CN110289980A - Using the method and system of learning automaton prediction pocket exchange network link - Google Patents
Using the method and system of learning automaton prediction pocket exchange network link Download PDFInfo
- Publication number
- CN110289980A CN110289980A CN201910395002.8A CN201910395002A CN110289980A CN 110289980 A CN110289980 A CN 110289980A CN 201910395002 A CN201910395002 A CN 201910395002A CN 110289980 A CN110289980 A CN 110289980A
- Authority
- CN
- China
- Prior art keywords
- node
- pair
- prediction
- learning automaton
- network link
- 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.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/147—Network analysis or design for predicting network behaviour
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W84/00—Network topologies
- H04W84/18—Self-organising networks, e.g. ad-hoc networks or sensor networks
Abstract
The present invention provides a kind of method and system using learning automaton prediction pocket exchange network link, this method comprises: obtaining the historical behavior information between opportunistic network interior joint pair；Type division is carried out according to the connection frequent degree to node pair according to historical behavior information, by the node in opportunistic network to being divided into live-vertex pair or inactive node pair；According to live-vertex to inactive node to building the network link prediction model based on learning automaton；Operational network link prediction model exports prediction result, to predict whether connection can be generated between opportunistic network interior joint pair.The present invention constructs network link prediction model by using based on the historical behavior information between opportunistic network interior joint pair with corresponding, it can more precisely predict a possibility that connection is generated between node pair, limitation caused by preventing the association attributes network-based topological attribute or node to be predicted, and support can be provided for upper layer Routing Protocol.
Description
Technical field
The present invention relates to opportunistic network technical fields, in particular to a kind of to predict pocket exchange network using learning automaton
The method and system of link.
Background technique
Opportunistic network (pocket exchange network) is a kind of new network framework developed by mobile ad-hoc network, it
The forwarded hop-by-hop of message can be realized using node motion bring chance of meeting under the network condition of segmentation, and finally deliver
To destination node.Concept of meeting refers to the primary connection occurred between node: when node enters within mutual communication range (ratio
In the point to point link of Wi-Fi or Bluetooth protocol distance) when, then it establishes the link, communicates, when the two is left each other
When communication range, then link disconnects, and stops communication.Due to the mobility of node, meeting in opportunistic network is opportunistic, and
Uncertainty, reliable end-to-end communication path is not present.
Opportunistic network carries out the message dilivery of multi-hop using the operating mode of " storage-carrying-forwarding ", if present node does not have
Have when encountering towards the next-hop node of destination node, with regard to buffered message, and with the suitable repeater-transmitter meeting of the mobile searching of node.
It improves the occasion the network architecture, infrastructure can not depended on or only need a small amount of infrastructure in remote highway, city
The functions such as message dilivery, content distribution, resource-sharing are realized under the scenes such as city's traffic, mobile social activity.But their performance and use
Family experience is largely dependent upon the transport services that opportunistic network can be provided.
In existing opportunistic network use process, the prediction of link is particularly important, at present both at home and abroad about using study
The research of automatic machine prediction pocket exchange network link can be mainly divided into: prediction technique based on similitude, based on time sequence
Prediction technique, the link prediction method based on machine learning for arranging analysis, but the above method is all based on the topological attribute of network
Or the association attributes of node are predicted the historical behavior feature without considering node, and then cause to predict to answer in physical link
With having some limitations in the process.
Summary of the invention
Based on this, the present invention provides a kind of method and system using learning automaton prediction pocket exchange network link,
For solving in the prior art, the link prediction application process caused by the historical behavior feature for not accounting for node is more
The problem of limitation.
In a first aspect, the present invention provides a kind of method using learning automaton prediction pocket exchange network link, institute
The method of stating includes:
Obtain the historical behavior information between opportunistic network interior joint pair；
Type division is carried out according to the connection frequent degree of node pair according to the historical behavior information, by the chance
Node in network is to being divided into live-vertex pair or inactive node pair；
The network link with the inactive node to building based on learning automaton is predicted according to the live-vertex
Model；
The network link prediction model is run, and exports prediction result, to predict the opportunistic network interior joint to it
Between whether can generate connection.
The above-mentioned method using learning automaton prediction pocket exchange network link, by using based on saving in opportunistic network
Historical behavior information of the point between constructs network link prediction model with corresponding, can more precisely predict to produce between node pair
A possibility that raw connection, it is therefore prevented that the office caused by being predicted the association attributes of network-based topological attribute or node
It is sex-limited, and can support effectively be provided for upper layer Routing Protocol.
Further, described that Type division is carried out according to the connection frequent degree of node pair according to the historical behavior information
The step of include:
Sliding-model control is carried out to the opportunistic network, and according to the corresponding corresponding figure of adjacency matrix building of network snapshots
Matrix；
Count the number in the figure matrix in each element for " 1 "；
According to activity threshold value by the node in the opportunistic network to be divided into the live-vertex to it is described non-live
Jump node pair.
Further, it is described according to the live-vertex to the inactive node to building based on learning automaton
The step of network link prediction model includes:
Respectively to the live-vertex to the inactive node to carry out random environment setting；
Operation is carried out to the random environment in the learning automaton according to default agent algorithms.
Further, it is described respectively to the live-vertex to the inactive node to carrying out random environment setting
Step includes:
Setting according to conventional topologies structural similarity index to the live-vertex to the random environment is carried out, it is described
Conventional topologies structural similarity index is common neighbours' index, AA index, resource allocation index or local path index；
Setting according to decision Tree algorithms to the inactive node to the random environment is carried out, the decision Tree algorithms
For ID3 algorithm, CART decision Tree algorithms or C4.5 algorithm.
Further, the random environment is a triple form, is expressed as<α, β, c>, wherein α is the random environment
Input set, β indicates that the random environment feeds back to the signal set of the learning automaton, and c expression acts select probability
Set.
Further, the default agent algorithms are as follows:
Wherein, parameter is rewarded in a expression, b indicates punishment parameter, r indicates the optional amount of action of the learning automaton agency,
β (k) indicates that the random environment feeds back to the signal of learning automaton, Pj(k) indicate that the k moment selects the probability of j-th of movement.
Second aspect, the present invention provides a kind of system using learning automaton prediction pocket exchange network link, packets
It includes:
Data obtaining module, for obtaining the historical behavior information between opportunistic network interior joint pair；
Node division module, for carrying out type according to the connection frequent degree of node pair according to the historical behavior information
It divides, by the node in the opportunistic network to being divided into live-vertex pair or inactive node pair；
Model building module, for automatic to study is based on to building with the inactive node according to the live-vertex
The network link prediction model of machine；
Data run module for running the network link prediction model, and exports prediction result, to predict the machine
Whether connection can be generated between nodes pair.
Further, the node division module is also used to:
Sliding-model control is carried out to the opportunistic network, and according to the corresponding corresponding figure of adjacency matrix building of network snapshots
Matrix；
Count the number in the figure matrix in each element for " 1 "；
According to activity threshold value by the node in the opportunistic network to be divided into the live-vertex to it is described non-live
Jump node pair.
The third aspect, the present invention provides a kind of mobile terminal, including memory, processor and storage are on a memory simultaneously
The computer program that can be run on a processor, the processor run the computer program so that the mobile terminal execution
The above-mentioned method using learning automaton prediction pocket exchange network link.
Fourth aspect, the present invention provides a kind of storage mediums, are stored thereon with computer program, and the program is by processor
The step of above-mentioned method that pocket exchange network link is predicted using learning automaton is realized when execution.
Detailed description of the invention
Fig. 1 is the method for predicting pocket exchange network link using learning automaton that first embodiment of the invention provides
Flow chart；
Fig. 2 is the method for predicting pocket exchange network link using learning automaton that second embodiment of the invention provides
Flow chart；
Fig. 3 to Fig. 5 is the opportunistic network link evolutionary process figure that second embodiment of the invention provides；
Fig. 6 is the system for predicting pocket exchange network link using learning automaton that third embodiment of the invention provides
Structural schematic diagram.
Specific embodiment
For the ease of more fully understanding the present invention, the present invention is carried out further below in conjunction with related embodiment attached drawing
It explains.The embodiment of the present invention is given in attached drawing, but the present invention is not limited in above-mentioned preferred embodiment.On the contrary, providing
The purpose of these embodiments be in order to make disclosure of the invention face more sufficiently.
Expression or logic and/or step described otherwise above herein in flow charts, for example, being considered use
In the order list for the executable instruction for realizing logic function, may be embodied in any computer-readable medium, for
Instruction execution system, device or equipment (such as computer based system, including the system of processor or other can be held from instruction
The instruction fetch of row system, device or equipment and the system executed instruction) it uses, or combine these instruction execution systems, device or set
It is standby and use.For the purpose of this specification, " computer-readable medium ", which can be, any may include, stores, communicates, propagates or pass
Defeated program is for instruction execution system, device or equipment or the dress used in conjunction with these instruction execution systems, device or equipment
It sets.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
The description of example " or " some examples " etc. means specific features, structure, material or spy described in conjunction with this embodiment or example
Point is included at least one embodiment or example of the invention.In the present specification, schematic expression of the above terms are not
Centainly refer to identical embodiment or example.Moreover, particular features, structures, materials, or characteristics described can be any
One or more embodiment or examples in can be combined in any suitable manner.
Referring to Fig. 1, predicting pocket exchange network link using learning automaton for what first embodiment of the invention provided
Method flow chart, comprising the following steps:
Step S10 obtains the historical behavior information between opportunistic network interior joint pair；
Wherein, the historical behavior information of node pair refers to the connection number between node pair in each network snapshots,
The mode of acquisition are as follows: whole network is sliced according to the size of timeslice, statistics is in each network snapshots interior nodes to it
Between connection number；
Step S20 carries out Type division according to the connection frequent degree of node pair according to the historical behavior information, will
Node in the opportunistic network is to being divided into live-vertex pair or inactive node pair；
Wherein, which can use the mode of alive threshold, the mode of active condition judgement to carry out the active section
Point to the division between the inactive node pair, and by the live-vertex to the area with the inactive node pair
Not, the foundation of subsequent network link prediction model is effectively facilitated；
Step S30, according to the live-vertex to the inactive node to building the network based on learning automaton
Link prediction model；
Wherein, learning automaton be a kind of adaptive decision-making unit by the interaction with random environment to improve its decision
Performance selects a series of more appropriate movements from one group of limited set of actions, so that its probability for being punished
The selection of minimum, movement selects more new formula to the movement of corresponding learning automaton based on each node, and the movement is made
For the input of random environment, random environment gives corresponding feedback according to the movement, last according to the feedback from random environment
Signal update acts select probability vector, repeatedly until meeting training termination condition.Learning automaton model is by two parts
Composition: random environment and learning automaton agency.Compared with traditional machine learning algorithm, the generalization ability of learning automaton is more
By force, link prediction precision is higher, specifically, the step can be by using the mode that random environment is arranged to carry out the network
The foundation of link prediction model；
Step S40 runs the network link prediction model, and exports prediction result, to predict in the opportunistic network
Whether connection can be generated between node pair；
Wherein, model output the result is that each in set of actions acts the probability value that is selected, be numeric type
Data；
In the present embodiment, net is constructed with corresponding by using based on the historical behavior information between opportunistic network interior joint pair
Network link prediction model can more precisely predict a possibility that connection is generated between node pair, it is therefore prevented that due to being based on network
Topological attribute or the association attributes of node predict caused limitation, and can effectively be mentioned for upper layer Routing Protocol
For support.
Fig. 2 to Fig. 5 is please referred to, pocket exchange network is predicted using learning automaton for what second embodiment of the invention provided
The flow chart of the method for link, the described method comprises the following steps:
Step S11 obtains the historical behavior information between opportunistic network interior joint pair；
Wherein, defining opportunistic network is G=(V, E), and wherein V is node set, and E is line set, on time by opportunistic network G
Between sequence be divided into a series of network snapshots, snapshot set G={ G1,G2,...,Gt, wherein Gt=(Vt,Et), GtWhen indicating t
The network topology structure figure at quarter, VtIndicate the set of t moment node, EtIndicate the set on t moment side；
Step S21 carries out sliding-model control to the opportunistic network according to the historical behavior information, and fast according to network
Corresponding figure matrix is constructed according to corresponding adjacency matrix；
Wherein, each of described figure matrix element is a string of binary characters, data handled by the model
Concentration contains node to the beginning Connection Time and terminates the Connection Time, in the present embodiment, according to the timeslice size of setting, i.e.,
Start to connect turn-off time difference between Connection Time and the last one node pair between first node pair in each timeslice
For the timeslice size, to obtain carrying out the opportunistic network effect of sliding-model control；
Step S31 counts the number in the figure matrix in each element for " 1 "；
Wherein, which, which counts the effect for the number that each element is 1, is: counting between each timeslice interior nodes pair
Connect number, effectively facilitate subsequent all nodes to be divided into live-vertex between inactive node pair draw
Divide, i.e. partitioning standards in the present embodiment are as follows: if the connection number between node pair is more than or equal to threshold in a timeslice
Value, then it is assumed that the node is on the contrary then belong to inactive node pair to belonging to live-vertex pair；
Step S41, according to activity threshold value by the node in the opportunistic network to be divided into the live-vertex to
The inactive node pair；
Wherein, different isochronous surface duration is also different for determination of the node to activity.In general, when single
Between slice duration it is longer, corresponding node activity threshold value also will be bigger.Since the sparse degree of different data collection network is different
Sample, so corresponding isochronous surface length is also different, the length of isochronous surface can be determined according to Chaotic Time Series Analysis.Mesh
Preceding its threshold value that can be considered determines formula are as follows:
Wherein N is network snapshots quantity, and TL indicates the corresponding duration of a network snapshots；
Step S51, respectively to the live-vertex to the inactive node to carry out random environment setting；
Wherein, it is described respectively to the live-vertex to the inactive node to carry out random environment setting the step of
Include:
Setting according to conventional topologies structural similarity index to the live-vertex to the random environment is carried out, it is described
Conventional topologies structural similarity index is common neighbours' index, AA index, resource allocation index or local path index；
Setting according to decision Tree algorithms to the inactive node to the random environment is carried out, the decision Tree algorithms
For ID3 algorithm, CART decision Tree algorithms or C4.5 algorithm；
Step S61 carries out operation to the random environment in the learning automaton according to default agent algorithms；
Wherein, the random environment is a triple form, is expressed as<α, β, c>, wherein α is the defeated of the random environment
Enter set, β indicates that the random environment feeds back to the signal set of the learning automaton, and c expression acts select probability collection
It closes；
Specifically, the default agent algorithms are as follows:
Wherein, parameter is rewarded in a expression, b indicates punishment parameter, r indicates the optional amount of action of the learning automaton agency,
β (k) indicates that the random environment feeds back to the signal of learning automaton, Pj(k) indicate that the k moment selects the probability of j-th of movement.
Step S71 runs the network link prediction model, and exports prediction result, to predict in the opportunistic network
Whether connection can be generated between node pair；
Wherein, the link prediction model algorithm pseudocode used in the present embodiment is described as follows:
Learning Automata Algorithm
Begin
Network adjacent matrix set M, maximum number of iterations Kmax, information entropy threshold Entromin, system information entropy EV,
Iteration count k, s, random environment Et, learning automaton hyper parameter a and b, selection movement 1 probability p1=0.5
While s<Smax do
While k<Kmax or EV<Entromin do
Each learning automaton is based on movement ProbabilityDistribution Vector and selects a movement j
The movement is acted on into random environment Et
Random environment feeds back to one signal of learning automaton
Select probability distribution vector is more preferably acted based on the feedback signal
Above procedure is repeated until meeting training termination condition
Continue to train into next stage, until the last stage
End While
Output: node is to the probability value for generating connection in future.
In the present embodiment, net is constructed with corresponding by using based on the historical behavior information between opportunistic network interior joint pair
Network link prediction model can more precisely predict a possibility that connection is generated between node pair, it is therefore prevented that due to being based on network
Topological attribute or the association attributes of node predict caused limitation, and can effectively be mentioned for upper layer Routing Protocol
For support.
Referring to Fig. 6, predicting pocket exchange network link using learning automaton for what third embodiment of the invention provided
System 100 structural schematic diagram, including data obtaining module 10, node division module 11, model building module 12 and data
Run module 13, in which:
Data obtaining module 10, for obtaining the historical behavior information between opportunistic network interior joint pair.
Node division module 11, for carrying out class according to the connection frequent degree of node pair according to the historical behavior information
Type divides, by the node in the opportunistic network to being divided into live-vertex pair or inactive node pair.
Further, the node division module 11 is also used to: to opportunistic network progress sliding-model control, and according to
The corresponding adjacency matrix of network snapshots constructs corresponding figure matrix；Count the number in the figure matrix in each element for " 1 "；
According to activity threshold value by the node in the opportunistic network to be divided into the live-vertex to the inactive node pair.
Model building module 12, for being based on study certainly to building to the inactive node according to the live-vertex
The network link prediction model of motivation.
In the present embodiment, the model building module 12 is also used to: respectively to the live-vertex to it is described inactive
Node is to progress random environment setting；The random environment is transported in the learning automaton according to default agent algorithms
It calculates.
In addition, the model building module 12 is also used to: according to conventional topologies structural similarity index to the active section
Setting of the point to the random environment is carried out, the conventional topologies structural similarity index are common neighbours' index, AA index, money
Source indicator of distribution or local path index；The inactive node is set to the random environment is carried out according to decision Tree algorithms
It sets, the decision Tree algorithms are ID3 algorithm, CART decision Tree algorithms or C4.5 algorithm.
Specifically, the random environment is a triple form, it is expressed as<α, β, c>, wherein α is the random environment
Input set, β indicate that the random environment feeds back to the signal set of the learning automaton, and c expression acts select probability collection
It closes.
The default agent algorithms are as follows:
Wherein, parameter is rewarded in a expression, b indicates punishment parameter, r indicates the optional amount of action of the learning automaton agency,
β (k) indicates that the random environment feeds back to the signal of learning automaton, Pj(k) indicate that the k moment selects the probability of j-th of movement.
Data run module 13 for running the network link prediction model, and exports prediction result, described in prediction
Whether connection can be generated between opportunistic network interior joint pair.
In the present embodiment, net is constructed with corresponding by using based on the historical behavior information between opportunistic network interior joint pair
Network link prediction model can more precisely predict a possibility that connection is generated between node pair, it is therefore prevented that due to being based on network
Topological attribute or the association attributes of node predict caused limitation, and can effectively be mentioned for upper layer Routing Protocol
For support.
The present embodiment additionally provides a kind of storage medium, is stored thereon with computer program, the program when being executed, including
Following steps:
Obtain the historical behavior information between opportunistic network interior joint pair；
Type division is carried out according to the connection frequent degree of node pair according to the historical behavior information, by the chance
Node in network is to being divided into live-vertex pair or inactive node pair；
The network link with the inactive node to building based on learning automaton is predicted according to the live-vertex
Model；
The network link prediction model is run, and exports prediction result, to predict the opportunistic network interior joint to it
Between whether can generate connection.The storage medium, such as: ROM/RAM, magnetic disk, CD.
Above embodiment described technical principles of the invention, and the description is merely to explain the principles of the invention, and
It cannot be construed to the limitation of the scope of the present invention in any way.Based on the explanation herein, those skilled in the art is not required to
Other specific embodiments of the invention can be associated by paying creative labor, these modes fall within of the invention
In protection scope.
The more specific example (non-exhaustive list) of computer-readable medium include the following: there are one or more wirings
Electrical connection section (electronic device), portable computer diskette box (magnetic device), random access memory (RAM), read-only memory
(ROM), erasable edit read-only storage (EPROM or flash memory), fiber device and portable optic disk are read-only
Memory (CDROM).In addition, computer-readable medium can even is that the paper that can print described program on it or other are suitable
Medium because can then be edited for example by carrying out optical scanner to paper or other media, interpret or when necessary with
Other suitable methods are handled electronically to obtain described program, are then stored in computer storage.
It should be appreciated that each section of the invention can be realized with hardware, software, firmware or their combination.Above-mentioned
In embodiment, software that multiple steps or method can be executed in memory and by suitable instruction execution system with storage
Or firmware is realized.It, and in another embodiment, can be under well known in the art for example, if realized with hardware
Any one of column technology or their combination are realized: having a logic gates for realizing logic function to data-signal
Discrete logic, with suitable combinational logic gate circuit specific integrated circuit, programmable gate array (PGA), scene
Programmable gate array (FPGA) etc..
Claims (10)
1. a kind of method using learning automaton prediction pocket exchange network link, which is characterized in that the described method includes:
Obtain the historical behavior information between opportunistic network interior joint pair；
Type division is carried out according to the connection frequent degree of node pair according to the historical behavior information, by the opportunistic network
In node to being divided into live-vertex pair or inactive node pair；
According to the live-vertex to the inactive node to building the network link prediction model based on learning automaton；
The network link prediction model is run, and exports prediction result, to predict to be between the opportunistic network interior joint pair
It is no to generate connection.
2. the method according to claim 1 using learning automaton prediction pocket exchange network link, which is characterized in that
It is described according to the historical behavior information according to node pair connection frequent degree carry out Type division the step of include:
Sliding-model control is carried out to the opportunistic network, and corresponding figure square is constructed according to the corresponding adjacency matrix of network snapshots
Battle array；
Count the number in the figure matrix in each element for " 1 "；
According to activity threshold value by the node in the opportunistic network to be divided into the live-vertex to the inactive section
Point pair.
3. the method according to claim 1 using learning automaton prediction pocket exchange network link, which is characterized in that
It is described according to the live-vertex to the inactive node to building the network link prediction model based on learning automaton
The step of include:
Respectively to the live-vertex to the inactive node to carry out random environment setting；
Operation is carried out to the random environment in the learning automaton according to default agent algorithms.
4. the method according to claim 3 using learning automaton prediction pocket exchange network link, which is characterized in that
It is described to include: to the step of carrying out random environment setting to the inactive node to the live-vertex respectively
Setting according to conventional topologies structural similarity index to the live-vertex to the random environment is carried out, the tradition
Topological structure similarity indices are common neighbours' index, AA index, resource allocation index or local path index；
Setting according to decision Tree algorithms to the inactive node to the random environment is carried out, the decision Tree algorithms are
ID3 algorithm, CART decision Tree algorithms or C4.5 algorithm.
5. the method according to claim 3 using learning automaton prediction pocket exchange network link, which is characterized in that
The random environment is a triple form, is expressed as<α, β, c>, wherein α is the input set of the random environment, and β is indicated
The random environment feeds back to the signal set of the learning automaton, and c expression acts select probability set.
6. the method according to claim 5 using learning automaton prediction pocket exchange network link, which is characterized in that
The default agent algorithms are as follows:
Wherein, a indicates that reward parameter, b indicate that punishment parameter, r indicate that the learning automaton acts on behalf of optional amount of action, β (k)
Indicate that the random environment feeds back to the signal of learning automaton, Pj(k) indicate that the k moment selects the probability of j-th of movement.
7. a kind of system using learning automaton prediction pocket exchange network link characterized by comprising
Data obtaining module, for obtaining the historical behavior information between opportunistic network interior joint pair；
Node division module is drawn for carrying out type according to the connection frequent degree of node pair according to the historical behavior information
Point, by the node in the opportunistic network to being divided into live-vertex pair or inactive node pair；
Model building module, for according to the live-vertex to the inactive node to building based on learning automaton
Network link prediction model；
Data run module for running the network link prediction model, and exports prediction result, to predict the chance net
Whether connection can be generated between network interior joint pair.
8. the system according to claim 7 using learning automaton prediction pocket exchange network link, which is characterized in that
The node division module is also used to:
Sliding-model control is carried out to the opportunistic network, and corresponding figure square is constructed according to the corresponding adjacency matrix of network snapshots
Battle array；
Count the number in the figure matrix in each element for " 1 "；
According to activity threshold value by the node in the opportunistic network to be divided into the live-vertex to the inactive section
Point pair.
9. the system according to claim 7 using learning automaton prediction pocket exchange network link, which is characterized in that
The model building module is also used to:
Respectively to the live-vertex to the inactive node to carry out random environment setting；
Operation is carried out to the random environment in the learning automaton according to default agent algorithms.
10. the system according to claim 9 using learning automaton prediction pocket exchange network link, feature exist
In the model building module is also used to:
Setting according to conventional topologies structural similarity index to the live-vertex to the random environment is carried out, the tradition
Topological structure similarity indices are common neighbours' index, AA index, resource allocation index or local path index；
Setting according to decision Tree algorithms to the inactive node to the random environment is carried out, the decision Tree algorithms are
ID3 algorithm, CART decision Tree algorithms or C4.5 algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910395002.8A CN110289980A (en) | 2019-05-13 | 2019-05-13 | Using the method and system of learning automaton prediction pocket exchange network link |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910395002.8A CN110289980A (en) | 2019-05-13 | 2019-05-13 | Using the method and system of learning automaton prediction pocket exchange network link |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110289980A true CN110289980A (en) | 2019-09-27 |
Family
ID=68001932
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910395002.8A Pending CN110289980A (en) | 2019-05-13 | 2019-05-13 | Using the method and system of learning automaton prediction pocket exchange network link |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110289980A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112929445A (en) * | 2021-02-20 | 2021-06-08 | 山东英信计算机技术有限公司 | Recommendation system-oriented link prediction method, system and medium |
| CN113300890A (en) * | 2021-05-24 | 2021-08-24 | 同济大学 | Self-adaptive communication method of networked machine learning system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101771964A (en) * | 2010-01-06 | 2010-07-07 | 北京航空航天大学 | Information correlation based opportunistic network data distributing method |
| CN104378229A (en) * | 2014-10-30 | 2015-02-25 | 东南大学 | Link prediction method for opportunity network |
| CN105792250A (en) * | 2016-01-14 | 2016-07-20 | 南昌航空大学 | Method for expressing opportunistic sensor network connectivity by whole network connectivity |
| US20180254958A1 (en) * | 2017-03-03 | 2018-09-06 | Nec Laboratories America, Inc. | Link prediction with spatial and temporal consistency in dynamic networks |
| CN109347697A (en) * | 2018-10-10 | 2019-02-15 | 南昌航空大学 | Opportunistic network link prediction method, apparatus and readable storage medium storing program for executing |
-
2019
- 2019-05-13 CN CN201910395002.8A patent/CN110289980A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101771964A (en) * | 2010-01-06 | 2010-07-07 | 北京航空航天大学 | Information correlation based opportunistic network data distributing method |
| CN104378229A (en) * | 2014-10-30 | 2015-02-25 | 东南大学 | Link prediction method for opportunity network |
| CN105792250A (en) * | 2016-01-14 | 2016-07-20 | 南昌航空大学 | Method for expressing opportunistic sensor network connectivity by whole network connectivity |
| US20180254958A1 (en) * | 2017-03-03 | 2018-09-06 | Nec Laboratories America, Inc. | Link prediction with spatial and temporal consistency in dynamic networks |
| CN109347697A (en) * | 2018-10-10 | 2019-02-15 | 南昌航空大学 | Opportunistic network link prediction method, apparatus and readable storage medium storing program for executing |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112929445A (en) * | 2021-02-20 | 2021-06-08 | 山东英信计算机技术有限公司 | Recommendation system-oriented link prediction method, system and medium |
| CN112929445B (en) * | 2021-02-20 | 2022-06-07 | 山东英信计算机技术有限公司 | Recommendation system-oriented link prediction method, system and medium |
| CN113300890A (en) * | 2021-05-24 | 2021-08-24 | 同济大学 | Self-adaptive communication method of networked machine learning system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Tong et al. | An energy-efficient multipath routing algorithm based on ant colony optimization for wireless sensor networks | |
| Iyengar et al. | Biologically inspired cooperative routing for wireless mobile sensor networks | |
| Zivan et al. | Explorative anytime local search for distributed constraint optimization | |
| Patel et al. | A hybrid ACO/PSO based algorithm for QoS multicast routing problem | |
| CN108989133B (en) | Network detection optimization method based on ant colony algorithm | |
| Wu et al. | Multicast routing with multiple QoS constraints in ATM networks | |
| Yin et al. | Niched ant colony optimization with colony guides for QoS multicast routing | |
| Qu et al. | Particle swarm optimization for the Steiner tree in graph and delay-constrained multicast routing problems | |
| CN106658539B (en) | Mobile path planning method for mobile data collector in wireless sensor network | |
| CN104378229A (en) | Link prediction method for opportunity network | |
| Ahmadi et al. | A hybrid algorithm for preserving energy and delay routing in mobile ad-hoc networks | |
| CN110289980A (en) | Using the method and system of learning automaton prediction pocket exchange network link | |
| Li et al. | An ant colony optimization metaheuristic for single-path multicommodity network flow problems | |
| Al-Hamid et al. | Vehicular intelligence: Towards vehicular network digital-twin | |
| CN107040884A (en) | A kind of mobile ad hoc network data transmission method based on neighborhood strong connectedness | |
| CN113259242A (en) | Method and device for networking field area network | |
| Król | On modelling social propagation phenomenon | |
| Chen | Simulation model of AI-assisted cognitive routing algorithm for the dynamic optical network in business | |
| CN103368770A (en) | Gateway level topology-based self-adaptive ALM overlay network constructing and maintaining method | |
| Poojary et al. | Ant colony optimization routing to mobile ad hoc networks in urban environments | |
| Guoying et al. | Multicast routing based on ant algorithm for delay-bounded and load-balancing traffic | |
| Rekik et al. | Geographic greedy routing with aco recovery strategy graco | |
| Morgan | An ant colony approach to regenerator placement with fault tolerance in optical networks | |
| Jun et al. | Improved method of ant colonies to search independent data transmission routes in WSN | |
| Mahseur et al. | Using bio-inspired approaches to improve the quality of service in a multicast routing |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190927 |
|
| RJ01 | Rejection of invention patent application after publication |