CN110019100A - For the big data analysis system of LTE network problem - Google Patents
For the big data analysis system of LTE network problem Download PDFInfo
- Publication number
- CN110019100A CN110019100A CN201711494416.3A CN201711494416A CN110019100A CN 110019100 A CN110019100 A CN 110019100A CN 201711494416 A CN201711494416 A CN 201711494416A CN 110019100 A CN110019100 A CN 110019100A
- Authority
- CN
- China
- Prior art keywords
- data
- layer
- big data
- lte network
- data analysis
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/10—File systems; File servers
- G06F16/18—File system types
- G06F16/182—Distributed file systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/25—Integrating or interfacing systems involving database management systems
- G06F16/254—Extract, transform and load [ETL] procedures, e.g. ETL data flows in data warehouses
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Artificial Intelligence (AREA)
- Genetics & Genomics (AREA)
- Physiology (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses the big data analysis system of a kind of pair of LTE network problem, which includes: data active layer, acquisition layer, convergence layer and application layer；Wherein, the data active layer is practical adjacent by physical location or non-conterminous several data sources form, to provide multi-source heterogeneous data information needed for system；The acquisition layer is responsible for accurately and rapidly completing the acquisition tasks from the data active layer, and clears up abnormal data, is the basis of big data analysis；The convergence layer is responsible for establishing the data sharing center based on HDFS using data provided by acquisition layer, and handles it；The application layer provides unified access and control interface using the processing result that the convergence layer provides for user and extension application.
Description
Technical field
The invention belongs to big data analysis technical fields, are related to a kind of big data analysis system for LTE network problem.
Background technique
With the development of 4G network, LTE gradually substitutes 2G/3G network as user's preferred network, forms and is with LTE
It is main, take into account the operating mode across manufacturer, cross-mode of 2G/3G network.Under the new situation, ever-increasing Optimization Work amount and phase
It is also more obvious to the contradiction between insufficient staffing.
In real network Optimization Work, the complexity of problem analysis is often determined using the accuracy of data, and
The validity of analysis method and the ease for use of external tool then determine the height of Resolving probiems efficiency.
Therefore, network data, reasonable analysis method and effective external tool how to be made full use of to alleviate optimization
Pressure and raising working efficiency, become network operator's major issue in need of consideration.
Summary of the invention
It is an object of that present invention to provide a kind of big data analysis systems for LTE network problem, in order to overcome new situations
Under contradiction between ever-increasing network optimization amount and the staffing of relative deficiency, using big data analysis technology,
Existing net multi-source data is accurately extracted with the mode of genetic algorithm and BP neural network by reasonable construction model, mark is added
Label are carried out data sharing and big data analysis in a manner of procedure, automation, are carried out in current LTE network by convergence
Network problem positioning, forms the very high automated solution of a set of feasibility, realizes the effect for reducing labor workload, thus
Change existing working method.
In order to solve the above technical problems, the present invention adopts the following technical scheme that: a kind of big number for LTE network problem
According to analysis system, which includes: data active layer, acquisition layer, convergence layer and application layer；Wherein, the data active layer is practical
It is adjacent by physical location or non-conterminous several data sources form, to provide multi-source heterogeneous data information needed for system；Institute
It states acquisition layer to be responsible for accurately and rapidly completing the acquisition tasks from the data active layer, and abnormal data is cleared up, be big
The basis of data analysis；The convergence layer is responsible for establishing using data provided by acquisition layer based on HDFS (distributed field system
System) data sharing center, and it is handled；The application layer is user using the processing result that the convergence layer provides
And extension application provides unified access and control interface.
Further, the convergence layer uses the big data processing mode of standard, and all original and machined data are united
One is stored in HBase, and completes data cleansing, data check to it, formulates label, aggregation scene and associating policy processing.
Further, the application layer is by customization standard interface module, system core function application module, process function
Module and authority management module composition.
Further, the customization standard interface module is responsible for extension application access data sharing center.
Further, the system core function application module includes label, scene, strategy, parameter management and assessment function
Energy.
Further, the process functional module is responsible for providing the production procedure based on permission and ensures network analysis result
Legitimacy in terms of automated execution.
Further, the authority management module is responsible for realizing stringent function control and user's control, lifting system peace
Entirely.
The present invention have compared with prior art it is below the utility model has the advantages that
The present invention program in order to overcome under the new situation the personnel of ever-increasing network optimization amount and relative deficiency match
Contradiction between standby, using big data analysis technology, by reasonable construction model, with the side of genetic algorithm and BP neural network
Formula accurately extracts existing net multi-source data, and label is added, and by convergence, data sharing is carried out in a manner of procedure, automation
And big data analysis, the very high automated solution of a set of feasibility is formed, realizes the effect for reducing labor workload.
Detailed description of the invention
Fig. 1 is the general frame figure for the big data analysis system of LTE network problem.
Fig. 2 is the main process figure for the big data analysis system of LTE network problem.
Fig. 3 is the big data analysis systematic difference algorithm flow chart for LTE network problem.
Specific embodiment
With reference to the accompanying drawing and specific embodiment to the present invention carry out in further detail with complete explanation.It is understood that
It is that described herein the specific embodiments are only for explaining the present invention, rather than limitation of the invention.
Referring to Fig.1, a kind of big data analysis system for LTE network problem of the invention, which includes: data source
Layer, acquisition layer, convergence layer and application layer；Wherein, the data active layer is practical adjacent or non-conterminous several by physical location
Data source composition, to provide multi-source heterogeneous data information needed for system；The acquisition layer be responsible for accurately and rapidly complete from
The acquisition tasks of the data active layer, and abnormal data is cleared up, it is the basis of big data analysis.
The convergence layer uses the big data processing mode of standard, supports isomery storage, establishes based on HDFS (distributed text
Part system) data sharing center, all initial data and machined data (MapReduce completion) are uniformly stored in
In HBase, provide necessary underlying services for ETL, and ETL complete data cleansing, data check, formulate label, aggregation scene,
The functions such as associating policy.
The application layer is presented using the processing result that the convergence layer provides as system, and the function of 4 aspects is provided:
Customization standard interface facilitates other systems to access data sharing center；System core functional application includes label, scene, plan
The contents such as summary, parameter management, assessment；Process function provides the production procedure based on permission, it is ensured that network analysis result is automatic
Change the legitimacy for executing aspect；Stringent function control and user's control, lifting system safety are realized in rights management.
The data sharing center of this system be towards multiple production systems, while need acquire, processing, share TB grades of data
Data warehouse.Data sharing center needs support never homologous ray and concurrently acquires mass data, since data source is in specific feelings
The problems such as increasing under condition there may be gather disappearance, collection period, this system supports mass data acquisition to appoint in addition to needing to concentrate
Outside the scheduling of business, data check and complicated data filling mining task schedule are also supported.
The task schedule of same data source uses the scheduling mode of concurrent multinode, while improving efficiency, effectively protects
Hinder the reliability of scheduling.Change can be artificially arranged in task schedule number of nodes, avoid the occurrence of hardware bottleneck.Data are acquired to draw by region
Point can dynamic regulation, avoid load unbalanced.Traffic control is combined using aggregative weighted algorithm and penalty mechanism, is avoided frequently
Start filling mining task.
Data sharing center extracts the different types of data for entering data warehouse, is cleaned, is processed, is summarized, is arranged,
By the data integration of various dimensions in not homologous ray to data warehouse, tag attributes are increased to all characteristics, pass through label-like
State determines that critical data, critical data are divided into according to thematic tract tissue and according to application or process demand according to network element dimension small
Area, base station, region, districts and cities；It is divided into hour, busy, day, week, the moon according to time dimension, and providing standardized interface is it
His system provides data and supports.
Data warehouse is found the problem on network element or time dimension by the critical data of label, is found in the form of scene
Problem, and combine existing tactful compositional optimization task issues solution by system, the entire problem-solving process of closed loop, most
End form is at automatic network optimization system.
This system be it is data-centered, by process through the functional automatic intelligent analysis system of institute.Therefore, it presses
Process can be divided into 3 sections by two key points (data loading and schemes generation) of data consumption:
1. data loading verifies process；
2. data characterization, which is extracted, judges that binding rule determines protocol procedures with fitting；
3. Automatic parameter modification process.
Main flow is made of several sub-processes, and definition passes through the interactive interfacing of process class in system, completes process and function
Combination；There is the movement of automation in function and influence the operation of existing net state, user right and function are solidified by process
Boundary completes information exchange by different process interfaces.Main flow design is as shown in Figure 2.
This system is attempted data being changed into information on the basis of realizing automation, and the characteristics of according to information itself,
Information attribute is taken out, information modeling is emulated, intelligence is done step-by-step by machine learning.It can divide according to flow chart of data processing
It is intelligent acquisition, pretreatment and statistics, data analysis and data mining respectively for 3 steps:
(1) intelligent acquisition
By arranging general acquisition frame, a plurality of types of data are stored by its own data characteristics and demand into number
According to Sharing Center.Frame itself provides the functions such as input, output, intermediate buffering, verifying verification, and supports to handle a variety of data knots
Structure.
(2) pretreatment and statistics
In face of effective analysis of mass data, the pretreatment mode for carrying out data by manually setting label main at present is real
It is existing.First stage of construction determined based on artificial it is pretreated in a manner of, method, later can be using being given and to user behavior analysis
The preprocess method recommended out.
(3) data analysis and data mining
Based on label before, data summarization, general analysis and characteristic are carried out to related data by scene and analyzed, with full
The analysis demand of sufficient user, basic analyzing method has verification, different grain size to summarize, otherness inspection, correlation analysis, degree of overlapping
Analysis, probability of interference analysis, factorial analysis, variance analysis, regression analysis, optimal subset analysis, clustering, multivariate analysis,
Principal component analysis, correlation rule, decision tree, genetic algorithm and neural network.
System on the basis of automation, gradually addition and optimization algorithm, given by partial analysis work and decision-making work
System is completed, and artificial intelligence is done step-by-step.
Positioning problems
Network performance could be accurately positioned in the apparent situation of directive property, it will usually by wireless environment, user location, use
The joint effect of many factors such as family business, user behavior and system resource is difficult to go quantitative analysis by simple factor
Network performance.
And influence there are many kinds of the factors of network performance, geological disaster is arrived greatly, it is small to move to ant, it is likely to cause net
Network performance shows entirely different performance, this can be regarded as a kind of extremely complex multivariate nonlinear function relationship.At this
In the case of kind, it is often necessary to which veteran expert is by being constantly trying to solve the problems, such as, promoting quality.And it throws into question
The reason range reachable far more than manpower, therefore be often in actual operation "retain the large, release the small", it can only solve more serious
Network performance problems can only then abandon problem stealthy or that coverage is smaller.
By the training analysis to multidimensional data, multiple nonlinear statistical method model and analysis method are constructed, to various dimensions
Label characteristics combined influence carries out mining analysis, is to judge that Same Scene network performance and one kind of its configuration parameter correlation have
Effect is attempted.
This system in such a way that BP neural network combines, is solved above-mentioned multi-parameter nonlinear model and asked using genetic algorithm
Topic.BP neural network can accurately predict challenge, but vulnerable to the influence overtrained, exist simultaneously training speed
Spend slow disadvantage；Genetic algorithm is global optimization search, and lookup optimal solution ability is good, and robustness is strong, but degree of fitting letter
Number changes greatly.The two is combined using BP neural network to weight optimal initialization, and uses the BP neural network method evolved,
With genetic algorithm substitution error against propagation algorithm, to avoid the defect of gradient descent method.It can effectively improve BP nerve in this way
The strong disadvantage of the structure randomness of network, enhances the Generalization Capability of BP neural network, and search does not require problem letter in space
Number is smooth can continuously to lead, as long as meeting objective function within the scope of corresponding constraint condition can solve, effect is much better than single method
It is generating as a result, even better than radial basis function network.
Network performance is that many factors are coefficient as a result, including parameter, device characteristics, user behavior, wireless environment
Etc. factors, and be not easy to prove the degree that these correlative factors work.But BP neural network is by very more and fairly simple
Processing neuron interactive connection constitute system, possess excellent fault-tolerant and learning performance.It is its global adaptability, non-linear
It is all very suitable to solve network performance problems with fault-tolerant ability, and utilizes genetic algorithm to the weight in BP neural network, threshold value
It is analyzed, then easily can fall into local solution to avoid itself, the defects of training speed is slow.The algorithm flow of modeling such as Fig. 3 institute
Show.
1. the characteristics of label and data itself for being generated according to data prediction, constructs the topology knot of BP neural network
Structure；
2. carrying out pre-training to BP neural network, and determine initial weight, threshold value；
3. genetic algorithm determines population quantity, the number of iterations etc. to initial weight, threshold process；
4. determining fitness function, and according to desired value error, calculate the fitness f (i) of individual i；
5. the selection operation of genetic computation calculates separately individual adaptation degree, process of aggregation.Calculate probability
6. the crossover operation of genetic computation is realized using arithmetic crossover method by crossover probability；
7. the mutation operation of genetic computation is realized using non-uniform mutation probability；
8. calculating the fitness of new population, constantly study is until meeting standard；
9. finally bringing optimal value into BP neural network as best initial weights, threshold value；
10. progress BP neural network training, forward direction, which calculates, is implied in output unit output, and calculating error reaches precision and then stops
It only trains, the connection weight and threshold value of reversed corrective networks structure interlayer, training setting maximum times.Training terminates, and obtains pre-
Survey result.
When that will be applied to the network optimization based on the BP neural network model of genetic algorithm, BP neural network can be related to network
The number of plies, neuron number, transmission function, the design of Learning Algorithms.
(1) the network number of plies
Unlimited mapping relations can be approached with arbitrary accuracy in view of 3 layers of BP neural network, therefore are used in practical applications
The mode of one hidden layer.
(2) hidden layer node number
During the network optimization, the selection of hidden layer neuron quantity is usually all true by way of implementing verifying
Recognize.According to previous Optimization Experience, the formula of following hidden layer and input, output layer neuron number is summarized
Y=log2i；
Wherein, y is hidden layer neuron number, and i, o represent input layer, output layer neuron number；R is constant according to different
Optimize scene to determine；K is sample number, and x is 0 to the constant between n.
(3) transmission function
Tansig function passes are chosen between input layer and hidden layer, select logsig function in hidden layer and output layer
Transmitting.
(4) training function
Trainlm Learning Algorithms are selected in terms of training function.
In systems in practice, test data, performance data, work parameter according to etc. multi-data sources data in daily morning conduct
" external information " is input to data sharing center, and Data Discretization is formed included judgement attribute by established rule label
Label is input to neural network input layer neuron, and is transferred to each neuron of hidden layer.Here " external information " is pre-
Exist in survey as impact factor.
In terms of weight and threshold optimization, initial population is selected according to practical problem.Fitness function is by sample
Data determine that learning error formula is as follows to the resulting learning error of population at individual training
Wherein, i indicates chromosome number, and m indicates that output node number, n indicate number of training.
The training requirement of BP neural network is that the MSE between reality output and desired output is minimum, then individual adaptation degree F (i)
For the inverse of E (i).
Initial population number, crossover probability and mutation probability determine that initial population number is bigger according to actual analysis problem, obtain
The probability of globally optimal solution is bigger, and the time is also longer, therefore is generally determined according to analyst coverage.It is then first if it is the whole province's grade analysis
Beginning population is districts and cities' number × n (n is [1,100])；If it is it is prefecture-level analysis then initial population be districts and cities' number × m (m for [1,
500])；Initial population is districts and cities' number × x if region class analysis (x is [1,2000]).The the frequency of crossover probability the big, receives
It is faster to hold back speed, it is contemplated that the general frequency of practical application request is determined by runing time.Mutation probability generally takes 0.0001~
0.1, the excessive possible triggering sample diversity of aberration rate, aberration rate is too small more difficult to find globally optimal solution.
The method that the present invention utilizes big data analysis is tested by analyzing magnanimity network optimization data as congested cell, power control
Two contents are made, by analysing in depth, input of 52 labels as BP neural network is devised, is used for training pattern, and pre-
Survey These parameters.
The above description is only a preferred embodiment of the present invention, is not intended to restrict the invention, for those skilled in the art
For, the invention can have various changes and changes.All any modifications made within the spirit and principles of the present invention are equal
Replacement, improvement etc., should all be included in the protection scope of the present invention.
Claims (7)
1. being directed to the big data analysis system of LTE network problem, which is characterized in that the system comprises: data active layer, acquisition
Layer, convergence layer and application layer；Wherein, the data active layer is practical by physical location is adjacent or non-conterminous several data source groups
At to provide multi-source heterogeneous data information needed for system；The acquisition layer is responsible for accurately and rapidly completing from the data
The acquisition tasks of active layer, and abnormal data is cleared up, it is the basis of big data analysis；The convergence layer is responsible for utilizing acquisition
Data provided by layer establish the data sharing center for being based on HDFS (distributed file system), and handle it；It is described
Application layer provides unified access and control and connects using the processing result that the convergence layer provides for user and extension application
Mouthful.
2. the big data analysis system according to claim 1 for LTE network problem, which is characterized in that the convergence
Layer uses the big data processing mode of standard, and all original and machined data are uniformly stored in HBase, and is completed to it
Data cleansing, data check formulate label, aggregation scene and associating policy processing.
3. the big data analysis system according to claim 1 for LTE network problem, which is characterized in that the application
Layer is made of customization standard interface module, system core function application module, process functional module and authority management module.
4. the big data analysis system according to claim 3 for LTE network problem, which is characterized in that the customization
Standard interface module is responsible for extension application access data sharing center.
5. the big data analysis system according to claim 3 for LTE network problem, which is characterized in that the system
Core function application module includes label, scene, strategy, parameter management and evaluation function.
6. the big data analysis system according to claim 3 for LTE network problem, which is characterized in that the process
Functional module is responsible for providing the production procedure based on permission and ensures legitimacy of the network analysis result in terms of automated execution.
7. the big data analysis system according to claim 3 for LTE network problem, which is characterized in that the permission
Management module is responsible for realizing stringent function control and user's control, lifting system safety.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711494416.3A CN110019100A (en) | 2017-12-31 | 2017-12-31 | For the big data analysis system of LTE network problem |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711494416.3A CN110019100A (en) | 2017-12-31 | 2017-12-31 | For the big data analysis system of LTE network problem |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110019100A true CN110019100A (en) | 2019-07-16 |
Family
ID=67187232
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711494416.3A Pending CN110019100A (en) | 2017-12-31 | 2017-12-31 | For the big data analysis system of LTE network problem |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110019100A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113191576A (en) * | 2020-12-31 | 2021-07-30 | 北京声迅电子股份有限公司 | Financial institution public security three-dimensional prevention and control system and construction method thereof |
| CN114443746A (en) * | 2022-01-21 | 2022-05-06 | 南京林科斯拉信息技术有限公司 | Expandable high-performance operation and maintenance index (ETL) tool and application method |
-
2017
- 2017-12-31 CN CN201711494416.3A patent/CN110019100A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113191576A (en) * | 2020-12-31 | 2021-07-30 | 北京声迅电子股份有限公司 | Financial institution public security three-dimensional prevention and control system and construction method thereof |
| CN114443746A (en) * | 2022-01-21 | 2022-05-06 | 南京林科斯拉信息技术有限公司 | Expandable high-performance operation and maintenance index (ETL) tool and application method |
| CN114443746B (en) * | 2022-01-21 | 2024-06-07 | 南京林科斯拉信息技术有限公司 | Expandable high-performance operation and maintenance index item ETL tool and application method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zeb et al. | Industrial digital twins at the nexus of NextG wireless networks and computational intelligence: A survey | |
| US10481629B2 (en) | Cognitive platform and method for energy management for enterprises | |
| WO2023246568A1 (en) | Digital twin architecture, model evaluation system, and model running method | |
| Danesh et al. | A systematic comparison of multi-criteria decision making methods for the improvement of project portfolio management in complex organisations | |
| Celik et al. | DDDAS-based multi-fidelity simulation framework for supply chain systems | |
| CN108924198A (en) | A kind of data dispatching method based on edge calculations, apparatus and system | |
| CN109088785A (en) | A kind of IDC data center apparatus management system | |
| Rahman et al. | Replica selection strategies in data grid | |
| Guim et al. | Autonomous lifecycle management for resource-efficient workload orchestration for green edge computing | |
| CN101299861A (en) | Base station system polling path automatization determination method based on shortest cycle | |
| JP2019075077A (en) | Workflow engine framework | |
| Okafor | Dynamic reliability modeling of cyber-physical edge computing network | |
| Wang et al. | Optimal deployment of cloudlets based on cost and latency in Internet of Things networks | |
| CN104363104A (en) | User-requirement-oriented mass multivariate data situation displaying system and method | |
| Li et al. | Service reliability modeling and evaluation of active-active cloud data center based on the IT infrastructure | |
| CN110019100A (en) | For the big data analysis system of LTE network problem | |
| CN109034403A (en) | A kind of urban rail transit equipment skeleton | |
| Rakkiannan et al. | An automated network slicing at edge with software defined networking and network function virtualization: a federated learning approach | |
| Ruiz et al. | Big data architecture for building energy management systems | |
| Almurshed et al. | Adaptive edge-cloud environments for rural AI | |
| CN105007176B (en) | A kind of cloud service QoS prediction technique based on layering Bayesian network model | |
| Zhang | Storage optimization algorithm design of cloud computing edge node based on artificial intelligence technology | |
| CN109636095A (en) | A kind of grid equipment Unified Model management system based on regulation cloud framework | |
| Garcia-Magarino et al. | Survivability strategies for emerging wireless networks with data mining techniques: A case study with NetLogo and RapidMiner | |
| Faraji-Mehmandar et al. | A self-learning approach for proactive resource and service provisioning in fog environment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190716 |