number of (2); c is a LEO satellite controller node set in the LEO satellite network; mu (mu) _i For LEO satellite controller node C _i Is the LEO satellite controller node C _i The sum of the flow request rates of all the managed LEO satellite switches;

is the average load of the LEO satellite controller nodes.

3) Constructing an LEO satellite network control link time delay model;

the LEO satellite network control link time delay mainly comprises two aspects of processing time delay of the LEO satellite controller node and propagation time delay between the LEO satellite switch node and the LEO satellite controller node, and the LEO satellite network control link time delay is expressed by the following formula:

wherein T is LEO satellite network control link time delay; i is LEO satellite controller node C _i Is the number of (2); c is a LEO satellite controller node set in the LEO satellite network; tcl (Tcl) _i Processing time delay for LEO satellite controller node; tcb _i The propagation delay between the LEO satellite switch node and the LEO satellite controller node is set; j is LEO satellite switch node S _j Is the number of (2); s is a set of all LEO satellite switch nodes in the LEO satellite network; d, d _ij For LEO satellite controller node C _i And switch node S _j Shortest distance of inter-links; vc is the speed of light in free space; lambda (lambda) _j Request rate for LEO satellite switch data stream；f _i For LEO satellite controller node C _i Maximum processing capacity of (2); h _ij As matrix elements, the relation between LEO satellite controller nodes and LEO satellite switch nodes uses M multiplied by N matrix H= [ H ] _ij ] _M×N Representation, wherein the matrix element H _ij Is a binary variable.

In order to ensure normal communication of space normal tasks, provide real-time, reliable and safe service, and complete space tasks with high efficiency and high standard, the invention takes the improvement of the reliability of inter-satellite links, the balance of the load of LEO satellite network and the reduction of the delay of LEO satellite network control links as optimization targets, constructs a multi-target optimization model, and the multi-target optimization model (objective function) is expressed by the following formula:

objective function = min [ FAP, T, B _a ]

Wherein FAP is the average failure probability of the LEO satellite network; t is LEO satellite network control link time delay; b (B) _a Is a load balancing coefficient of the LEO satellite network.

S300: and solving the multi-objective optimization model (the inter-satellite link reliability model, the LEO satellite network load balancing model and the LEO satellite network control link time delay model) based on an improved NSGA-III algorithm to output a Pareto optimal solution set, and decoding based on the optimal solution set to obtain a group of multi-objective optimal SDN multi-controller deployment scheme.

S310: coding LEO satellite controller nodes and LEO satellite exchanger nodes in the LEO satellite network simultaneously, and initializing a population to generate an initialized population;

(1) Encoding LEO satellite controller nodes and LEO satellite switch nodes in the LEO satellite network;

the SDN controller deployment method mainly solves the problem of how to distribute controllers to process Packet-in messages sent by different switches, different distribution strategies can form different controller deployment schemes, and one scheme can intuitively reflect the management and control relationship between the controllers and the switches; because the controller node can process the Packet-in message sent by the switch according to the sequence of the message sent by the switch, the switch node with a later message sending process can have a queuing waiting process, so the invention adopts a double-layer coding mode to form a deployment scheme, the LEO satellite controller node and the LEO satellite switch node are coded simultaneously, and then the switch is distributed to the controller according to the coding sequence.

The scheme of distributing N switches to M controllers can form a controller code sequence X _C ＝[C ₁ ,C ₃ ,C ₄ ...C _M ...C ₂ ]And a switch code sequence X _S ＝[S ₅ ,S ₂ ,S ₁ ...S _N ...S ₃ ]The coding mode is shown in fig. 5, and the coding lengths are all N; s is S ₅ Representing LEO satellite exchanger node with number 5, according to the number of iridium satellite base satellites being 66, the value of N is 66; c (C) ₁ Representing the LEO satellite controller node with controller number 1.

(2) Initializing a population according to the coding mode to generate an initialized population Y, namely a plurality of controller deployment schemes;

corresponding to the double coding sequences (the controller coding sequences and the switch coding sequences), performing LEO satellite controller node deployment allocation, wherein the LEO satellite controller node deployment allocation is to allocate LEO satellite switch nodes to LEO satellite controller nodes, the allocation result is shown in fig. 6, and LEO satellite controller node C in fig. 6 ₁ Processing LEO satellite switch nodes S one by one according to deployment scheme ₅ ，S ₄ Equal LEO satellite controller node C ₃ Processing LEO satellite switch nodes S one by one according to deployment scheme ₂ And so on, in this order, switch nodes are allocated one by one.

The population is constructed by adopting random initialization, one deployment scheme is determined by the switch coding sequence and the controller coding sequence at the same time because one individual in the population corresponds to one deployment scheme, the pop group controller and the switch coding sequence meeting the constraint conditions can be formed according to the number pop of the population, and then the random initialization population Y is formed, wherein the random initialization population Y is shown in figure 7.

S320: calculating the fitness value of each individual in the initialized population, and performing rapid non-dominant sorting based on the fitness value of each individual to divide the initialized population into different levels;

calculating the fitness value of each individual in the initialized population, wherein the fitness value of each individual consists of three objective function values, namely the three objective function values of the reliability of the inter-satellite link of the controller deployment scheme, the LEO satellite network load balance and the LEO satellite network control link time delay;

then, according to the fitness value of each individual, quick non-dominant sorting is carried out, specifically comprising:

according to the fitness value of each individual in the initialized population Y, the fitness value of each individual in the initialized population Y is subjected to dominance judgment one by one with the fitness values of other individuals, and two parameters n of each individual Y of the initialized population Y need to be calculated _y And g _y ，n _y G represents the number of individuals, g, of the initial population Y, which dominate the population Y _y For initializing a set of individuals in the population Y, which are dominated by the individuals Y, sequentially recording two parameters of each individual in the population Y; according to n of each individual _y Non-dominant stratification is performed: (1) Find all n in population Y _y Individual=0, and is held at the first level of the current non-dominant level, set F ₁ In (a) and (b); (2) For the current set F ₁ Each individual z of (a) whose dominant set of individuals is g _z Traversing g _z Performs a simple operation n 'on each individual h of (a)' _h ＝n _h -1, if n' _h Individual H is kept in pool H, n =0' _h Is a simple post-operation set F ₁ Number of individuals that innervate individual h; n is n _h For easy pre-operation set F ₁ Number of individuals that innervate individual h; (3) Record F ₁ The obtained individuals are the individuals with the first non-dominant level, H is taken as the current set, the operations (2) and (3) are repeated, and the operations are repeated, and layering is carried out until the whole initialization population Y is divided into a plurality of level sets F as shown in figure 8 ₁ Collection F ₂ Collection F ₃ Etc.

S330: selecting elite individuals in the population Y after the rapid non-dominant ranking;

as shown in fig. 8, from the fast non-dominant orderingCollection F of population Y ₁ (first layer non-dominant level), by arbitrarily generating an integer k (integer k ranges from 1-A, A is set F ₁ Total number of individuals in) a set F is selected according to an integer k ₁ The kth individual of (a) is designated as elite individual 1; in the same way from the set F of fast non-dominant ordered populations Y ₂ (second non-dominant level), an integer U is arbitrarily generated (the range of the integer U is 1-B, B is a set F ₂ Total number of individuals in) a set F is selected according to an integer U ₂ The U-th individual in (2) is taken as an elite individual; two individuals, elite individual number 1 and elite individual number 2, are considered herein to be currently preferred two different deployment schemes; record number 1 elite individual E ₁ ，E ₁ Respectively E is the three objective function values of (2) _{1_} a (i.e. objective function value of reliability of inter-satellite link of elite number 1), E _{1_} b (i.e. load balancing objective function value of elite number 1 individual LEO satellite network), E _{1_} c (namely, a delay objective function value of a control link of the elite individual LEO satellite network 1); record number 2 elite individuals E ₂ ，E ₂ Respectively E is the three objective function values of (2) _{2_} a (i.e. objective function value of reliability of inter-satellite link of elite number 2), E _{2_} b (i.e. load balancing objective function value of elite number 2 individual LEO satellite network), E _{2_} c (i.e., elite number 2 individual LEO satellite network control link delay objective function value).

S340: selecting individuals participating in crossing from elite individuals, and performing crossing operation on the individuals participating in crossing in a multi-point mixed crossing mode; randomly generating a number, and executing mutation operation on the individuals after the cross operation when the random number is smaller than the self-adaptive mutation probability; when the random number is greater than or equal to the adaptive mutation probability, not executing mutation operation on the individuals after the cross operation; constraint restoration is carried out on the offspring population after the crossover or mutation operation;

(1) Selecting individuals participating in crossover from elite individuals;

selecting individuals participating in crossing from elite individuals 1 and elite individuals 2 in a ratio mode, and marking the individuals as W;

reading population individuals X from the initial population Y one by one, and recording three objective function values of the individuals X as X _{_} a、X _{_} b、X _{_} c, combining the objective function value of the individual X with three objective function values E of the elite individual No. 1 _{1_} a、E _{1_} b、E _{1_} c is divided one by one to obtain three ratios of a, b and c, and if more than two values of the three ratios of a, b and c are smaller than or equal to 1, the method indicates that more than two (including two) objective function values in three objective function values in the population X are superior to or equal to the past elite individual 1, the elite individual 1 is selected as the individual W (namely, the intersecting individual); if more than two of the three ratios of a, b and c are greater than 1, selecting the elite individual No. 2 as the individual W; wherein, the three ratios of a, b and c are respectively expressed by the following formulas:

wherein a, b and c are all ratios; x is X _{_} a、X _{_} b and X _{_} c is the objective function value of the reliability of the inter-satellite link of the individual X, the load balancing objective function value of the LEO satellite network and the delay objective function value of the LEO satellite network control link respectively; e (E) _{1_} a、E _{1_} b、E _{1_} c are elite individuals E No. 1 respectively ₁ An objective function value of the reliability of the inter-satellite link, a load balancing objective function value of the LEO satellite network, and a LEO satellite network control link delay objective function value.

(2) Controller coding sequence X in individuals W involved in crossover _C Setting three crossing points, and carrying out crossing operation on individuals participating in crossing in a multi-point mixed crossing mode;

the cross operator adopted by the invention is as follows: randomly selecting an integer q with the value range of 3-64, and encoding a sequence X in a controller of an individual W participating in crossing _C Three crossing points are arranged and are respectively X _C The (q-2) th position, the (q+2) th position, X of the individual W by a multi-point mixed crossing manner _C X crossing the (q-2) th element of (b) to the individual X _C The q-th element of (B), X of individual W _C X crossing the q-th element of (B) to individual X _C In (q+2) th element, X of individual W _C The (q+2) th element of (a) crosses to X of individual X _C The (q-2) th element of the sequence is shown in FIG. 9, when q is 4, the three crossing points are respectively 2,4,6 positions, due to X _C The three-point hybrid intersection is carried out, so that a new distribution processing sequence of the multi-controller deployment scheme is formed, and the constraint limit of the original individual is destroyed, therefore, the constraint restoration is carried out on the individual after the intersection operation, and the constraint restoration refers to the random generation of a group of new multi-controller deployment strategies meeting the constraint limit.

(3) Randomly generating a number, and executing mutation operation on the individuals after the cross operation when the random number is smaller than the self-adaptive mutation probability; when the random number is greater than or equal to the self-adaptive variation probability, the variation operation is not executed on the individual after the cross operation;

randomly generating a number, namely generating a random number, wherein the value range of the random number is 0-1, and when the random number is smaller than the self-adaptive variation probability Ap _m When the random number is larger than or equal to the self-adaptive mutation probability, the mutation operation is not carried out on the individuals after the cross operation; ap as shown in FIG. 10 _m The value range of the algorithm is between 0.1 and 0.5, the global search is focused at the early stage of algorithm search, and Ap is focused on _m The value is smaller, the change amplitude is slower, and as the iteration number increases, ap is in the later searching process of the algorithm _m Quick improvement, local search in excellent individuals to improve search accuracy, ap _m Calculated by the following formula:

wherein Ap _m Is the adaptive mutation probability; gen is the current iteration number; maxgen is the maximum iteration number, and the maxgen takes a value of 200 in the invention.

The mutation probability of the traditional NSGA-III algorithm is always stable and unchanged, however, when the controller deployment strategy search is carried out, a large number of excellent populations are considered to be generated in the later period of the search, local search is required to be focused, and the population quality is optimized, so the invention introduces the self-adaptive mutation probability Ap _m 。

The mutation operation comprises the following steps:

the mutation of the traditional NSGA-III algorithm has strong randomness, so the invention changes the traditional mutation strategy, adopts the method of inserting mutation, and as shown in figure 11, codes the sequence X from the exchange of the individual X after the cross operation _S Randomly generating two insertion variation point positions P and K, and performing cross operation on X of the individual X _S The element at the P position in the array fragment is inserted into the K position, and the element arrays from the (P+1) fragment to the K fragment in the array fragment are moved towards the P direction.

For example, the insertion variation point positions p=3, k=6 in fig. 11, the switch code sequence X of the individual X after the crossover operation will be _S Performing insertion mutation operation to form a new coding sequence; the constraint limit of individuals in the initial population is broken through by a series of crossed and mutated child populations, and further constraint restoration (constraint restoration refers to a new multi-controller deployment strategy which meets the constraint limit and is randomly generated) is needed to be carried out on the crossed and mutated child populations, so that convenience is brought to the next elite selection strategy.

The invention improves the mutation and crossover operators of the traditional NSGA-III algorithm to enhance the exploration capability of the algorithm.

Elite selection strategies include:

marking the offspring population after the intersecting and mutation operation as Y', the population quantity as pop, calculating the fitness value of each individual in the offspring population, and carrying out population combination on the offspring population and the father population (namely, the initialized population Y); in order to avoid the influence of repeated individuals on the final non-dominant solution set, the combined population is required to be subjected to a deduplication operation, and the repeated individuals are deleted from the combined population; then, based on the fitness value of each individual, carrying out rapid non-dominant ranking on the combined population, selecting according to the limitations of non-dominant level and population scale, and keeping the individual (the individual after elite selection) in Y'; when screening a certain level, the number of stored individuals exceeds the population scale, at the moment, the individuals with sparse distribution of the level are screened out by adopting a reference point-based method and are reserved in Y 'until the population number pop is met, and finally Y is updated through the individuals Y' after elite selection.

S360: updating elite individuals based on the elite selected individuals;

and carrying out control judgment on the selected elite individuals and the current

elite individuals

1 and 2 one by one, if the selected elite individuals can control the

elite individuals

1 or 2, further updating the

elite individuals

1 and 2, otherwise, not updating.

S370: outputting a non-dominant (Pareto) optimal solution set and decoding to obtain an optimal SDN multi-controller deployment scheme;

repeating the steps S340-S360 to continuously optimizing and updating the population until the maximum iteration times maxgen are reached, decoding the Pareto optimal solution set of the updated population, and finally outputting the optimal SDN multi-controller deployment scheme with multi-objective optimization.

S400: and deploying the satellite network based on the optimal SDN multi-controller deployment scheme.

The beneficial effects of the technical scheme of the invention are described below by combining simulation experiments:

the invention aims to reduce the delay of LEO satellite network control links, improve the reliability of inter-satellite links and maintain the load balance of the LEO satellite network, and provides a multi-objective optimization model based on an improved NSGA-III algorithm (INSSGA-III)The deployment method of the satellite network SDN multi-controller is suitable for solving the deployment problem of the satellite network SDN multi-controller under a normal task by improving the traditional NSGA-III algorithm; and comparing with modified NSGA-II algorithm (INSSGA-II) and traditional NSGA-III algorithm; in the simulation experiment, the iteration number maxgen of each algorithm is 200, the population pop is 100, and the crossover probability P is high _c 0.5, variation probability P _m 0.5.

When the number of the set controllers is 10, pareto front edges formed by Pareto optimal solution sets of the three multi-objective genetic algorithms are shown in fig. 12; in fig. 12, the hypersurface formed by the Pareto optimal solution set of the INSGA-iii algorithm corresponding to the objective function value is at the forefront of other multi-objective genetic algorithms, and is the result of a detailed analysis simulation experiment, and the three-dimensional coordinate graph of fig. 12 is converted into a two-dimensional coordinate graph as shown in fig. 13-15; wherein, fig. 13 is a time delay and load balancing analysis chart, fig. 14 is a load balancing and reliability coefficient analysis chart, and fig. 15 is a reliability coefficient and time delay analysis chart; from fig. 12 to 15, it can be known that the INSGA-III algorithm is superior to other algorithms in terms of load balancing, reliability and time delay, and the effectiveness of the algorithm is fully proved.

In order to evaluate the overall optimizing performance of the algorithm, three targets are given different weights and then weighted and summed, and an overall performance objective function is established, wherein the overall performance objective function is expressed by the following formula:

W＝α×FAP+β×T+δ×B _a

wherein W is an overall performance objective function; FAP is the average failure probability of LEO satellite network; t is LEO satellite network control link time delay; b (B) _a Load balancing coefficients for the LEO satellite network; the alpha, the beta and the delta are respectively the average failure probability of the LEO satellite network, the delay of a control link of the LEO satellite network and the weight of a load balancing coefficient of the LEO satellite network, and the alpha, the beta and the delta epsilon (0, 1), wherein the alpha, the beta and the delta=1 are satisfied, and the weight can be adjusted according to the actual situation of the problem.

The simulation experiment of the invention sets the weight of each single target as alpha=beta=delta=1/3, the overall performance analysis table of each algorithm Pareto optimal solution set is shown in table 1, the running times in table 1 represent each running five times of each algorithm, if the optimal condition of the single target exists in three target values of each deployment scheme of each algorithm in each running result, the optimal condition of the single target is recorded, and the three target values are converted into the overall performance target function values as the numerical values in table 1; as can be seen from the data in Table 1, the INSSGA-III algorithm has better optimizing performance in the aspect of local single objective and has stronger optimizing capability than other algorithms in the aspect of the result analysis of global overall performance objective function values compared with the INSSGA-II algorithm and the NSGA-III algorithm.

Table 1 overall performance analysis table for Pareto optimal solution set for each algorithm

The INSSGA-III algorithm is further analyzed by setting different numbers of controllers; fig. 16 is a Pareto front-end diagram of each algorithm when the number of controllers is 12, and fig. 17 is a Pareto front-end diagram of each algorithm when the number of controllers is 8; it can be seen from a combination of fig. 16 and 17 that under different numbers of controllers, the INSGA-iii algorithm (modified NSGA-iii algorithm) is still superior to the INSGA-ii and NSGA-iii in terms of optimizing strategy, which fully demonstrates that the multi-controller deployment method is robust.

Example two

Fig. 18 is a schematic diagram of a deployment system for multiple controllers of a satellite network SDN according to an embodiment of the present invention, including a satellite network erection module, a multi-objective optimization model construction module, a calculation module, and a deployment module;

the deployment module is used for deploying the satellite network based on the optimal SDN multi-controller deployment scheme.

The computing module performs the following operations:

In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of the units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other form.

The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit.

The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a usb disk, a removable hard disk, a ROM, a RAM, a magnetic disk, or an optical disk, etc.

The foregoing is merely illustrative of the present invention, and the present invention is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present invention. Therefore, the protection scope of the invention is subject to the protection scope of the claims.

Claims

1. The deployment method of the satellite network SDN multi-controller is characterized by comprising the following steps of:

s100: erecting a satellite network architecture based on SDN;

wherein the constructing the multi-objective optimization model comprises:

constructing a reliability model of an inter-satellite link, an LEO satellite network load balancing model and an LEO satellite network control link time delay model;

the step S300 comprises the sub-steps of:

s340: selecting individuals participating in crossing from the elite individuals, and performing crossing operation on the individuals participating in crossing in a multi-point mixed crossing mode; randomly generating a random number, and executing mutation operation on the individuals after the cross operation when the random number is smaller than the self-adaptive mutation probability; when the random number is greater than or equal to the self-adaptive mutation probability, the mutation operation is not carried out on the individuals after the cross operation; constraint restoration is carried out on the offspring population after the crossover or mutation operation; the adaptive mutation probability is calculated by the following formula:

wherein Ap _m Is the adaptive mutation probability; gen is the current iteration number; maxgen is the maximum number of iterations;

2. The satellite network SDN multi-controller deployment method of claim 1, wherein a reliability model of inter-satellite links is formulated by:

wherein FAP is the average failure probability of the LEO satellite network; n is the number of LEO satellite exchanger nodes; i is LEO satellite controller node C _i Is the number of (2); j is LEO satellite switch node S _j Is the number of (2); c is a LEO satellite controller node set in the LEO satellite network; s is a set of all LEO satellite switch nodes in the LEO satellite network; AP (Access Point) _ij For LEO satellite controller node C _i And exchangeNode S _j Probability of normal communication between; h _ij As matrix elements, the relation between LEO satellite controller nodes and LEO satellite switch nodes uses M multiplied by N matrix H= [ H ] _ij ] _M×N And (3) representing.

3. The satellite network SDN multi-controller deployment method of claim 1, wherein LEO satellite network load balancing model is represented by the following formula:

is the average load of the LEO satellite controller nodes.

4. The satellite network SDN multi-controller deployment method of claim 1, wherein LEO satellite network control link delay model is formulated by:

wherein T is LEO satellite network control link time delay; i is LEO satellite controller node C _i Is the number of (2); c is a LEO satellite controller node set in the LEO satellite network; tcl (Tcl) _i Processing time delay for LEO satellite controller node; tcb _i The propagation delay between the LEO satellite switch node and the LEO satellite controller node is set; j is LEO satellite switch node S _j Is the number of (2); s is all LEO satellite switch nodes in LEO satellite networkA collection; d, d _ij For LEO satellite controller node C _i And switch node S _j Shortest distance of inter-links; vc is the speed of light in free space; lambda (lambda) _j Request rate for LEO satellite switch data stream; f (f) _i For LEO satellite controller node C _i Maximum processing capacity of (2); h _ij As matrix elements, the relation between LEO satellite controller nodes and LEO satellite switch nodes uses M multiplied by N matrix H= [ H ] _ij ] _M×N And (3) representing.

5. The method for deploying the satellite network SDN multiple controllers of claim 1, wherein the mutation operation is performed in step S340 by a method of inserting mutation.

6. The satellite network SDN multi-controller deployment method of claim 1, wherein the elite selection policy in step S350 comprises:

7. The system is characterized by comprising a satellite network erection module, a multi-objective optimization model construction module, a calculation module and a deployment module;

the multi-objective optimization model construction module is used for constructing a multi-objective optimization model based on a satellite network; the constructing the multi-objective optimization model comprises the following steps: constructing a reliability model of an inter-satellite link, an LEO satellite network load balancing model and an LEO satellite network control link time delay model;

wherein the computing module performs the following operations: