CN116801288A - Self-organizing network topology optimization method and system based on particle swarm and genetic algorithm - Google Patents
Self-organizing network topology optimization method and system based on particle swarm and genetic algorithm Download PDFInfo
- Publication number
- CN116801288A CN116801288A CN202310747381.9A CN202310747381A CN116801288A CN 116801288 A CN116801288 A CN 116801288A CN 202310747381 A CN202310747381 A CN 202310747381A CN 116801288 A CN116801288 A CN 116801288A
- Authority
- CN
- China
- Prior art keywords
- node
- particle
- nodes
- particles
- self
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 239000002245 particle Substances 0.000 title claims abstract description 203
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000005457 optimization Methods 0.000 title claims abstract description 38
- 230000002068 genetic effect Effects 0.000 title claims abstract description 34
- 230000035772 mutation Effects 0.000 claims abstract description 28
- 238000004891 communication Methods 0.000 claims description 17
- 230000006870 function Effects 0.000 claims description 16
- 238000004590 computer program Methods 0.000 claims description 9
- 238000010606 normalization Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 8
- 230000000694 effects Effects 0.000 description 2
- 238000002922 simulated annealing Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000007257 malfunction Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000007480 spreading Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a self-organizing network topology optimization method and a self-organizing network topology optimization system based on a particle swarm and a genetic algorithm, wherein in the method, particles are initialized through an initialization function, the fitness value of each particle is calculated, and the optimal particles in the population are updated; then, performing cross operation and mutation operation on each particle according to the optimal particles to obtain new particles, and recalculating the fitness value to update the optimal particles; and finally, judging whether a termination condition is reached, if so, the optimal particles are the final optimization scheme, and if not, performing the next iteration, thereby achieving the aim of self-organizing network topology optimization.
Description
Technical Field
The invention relates to the technical field of wireless communication, in particular to a self-organizing network topology optimization method and system combining particle swarm and genetic algorithm.
Background
The smart grid is used as a key infrastructure for guaranteeing normal operation of socioeconomic performance and is easy to be threatened by various security. The self-organizing network has wide points and complex structure, and the nodes in the network may malfunction due to a plurality of reasons, such as natural disasters, physical faults, malicious attacks and the like. Because of the high coupling of the self-organizing network information system and the physical system, the fault of one node may cause the load and energy consumption of other nodes to increase, thereby causing cascading failure, and finally spreading to the whole network, causing the whole network to collapse, and causing large-area power failure. Therefore, the ad hoc network needs to ensure stable operation under normal conditions, and needs to have a certain fault tolerance capability, so that the ad hoc network can maintain necessary functions under disturbance conditions. Robustness is a key attribute of a network system that represents the ability of a network to continue to maintain its basic functionality in the face of a fault or attack. To combat failures and attacks, and to increase the robustness of the ad hoc network, the network topology needs to be optimized.
Many students have made extensive studies on the problem of network topology and many network optimization methods have been proposed. For the NP-hard optimization problem which is discrete and nonlinear and has uncertainty, the optimal solution is difficult to obtain by adopting accurate mathematics, and an evolutionary algorithm is an effective tool for solving the problem, such as genetic algorithm, ant colony optimization, simulated annealing, particle swarm algorithm and the like. The particle swarm algorithm is based on a swarm intelligent algorithm, and has the characteristics of capability of solving complex problems, high convergence speed and good universality for different problems. The traditional particle swarm optimization is mainly used for continuous optimization, has the limitation of being in local optimization, and is not suitable for the scene of self-organizing network topology optimization.
Disclosure of Invention
The invention aims to provide a self-organizing network topology optimization method and system based on particle swarm and genetic algorithm, which solve the problem of maximally improving the self-organizing network topology robustness when a network is attacked.
In order to solve the technical problems, the technical scheme of the invention is as follows:
in a first aspect, a method for optimizing a self-organizing network topology based on a particle swarm and a genetic algorithm is provided, comprising: step S100: encoding each particle, each particle representing an optimized edge set for self-organizing network edge; step S200: initializing each particle;
step S300: calculating the fitness value of each particle based on the fitness function, and updating the optimal particles in the population;
step S400: and carrying out genetic algorithm operation on each particle according to the optimal particle to obtain a new particle, and recalculating the fitness value to update the optimal particle to obtain the optimal borderline scheme.
Further, in step S100, each particle is encoded by an integer, to obtain an encoded array of particles:
each particle corresponds to an array of length Γl, Γ representing the number q of attacks y After the nodes, the node number |G (q y ) When the I is smaller than a node number critical threshold delta|G| of the network capable of working normally, the number of effective nodes in the network is L, and L represents a sufficiently large constant; the length of each node in the array is L, and the position of (i-1) L of the array represents the ID of the node; the interval of (i-1) l+1 to i L represents connection to the node v i Is represented at node v i An optimized edge is added between the node v and other nodes, when the node v is connected i When the number of (2) is smaller than L, the remaining positions are filled with 0.
Further, in step S200, the input in the particle initializing step is the valid node set V e The node number is Γ, L is a constant, S is the total number of particles in the particle swarm; step S200 includes:
step S210: acquiring neighbor node set NB of nodes in effective node set i ；
Step S220: when initializing the front particle l, randomly selecting rd nodes from Γ effective nodes, initializing an optimized edge set of all nodes in the particle into an empty set, and then node v i From a set of neighbor nodes NB i In which the node v connected thereto is randomly selected j And updating their optimized edge sets;
step S230: updating the coding array of the particles based on the optimized edge sets of all the nodes; after all the particles are initialized, an initialized set of S particles is generated.
Further, step S210 specifically includes:
considering the limitation of node communication distance and the multimode communication characteristic of the self-organizing network;
when using the PLC communication mode, consideration needs to be given to the adjacent nodesWhen node v i And v j SINR between them is greater than threshold value of reliable communication of PLC link +.>When the node pair is considered to be capable of communicating through the PLC link, the alternative node v j Added to node v i PLC neighbor set NB of (E) i ' in>
In addition, the wireless communication mode also needs to consider the SINR between nodes, when node v i And v j Between (a) and (b)Greater than threshold->When the alternative node v is to j Added to node v i Wireless neighbor set NB i "in>
Thus, node v i Is NB i ＝NB i '∪NB i "next hop node set slave NB that adds an optimized edge to each selected node i Is selected from the group consisting of a plurality of combinations of the above.
Further, in step S300, the fitness function is:
wherein N is the total node number in the network, N-1 is a normalization factor, LC n The maximum connected subnetwork node number after removing n nodes in each attack is that Ft has a value range of (0, 0.5)]In the range, the network corresponding to the maximum value of 0.5 is a fully connected network;
recording the current optimal value G best =max (fit (l)), where fit (l) is the fitness value of particle l calculated by fitness function Ft.
Further, the step S400 specifically includes:
step S410: performing cross operation on each particle according to the optimal particle;
step S420: performing mutation operation on each particle;
step S430: according to the new particles obtained by the cross operation and the mutation operation, calculating the fitness value of the new particles again and updating the optimal particles;
step S440: outputting the global optimal fitness value G after the iteration times are reached best The corresponding optimal particles are the optimal scheme.
Further, in step S410, the cross operation is to adjust the current particles according to the overall optimal particles in the population;
from the active node set V e In selecting lambda different nodes v i Respectively obtaining a node v based on the current particle and the optimal particle i Lk of (2) i Andoptimizing edge setsThe method comprises the steps of carrying out a first treatment on the surface of the Wherein λ is less than or equal to Γ, i=1, 2 1 ,Lk 2 ,...,Lk λ Optimized edge set representing lambda nodes in the current particle,/->An optimized edge set representing lambda nodes in the optimal particle;
there are three situations in total during crossover operations:
case one: if the current node v i Is an optimized edge set of (1)And is different from the optimized edge in the optimal particleDisconnect node v i And an optimized edge set Lk i Connection of the intermediate node, then at node v i And optimize edge set->Establishing connection between nodes in the network;
and a second case: if the current node v i Is an optimized edge set of (1)And->Skipping the current node, and continuing to select other nodes for crossing;
case three: if the current node v i Is an optimized edge set of (1)Then directly at node v i And optimize edge set->A connection is established between the nodes in (a).
When lambda nodes complete the operation, new particles are formed.
Further, in step S420, the mutation operation is the adjustment of the particle itself to provide the diversity of solutions; when each particle executes mutation operation, mu different nodes are randomly selected as mutation nodes, and Lk 1 ,Lk 2 ,...,Lk μ An optimized edge set representing μ nodes in the current particle, then the following is performed on the μ nodes:
if the node v is selected i Is an optimized edge set of (1)And its neighbor set->At the time, slave node v i Neighbor set NB of (1) i Randomly selecting one node for connection to form a new Lk i A collection of edges;
if the node v is selected i Neighbor setWhen the node is not in operation, skipping the node;
if the node v is selected i When optimizing edge sets of (a)And its neighbor set->When Lk is i Disconnect between nodes in the set and then slave node v i Neighbor set NB of (1) i Selecting nodes to connect to form new Lk i A collection of edges; the selected node cannot be associated with Lk i Repeating the nodes in (a);
after all the mu nodes finish the operation, the mutation operation of the current particle is finished, and a new particle is formed.
In a second aspect, a self-organizing network topology optimization system based on particle swarm and genetic algorithm is provided, which is characterized in that: comprising a memory and a processor; the memory stores a computer program, and the program can realize the self-organizing network topology optimization method based on particle swarm and genetic algorithm when being executed by the processor.
In a third aspect, a computer readable storage medium is provided, on which a computer program is stored, where the computer program, when executed by a processor, implements the above-mentioned method for optimizing an ad hoc network topology based on particle swarm and genetic algorithm.
The invention has the following beneficial effects:
the self-organizing network topology optimization method based on the combination of the particle swarm and the genetic algorithm provided by the invention has the advantages that as robustness is a key attribute of a network system, the robustness represents the capability of the network to continuously maintain the basic functions when facing faults or attacks, and the network topology needs to be optimized in order to resist the faults and attacks and improve the robustness of the self-organizing network; after the self-organizing network topology optimization method provided by the invention is optimized, the number of added optimized edges is smaller than that of other algorithms while the optimization effect is higher than that of other algorithms. The invention can also provide a certain guide for enhancing the network topology elasticity and relieving the influence caused by malicious attack.
Drawings
FIG. 1 is a flow chart of an optimization method of the present invention;
FIG. 2 is a flowchart showing an optimization method according to the present invention;
FIG. 3 is a schematic diagram of a crossover operation in accordance with an embodiment of the present invention;
FIG. 4 is a schematic diagram showing mutation operation in the embodiment of the present invention;
FIG. 5 is a diagram of a network optimization model under malicious attack in an embodiment of the present invention; wherein the method comprises the steps of
FIG. 5 (a) is an initial network model diagram;
FIG. 5 (b) is a diagram of a network model after a malicious attack;
FIG. 5 (c) is a schematic diagram of adding an optimized edge;
fig. 5 (d) is an optimized network model diagram.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
Taking fig. 5 as an example, the present invention first shows the situation of network disintegration under a malicious attack in a real self-organizing network, and performs the edge optimization effect based on the network after the attack. To maximize the robustness of the enhanced network, a set of optimal edge sets needs to be searched, thereby creating an optimization problem:
wherein G is l ,G 0 Representing the addition of l optimized networks and the initial network, respectively, where l is an uncertainty value and constraint C1 represents that each optimized edge added has a gain on the robustness of the network. Constraint C2 indicates that the length of the optimized edge is limited, provided that the SINR between the node pair must be greater than a threshold. Constraint C3 indicates that the capacity of the nodes in the optimized network cannot exceed its maximum capacity. Constraint C4 indicates that the number of optimized edges is a positive integer and cannot be infinite, being less than a threshold. For the NP-hard optimization problem which is discrete and nonlinear and has uncertainty, the optimal solution is difficult to obtain by adopting accurate mathematics, and an evolutionary algorithm is an effective tool for solving the problem, such as genetic algorithm, ant colony optimization, simulated annealing, particle swarm algorithm and the like. The particle swarm algorithm is based on a swarm intelligent algorithm, and has the characteristics of capability of solving complex problems, high convergence speed and good universality for different problems. Therefore, the invention adopts the particle swarm algorithm to combine with the genetic algorithm to carry out global optimization, and provides the self-organizing network topology optimization method based on the combination of the particle swarm and the genetic algorithm, and a group of optimal edge sets are searched and added on the basis of the network after attack, so that the robustness of the self-organizing network is improved to the maximum.
Referring to fig. 1 and 2, the present invention provides a self-organizing network topology optimization method based on particle swarm and genetic algorithm, which comprises:
step S100: encoding each particle, each particle representing an optimized edge set for self-organizing network edge;
step S200: initializing each particle;
step S300: calculating the fitness value of each particle based on the fitness function, and updating the optimal particles in the population;
step S400: and carrying out genetic algorithm operation on each particle according to the optimal particle to obtain a new particle, and recalculating the fitness value to update the optimal particle to obtain the optimal borderline scheme.
The respective steps in fig. 1 are specifically described below.
In step S100, each particle is encoded by an integer to obtain an encoded array of particles; the method comprises the following steps:
each particle represents a solution, i.e., an addition scheme that optimizes the edge set; each particle corresponds to an array of length Γl, Γ representing the number q of attacks y After the nodes, the node number |G (q y ) When the I is smaller than a node number critical threshold delta|G| of the network capable of working normally, the number of effective nodes in the network is L, and L represents a sufficiently large constant; the length of each node in the array is L, and the position of (i-1) L of the array represents the ID of the node; the interval of (i-1) l+1 to i L represents connection to the node v i Is represented at node v i An optimized edge is added between the node v and other nodes, when the node v is connected i When the number of (2) is smaller than L, the remaining positions are filled with 0, and the length consistency of the particles is maintained.
In step S200, each particle is initialized; the initialization of each particle is to randomly generate a group of feasible solutions, when the particle l is initialized, rd nodes are randomly selected from Γ effective nodes, then one node is selected from neighbor node sets to be connected with the nodes, and the nodes are all in the communication range as the end points of the optimized edge, so that the feasibility of the solutions is ensured; the input in the particle initialization step is an effective node set V e The node number is Γ, L is a constant, S is in the particle swarmIs the total number of particles; specifically, step S200 includes:
step S210: acquiring neighbor node set NB of nodes in effective node set i The method comprises the steps of carrying out a first treatment on the surface of the The method comprises the following steps:
when adding edge sets among nodes, the limitation of node communication distance and the multimode communication characteristic of the self-organizing network need to be considered; with node v i For example, when using the PLC communication mode, consideration needs to be given to the between neighboring nodesWhen node v i And v j SINR between them is greater than threshold value of reliable communication of PLC link +.>When the node pair is considered to be capable of communicating through the PLC link, the alternative node v j Added to node v i PLC neighbor set NB of (E) i ' in>In addition, the wireless communication mode also needs to consider the SINR between nodes, when node v i And v j Between->Greater than threshold->When the alternative node v is to j Added to node v i Wireless neighbor set NB i "in>Thus, node v i Is NB i ＝NB i '∪NB i "next hop node set slave NB that adds an optimized edge to each selected node i Is selected from the group consisting of a plurality of combinations of the above.
Step S220: when the current particle l is initialized, selecting rd nodes randomly from Γ effective nodes, and optimizing edges of all nodes in the particleSet initialization as an empty set, then node v i From a set of neighbor nodes NB i In which the node v connected thereto is randomly selected j And updating their optimized edge sets;
step S230: updating the coding array of the particles based on the optimized edge sets of all the nodes; after all the particles are initialized, an initialization set of S particles is generated, and an initialization coding set Ans of the particles is calculated.
Therefore, the algorithm of step S200 is as follows:
wherein Ans is the coding array of the particles, the coding array of the particles is updated according to the optimized edge set added by the node, the node number corresponding to the other end of the optimized edge is added at the corresponding position of the node in the array, and the insufficient position is supplemented with 0.
In step S300, calculating an fitness value of each particle based on the fitness function, and updating the optimal particles in the population;
in the invention, the relative quality of the particle solution is evaluated according to the calculated fitness value, and in the updating iteration process of the algorithm, each particle needs to calculate the fitness value through the fitness function, so that the selection of a proper fitness function is very important. The robustness in the network is usually expressed by using the maximum connected subnetwork, and the objective function of the optimization problem is rewritten into the fitness function, where the fitness function is:
wherein N is the total node number in the network, N-1 is a normalization factor, LC n Is the node of the maximum connected subnetwork after removing n nodes by each attackThe value of Ft is within the range of (0, 0.5]In the range, the network corresponding to the maximum value of 0.5 is a fully connected network;
for each particle l=1:s, a particle fitness value fit (l) is calculated, the current optimum value G is recorded best =max (fit (l)), where fit (l) is the fitness value of particle l calculated by fitness function Ft.
In step S400, the iteration number is recorded by a loop variable t, then genetic algorithm operation is sequentially performed on each particle to obtain a new particle, and then the fitness value of the new particle is recalculated, if the fitness value is greater than the fitness value of the current particle, the new fitness value is replaced; the optimal particle G output after Max_Ite iteration is completed best The method is a final optimal scheme; the method has the advantages that the codes of the particles are discrete, so that the particle updating formula of the traditional particle swarm algorithm is replaced by introducing crossover and mutation operations in a genetic algorithm, information is exchanged between the particles and optimal particles through the crossover operation, and new particles are generated through the mutation operation, so that population diversity is improved. The step S400 specifically includes:
step S410: performing cross operation on each particle according to the optimal particle; the cross operation is to adjust the current particles according to the overall optimal particles in the population; the input being the active node set V e The node number selected by the cross operation is lambda, opt is the optimal particle, and l is the current particle;
first from the active node set V before performing the interleaving operation e In selecting lambda different nodes v i Constituting a set of intersecting nodes V c Then respectively obtaining a cross node set V based on the optimal particle opt and the current particle l c Middle node v i Lk of (2) i Andoptimizing an edge set; wherein λ represents the number of crossover operation nodes, λ is less than or equal to Γ, i=1, 2 i Representing particle v i Optimized edge set of->Representing the particle in the optimal particleSon v i Is defined by a set of optimized edges; lk (Lk) 1 ,Lk 2 ,...,Lk λ Optimized edge set representing lambda nodes in the current particle,/->An optimized edge set representing lambda nodes in the optimal particle;
there are three situations in total during crossover operations:
case one: if the current node v i Is an optimized edge set of (1)And is different from the optimized edge in the optimal particleDisconnect node v i And an optimized edge set Lk i Connection of the intermediate node, then at node v i And optimize edge set->Establishing connection between nodes in the network;
and a second case: if the current node v i Is an optimized edge set of (1)And->Skipping the current node, and continuing to select other nodes for crossing;
case three: if the current node v i Is an optimized edge set of (1)Then directly at node v i And optimize edge set->A connection is established between the nodes in (a).
When lambda nodes complete the operation, new particles are formed.
Therefore, the algorithm steps of the crossover operation of step S410 are:
specifically, the updated particle array Ans (l) =cross (V e Lambda, opt, l). Fig. 3 is a diagram illustrating a case in which the crossover operation is the most complicated with a simple example of three nodes, and a yellow position indicates the ID of a node, l=3, i.e., 3 association positions are reserved for each node to store the node IDs for establishing a connection with the current node. Suppose node v i Is selected as the crossover node, node v in the current particle i Connected is node v j In the optimal particles, v i And node v k And is connected with each other. At the time of crossing, node v is disconnected i And v j Connection between each other, node v in the code array of the particle i Finding v from the associated position of (a) j And delete, add v in the associated position k Simultaneously deleting node v j V in the associated position of (2) i And at node v k Adding v in the associated position of (2) i The particles obtained after the operation are updated.
Step S420: performing mutation operation on each particle; mutation manipulation is the adjustment of the particle itself to provide diversity of solutions; when each particle executes mutation operation, mu different nodes are randomly selected as mutation nodes, and Lk 1 ,Lk 2 ,...,Lk μ An optimized edge set representing μ nodes in the current particle, then the following is performed on the μ nodes:
if the node v is selected i Is an optimized edge set of (1)And its neighbor set->At the time, slave node v i Neighbor set NB of (1) i Randomly selecting one node for connection to formNew Lk i A collection of edges;
if the node v is selected i Neighbor setWhen the node is not in operation, skipping the node;
if the node v is selected i When optimizing edge sets of (a)And its neighbor set->When Lk is i Disconnect between nodes in the set and then slave node v i Neighbor set NB of (1) i Selecting nodes to connect to form new Lk i A collection of edges; the selected node cannot be associated with Lk i Repeating the nodes in (a);
after all the mu nodes finish the operation, the mutation operation of the current particle is finished, and a new particle is formed.
Therefore, the algorithm steps of the mutation operation of step S420 are as follows:
the specific meaning of the algorithm is as follows: the input being the active node set V e The number of nodes selected by the mutation operation is mu, l is the current particle, and a counting variable k is firstly set for recording the number of nodes for executing the mutation operation. When k < mu, first from V e Randomly selecting an unselected node v i Then obtain node v i Is (are) optimized edge set Lk i And the number of nodes len and neighbor set NB of nodes therein i And the number of neighbor nodes num. When node v i And NB (node B) i All of (3)Nodes all have connections or NB i When all nodes in the network are not connected, the cycle is ended, and one node is selected again. From NB i Is selected to form new Lk i '，Lk i The node in' is not present in Lk i Is a kind of medium. If len+.0, disconnect from Lk i Connection of the middle node and establishment of the connection with the Lk i Connection of the nodes in'. The code array of particle/is then updated. After the above operations are performed on the μ nodes, the original particles complete the Mutation operation to obtain an updated particle array Ans (l) =mutation (V e ，μ，l)。
FIG. 4 gives a simple example of a mutation operation, assuming node v is selected i As abrupt node, first disconnect the current associated location node, i.e. disconnect v j Then from v i Neighbor set NB of (1) i Is selected at random from another node v k Make the connection and at v k Associated position addition v of (2) i The nodes then get new particles after mutation.
Step S430: calculating the fitness value fit ' of the new particle again according to the new particle obtained by the crossover operation and the mutation operation, and if fit ' (l) > fit (l), updating fit (l) +% fit ' (l); wherein fit is the fitness value set of the particles.
Step S440: after all the particles finish the operation, updating the optimal particles of the population according to the fitness value; if G best <max (fit (l)), the current optimum value G is updated best =max (fit (l)), and outputs a global optimum fitness value G when the maximum number of iterations max_ite is reached best The corresponding optimal particles are the optimal scheme.
In other embodiments, the invention further provides a self-organizing network topology optimization system based on particle swarm and genetic algorithm, comprising a memory and a processor; the memory stores a computer program, and the program can realize the self-organizing network topology optimization method based on particle swarm and genetic algorithm when being executed by the processor.
In other embodiments, the present invention further provides a computer readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the above-mentioned self-organizing network topology optimization method based on particle swarm and genetic algorithm.
The invention is not related in part to the same or implemented in part by the prior art.
The foregoing is a further detailed description of the invention in connection with specific embodiments, and it is not intended that the invention be limited to such description. It will be apparent to those skilled in the art that several simple deductions or substitutions may be made without departing from the spirit of the invention, and these should be considered to be within the scope of the invention.
Claims (10)
1. The self-organizing network topology optimization method based on the particle swarm and the genetic algorithm is characterized by comprising the following steps of: comprising
Step S100: encoding each particle, each particle representing an optimized edge set for self-organizing network edge;
step S200: initializing each particle;
step S300: calculating the fitness value of each particle based on the fitness function, and updating the optimal particles in the population;
step S400: and carrying out genetic algorithm operation on each particle according to the optimal particle to obtain a new particle, and recalculating the fitness value to update the optimal particle to obtain the optimal borderline scheme.
2. The method for optimizing the self-organizing network topology based on the particle swarm and the genetic algorithm according to claim 1, wherein the method comprises the following steps: in step S100, each particle is encoded by an integer, and an encoded array of particles is obtained:
each particle corresponds to an array of length Γl, Γ representing the number q of attacks y After the nodes, the node number |G (q y ) The number of effective nodes in the network when the number is smaller than the threshold delta|G| of the nodes of the network which can work normally, and L represents oneA sufficiently large constant; the length of each node in the array is L, and the position of (i-1) L of the array represents the ID of the node; the interval of (i-1) l+1 to i L represents connection to the node v i Is represented at node v i An optimized edge is added between the node v and other nodes, when the node v is connected i When the number of (2) is smaller than L, the remaining positions are filled with 0.
3. The method for optimizing the self-organizing network topology based on the particle swarm and the genetic algorithm according to claim 1, wherein the method comprises the following steps: in step S200, the input in the particle initialization step is the active node set V e The node number is Γ, L is a constant, S is the total number of particles in the particle swarm; step S200 includes:
step S210: acquiring neighbor node set NB of nodes in effective node set i ；
Step S220: when initializing the front particle l, randomly selecting rd nodes from Γ effective nodes, initializing an optimized edge set of all nodes in the particle into an empty set, and then node v i From a set of neighbor nodes NB i In which the node v connected thereto is randomly selected j And updating their optimized edge sets;
step S230: updating the coding array of the particles based on the optimized edge sets of all the nodes; after all the particles are initialized, an initialized set of S particles is generated.
4. The method for optimizing the topology of the self-organizing network based on the particle swarm and the genetic algorithm according to claim 3, wherein the method comprises the following steps: the step S210 specifically includes:
considering the limitation of node communication distance and the multimode communication characteristic of the self-organizing network;
when using the PLC communication mode, consideration needs to be given to the adjacent nodesWhen node v i And v j SINR between them is greater than threshold value of reliable communication of PLC link +.>When the node pair is considered to be capable of communicating through the PLC link, the alternative node v j Added to node v i PLC neighbor set NB' i In (I)>
In addition, the wireless communication mode also needs to consider the SINR between nodes, when node v i And v j Between (a) and (b)Greater than a threshold valueWhen the alternative node v is to j Added to node v i Wireless neighbor set NB i In (I)>
Thus, node v i Is NB i ＝NB′ i ∪NB″ i Next hop node set slave NB that adds an optimized edge to each selected node i Is selected from the group consisting of a plurality of combinations of the above.
5. The method for optimizing the self-organizing network topology based on the particle swarm and the genetic algorithm according to claim 1, wherein the method comprises the following steps: in step S300, the fitness function is:
wherein N is the total node number in the network, N-1 is a normalization factor, LC n The maximum connected subnetwork node number after removing n nodes in each attack is that Ft has a value range of (0, 0.5)]In the range, the network corresponding to the maximum value of 0.5 is a fully connected network;
recording the current optimal value G best =max (fit (l)), where fit (l) is the fitness value of particle l calculated by fitness function Ft.
6. The method for optimizing the self-organizing network topology based on the particle swarm and the genetic algorithm according to claim 1, wherein the method comprises the following steps:
the step S400 specifically includes:
step S410: performing cross operation on each particle according to the optimal particle;
step S420: performing mutation operation on each particle;
step S430: according to the new particles obtained by the cross operation and the mutation operation, calculating the fitness value of the new particles again and updating the optimal particles;
step S440: outputting the global optimal fitness value G after the iteration times are reached best The corresponding optimal particles are the optimal scheme.
7. The method for optimizing the self-organizing network topology based on the particle swarm and the genetic algorithm according to claim 6, wherein the method comprises the following steps: in step S410, the cross operation is to adjust the current particles according to the overall optimal particles in the population;
from the active node set V e In selecting lambda different nodes v i Respectively obtaining a node v based on the current particle and the optimal particle i Lk of (2) i Andoptimizing an edge set; wherein λ is less than or equal to Γ, i=1, 2 1 ，Lk 2 ，...，Lk λ Optimized edge set representing lambda nodes in the current particle,/->An optimized edge set representing lambda nodes in the optimal particle;
there are three situations in total during crossover operations:
case one: if the current nodev i Is an optimized edge set of (1)And is different from the optimized edge in the optimized particle, i.e. +.>Disconnect node v i And an optimized edge set Lk i Connection of the intermediate node, then at node v i And optimize edge set->Establishing connection between nodes in the network;
and a second case: if the current node v i Is an optimized edge set of (1)Eye(s) for the treatment of a person suffering from a disorder>Skipping the current node, and continuing to select other nodes for crossing;
case three: if the current node v i Is an optimized edge set of (1)Then directly at node v i And optimize edge set->A connection is established between the nodes in (a).
When lambda nodes complete the operation, new particles are formed.
8. The method for optimizing the self-organizing network topology based on the particle swarm and the genetic algorithm according to claim 6, wherein the method comprises the following steps: in step S420, the mutation operation is the adjustment of the particle itself to provide the diversity of solutions; when each particle performs mutation operation, mu different nodes are randomly selected as mutationNode, lk 1 ，Lk 2 ，...，Lk μ An optimized edge set representing μ nodes in the current particle, then the following is performed on the μ nodes:
if the node v is selected i Is an optimized edge set of (1)And its neighbor set->At the time, slave node v i Neighbor set NB of (1) i Randomly selecting one node for connection to form a new Lk i A collection of edges;
if the node v is selected i Neighbor setWhen the node is not in operation, skipping the node;
if the node v is selected i When optimizing edge sets of (a)And its neighbor set->When Lk is i Disconnect between nodes in the set and then slave node v i Neighbor set NB of (1) i Selecting nodes to connect to form new Lk i A collection of edges; the selected node cannot be associated with Lk i Repeating the nodes in (a);
after all the mu nodes finish the operation, the mutation operation of the current particle is finished, and a new particle is formed.
9. The self-organizing network topology optimization system based on the particle swarm and the genetic algorithm is characterized in that: comprising a memory and a processor;
wherein the memory stores a computer program which when executed by the processor is capable of implementing the self-organizing network topology optimization method based on particle swarm and genetic algorithm according to any of claims 1-8.
10. A computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the particle swarm and genetic algorithm based ad hoc network topology optimization method according to any of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310747381.9A CN116801288B (en) | 2023-06-25 | 2023-06-25 | Self-organizing network topology optimization method and system based on particle swarm and genetic algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310747381.9A CN116801288B (en) | 2023-06-25 | 2023-06-25 | Self-organizing network topology optimization method and system based on particle swarm and genetic algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116801288A true CN116801288A (en) | 2023-09-22 |
| CN116801288B CN116801288B (en) | 2024-01-26 |
Family
ID=88034243
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310747381.9A Active CN116801288B (en) | 2023-06-25 | 2023-06-25 | Self-organizing network topology optimization method and system based on particle swarm and genetic algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116801288B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105160580A (en) * | 2015-07-13 | 2015-12-16 | 西安电子科技大学 | Symbol network structure balance of multi-objective particle swarm optimization based on decomposition |
| CN105430706A (en) * | 2015-11-03 | 2016-03-23 | 国网江西省电力科学研究院 | WSN (Wireless Sensor Networks) routing optimization method based on improved PSO (particle swarm optimization) |
| CN107229972A (en) * | 2017-03-10 | 2017-10-03 | 东莞理工学院 | A kind of global optimization based on Lamarch inheritance of acquired characters principle, search and machine learning method |
| CN108183860A (en) * | 2018-01-19 | 2018-06-19 | 东南大学 | Two-dimentional network-on-chip adaptive routing method based on particle cluster algorithm |
| US20180357584A1 (en) * | 2017-06-12 | 2018-12-13 | Hefei University Of Technology | Method and system for collaborative scheduling of production and transportation in supply chains based on improved particle swarm optimization |
| US20200191943A1 (en) * | 2015-07-17 | 2020-06-18 | Origin Wireless, Inc. | Method, apparatus, and system for wireless object tracking |
| CN113935275A (en) * | 2021-10-12 | 2022-01-14 | 福州大学 | Construction method of barrier X structure Steiner minimum tree under time sequence relaxation constraint |
| CN115952691A (en) * | 2023-03-10 | 2023-04-11 | 南京雷电信息技术有限公司 | Optimized station distribution method and device of multi-station passive time difference cross joint positioning system |
| US20230112223A1 (en) * | 2020-05-14 | 2023-04-13 | Fuzhou University | Multi-stage fpga routing method for optimizing time division multiplexing |
-
2023
- 2023-06-25 CN CN202310747381.9A patent/CN116801288B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105160580A (en) * | 2015-07-13 | 2015-12-16 | 西安电子科技大学 | Symbol network structure balance of multi-objective particle swarm optimization based on decomposition |
| US20200191943A1 (en) * | 2015-07-17 | 2020-06-18 | Origin Wireless, Inc. | Method, apparatus, and system for wireless object tracking |
| CN105430706A (en) * | 2015-11-03 | 2016-03-23 | 国网江西省电力科学研究院 | WSN (Wireless Sensor Networks) routing optimization method based on improved PSO (particle swarm optimization) |
| CN107229972A (en) * | 2017-03-10 | 2017-10-03 | 东莞理工学院 | A kind of global optimization based on Lamarch inheritance of acquired characters principle, search and machine learning method |
| US20180357584A1 (en) * | 2017-06-12 | 2018-12-13 | Hefei University Of Technology | Method and system for collaborative scheduling of production and transportation in supply chains based on improved particle swarm optimization |
| CN108183860A (en) * | 2018-01-19 | 2018-06-19 | 东南大学 | Two-dimentional network-on-chip adaptive routing method based on particle cluster algorithm |
| US20230112223A1 (en) * | 2020-05-14 | 2023-04-13 | Fuzhou University | Multi-stage fpga routing method for optimizing time division multiplexing |
| CN113935275A (en) * | 2021-10-12 | 2022-01-14 | 福州大学 | Construction method of barrier X structure Steiner minimum tree under time sequence relaxation constraint |
| CN115952691A (en) * | 2023-03-10 | 2023-04-11 | 南京雷电信息技术有限公司 | Optimized station distribution method and device of multi-station passive time difference cross joint positioning system |
Non-Patent Citations (2)
| Title |
|---|
| WEI REN等: "A Novel Control Plane Optimization Strategy for Important Nodes in SDN-IoT Networks", IEEE INTERNET OF THINGS JOURNAL, vol. 6, no. 2, XP011723697, DOI: 10.1109/JIOT.2018.2888504 * |
| 樊梦;: "基于WS小世界网络拓扑特性的自组织PSO算法", 大众科技, no. 10 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116801288B (en) | 2024-01-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101478426B (en) | Generation method and system for network topology model | |
| CN111935724B (en) | Wireless sensor network topology optimization method based on asynchronous deep reinforcement learning | |
| CN108494678A (en) | Alternate routing configuration method based on genetic algorithm and system | |
| CN114070773A (en) | Space network routing strategy based on shortest path length | |
| Hayashi | A new design principle of robust onion-like networks self-organized in growth | |
| CN116647459A (en) | Multi-agent co-evolution topological robustness optimization method for Internet of things | |
| CN115277115A (en) | Method and system for solving robust information propagation problem on network | |
| CN116801288B (en) | Self-organizing network topology optimization method and system based on particle swarm and genetic algorithm | |
| CN106658639A (en) | QG-OLSR routing method based on quantum genetic strategy | |
| CN111526555B (en) | Multi-hop routing path selection method based on genetic algorithm | |
| Gudla et al. | A reliable routing mechanism with energy-efficient node selection for data transmission using a genetic algorithm in wireless sensor network | |
| Nimbalkar et al. | Trust based energy efficient clustering using genetic algorithm in wireless sensor networks (teecga) | |
| Szlachcic | Fault tolerant topological design for computer networks | |
| Raei et al. | Optimal distributed algorithm for minimum connected dominating sets in wireless sensor networks | |
| Khelifa et al. | Enhancing Energy Efficiency in Mobile Ad hoc Networks using Genetic Algorithms | |
| Gao et al. | Exploiting social tie and SIR characteristic for cooperative jamming: A game model in a two-layer network | |
| Khalil et al. | CDS based reliable topology control in WSNs | |
| Balaji et al. | An efficient heuristic for the minimum connected dominating set problem on ad hoc wireless networks | |
| Oikonomou et al. | Probabilistic flooding coverage analysis in large scale wireless networks | |
| Akram et al. | Genetic algorithm-based optimized fuzzy adaptive path selection in wireless sensor networks | |
| Lin et al. | Reinforcement based communication topology construction for decentralized learning with Non-IID data | |
| CN115802390B (en) | Multi-channel wireless ad hoc network optimization method and device based on common channel neighbor ratio | |
| CN115883577B (en) | Block chain network clustering and data transmission method | |
| CN114599068B (en) | Wireless sensor network routing algorithm based on plant community behaviors | |
| Mutawa et al. | A novel steady‐state genetic algorithm approach to the reliability optimization design problem of computer networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |