CN103888285A - Cognitive network resource intelligent management method - Google Patents
Cognitive network resource intelligent management method Download PDFInfo
- Publication number
- CN103888285A CN103888285A CN201310631099.0A CN201310631099A CN103888285A CN 103888285 A CN103888285 A CN 103888285A CN 201310631099 A CN201310631099 A CN 201310631099A CN 103888285 A CN103888285 A CN 103888285A
- Authority
- CN
- China
- Prior art keywords
- network
- module
- degree
- depth belief
- business
- 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
Images
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention relates to a cognitive network resource intelligent management method. Aiming at a problem of business-oriented end-to-end service quality assurance in a cognitive network, service quality parameter values of network elements and links which are expected to be used by the business are firstly acquired by the method. Resource availability assessment is performed by using a BP-depth belief network. A decision scheme is confirmed by using case inference according to the assessment result, and network resource reconfiguration is performed according to the decision scheme so that business end-to-end service quality in the cognitive network is guaranteed.
Description
Technical field
The present invention relates to Internet resources intelligent management field, relate in particular to the resource intelligent management method based on BP-degree of depth belief network in a kind of cognition network.
Background technology
In recent years, popularizing fast of the fast development of Internet network technology and multimedia application, makes Network and number of users be explosive increase, the important channel that network becomes people's obtaining information, releases news.In cognition network, Internet resources intelligent management can provide safeguard for stability and the reliability of the network operation.
Internet resources intelligent management utilizes intelligent algorithm to realize network resources automatic configuration, meets network multi-service service quality, makes the resource in network obtain more effectively utilizing when QoS demand.Originally management resource network management is mainly according to network feedback information labor management network, and network resource management efficiency is low, and lacks real-time, is one of bottleneck improving network service quality.Along with the appearance of autonomous management concept, intelligent network resource management is more and more subject to people's attention.When intelligence resource intelligent management method can be according to network implementation, qos parameter is realized Internet resources configuration and automatically adjusting automatically.Cognitive network resource intelligent management can be according to the variation of customer service, considers current network resources condition, and dynamic recognition Internet resources, takes into account the QoS of business when having improved the real-time of network resource management and efficiency.
Summary of the invention
Technical problem solved by the invention is:
Improve network resource management efficiency low, and lack the problem of real-time, realize the awareness of network, improve efficiency and the service quality of network resource management.
In order to solve the problems of the technologies described above, the technical solution adopted in the present invention is:
Adopt a kind of resource intelligent management method of cognition network, mainly by secure communication interface module, stock assessment module, policy management module, database module and network router are realized:
Wherein, secure communication interface module is responsible for communicating by letter of fulfillment database, policy management module, stock assessment module and network router;
Stock assessment module is responsible for according to the availability of the qos parameter value assessment current network of the network element of business expection use and link, and wherein qos parameter can comprise bandwidth, time delay, shake, packet loss.This module is one of core of resource intelligent administrative center, and BP-degree of depth belief network can be realized the algorithm of this functions of modules.
Policy management module is responsible for the reasoning according to stock assessment result use case and is carried out decision-making, and reshuffles according to result of decision control Internet resources, one of core of this module Yi Shi resource intelligent administrative center.
Database module is responsible for Internet resources parameter and the case of storage of collected, is stock assessment module and decision-making management module service.
BP-degree of depth belief network algorithm steps is as follows:
1, adopt four layers of BP-degree of depth belief network architecture, a degree of depth belief network of the three layers of formation in bottom, top layer is output layer.Be input as the qos parameter value of network element and link, output represents current network resources availability;
2, ignore output layer, three layer depth belief networks form a reverse generation model, by the greediness based on restriction Boltzmann machine RBM (Restricted Boltzmann Machine) successively unsupervised learning algorithm complete degree of depth belief network and train in advance.Training process briefly in two steps in advance: (1) obtains the initial parameter collection of degree of depth trust network (DBN) top layer first order RBM by RBM training algorithm; (2) output of first order RBM is as the input of second level RBM, same initiation parameter collection;
3, obtain the parameter set of degree of depth belief network by pre-training, then four layers of BP-degree of depth belief network are carried out to accurate adjustment as four layers of common neural net.
The invention has the beneficial effects as follows:
When intelligence resource intelligent management method can be according to network implementation, qos parameter is realized Internet resources configuration and automatically adjusting automatically.Cognitive network resource intelligent management can be according to the variation of customer service, considers current network resources condition, and dynamic recognition Internet resources, takes into account the QoS of business when having improved the real-time of network resource management and efficiency
Fig. 1 is the resource intelligent administrative model structural representation based on degree of depth belief network of the embodiment of the present invention.
Fig. 2 is the resource intelligent administrative center design composition structural representation that the present invention is based on a kind of cognition network.
Fig. 3 is the flow chart that the present invention is based on a kind of resource intelligent management method of cognition network.
Embodiment
Below in conjunction with Figure of description, the present invention is further illustrated.
As shown in Figure 1, cognitive territory party A-subscriber is to cognitive territory B request resource, and in territory, the cognitive center A of business obtains the information such as business, request object, to the Resource Availability assessment of resource intelligent administrative center request service-oriented.If Internet resources can meet QoS of survice demand, return to feasible instruction to the cognitive center A of business in territory,, user is notified at business cognition center in the territory in.If Internet resources can not meet QoS of survice demand, return etc. to be instructedly to cognitive center A in territory, according to strategy, instruction is carried out Internet resources and is reshuffled simultaneously, such as heavy-route, automatic queue management etc.Reshuffle the cognitive center A of business in the backward territory of end and return to feasible instruction, in territory, user is notified at the cognitive center of business.Wherein resource intelligent administrative center as shown in Figure 2, adopts the resource intelligent management method based on BP-degree of depth belief network to realize.
As shown in Figure 2 and Figure 3, secure communication interface module is responsible for communicating by letter of fulfillment database, policy management module, stock assessment module and network router; Wherein stock assessment module is responsible for according to the availability of the qos parameter value assessment current network of the network element of business expection use and link, wherein qos parameter can comprise bandwidth, time delay, shake, packet loss, the algorithm of realizing this module estimation network resource availability is BP-degree of depth belief network, adopt four layers of BP-degree of depth belief network architecture, a degree of depth belief network of the three layers of formation in bottom, top layer is output layer.Be input as the qos parameter value of network element and link, output represents current network resources availability; Ignore output layer, three layer depth belief networks form a reverse generation model, by the greediness based on RBM successively unsupervised learning algorithm complete degree of depth belief network and train in advance.Training process briefly in two steps in advance: (1) obtains the initial parameter collection of DBN top layer first order RBM by RBM training algorithm; (2) output of first order RBM is as the input of second level RBM, and same initiation parameter collection finally obtains the parameter set of degree of depth belief network by pre-training, then four layers of BP-degree of depth belief network are carried out to accurate adjustment as four layers of common neural net.Policy management module is carried out decision-making according to the output valve use case reasoning of stock assessment module, and reshuffles according to result of decision control Internet resources.
More than show and described general principle of the present invention and principal character and advantage of the present invention.The technical staff of the industry should understand; the present invention is not restricted to the described embodiments; that in above-described embodiment and specification, describes just illustrates principle of the present invention; without departing from the spirit and scope of the present invention; the present invention also has various changes and modifications, and these changes and improvements all fall in the claimed scope of the invention.The claimed scope of the present invention is defined by appending claims and equivalent thereof.
Claims (3)
1. a cognitive network resource intelligent management, it is characterized in that: comprise secure communication interface module, stock assessment module, policy management module, database module and isdn router, wherein secure communication interface module respectively with stock assessment module, policy management module, database module, isdn router module is connected, be responsible for realizing stock assessment module, policy management module, the communication of database and network router, stock assessment module is responsible for according to the availability of the QoS parameter value assessment current network resources of the network element of business expection use and link, policy management module is responsible for the reasoning according to stock assessment result use case and is carried out decision-making, and carry out Internet resources according to the result of decision and reshuffle, database module is responsible for Internet resources parameter and the case of storage of collected, for stock assessment module and decision-making management module service.
2. according to a kind of cognitive network resource intelligent management described in claim l, it is characterized in that: described stock assessment module is assessed in the process of current network resources availability at network element and link quality of service parameter value that according to business, expection is used, network element and link quality of service parameter value that the business expection of adopting is used are bandwidth, time delay, shake and packet loss.
3. a kind of cognitive network resource intelligent management according to claim 1, it is characterized in that: described stock assessment module is assessed in the process of current network resources availability at network element and link quality of service parameter value that according to business, expection is used, adopt BP-degree of depth belief network to realize the algorithm of this functions of modules, this algorithm steps is:
Adopt four layers of BP-degree of depth belief network architecture, a degree of depth belief network of the three layers of formation in bottom, top layer is output layer, is input as the QoS parameter value of network element and link, output represents current network resources availability;
Ignore output layer, three layer depth belief networks form a reverse generation model, by the greediness based on restriction Boltzmann machine successively unsupervised learning algorithm complete degree of depth belief network and train in advance.Pre-training process briefly in two steps, first obtain the initial parameter collection of degree of depth trust network top first order restriction Boltzmann machine by restriction Boltzmann machine training algorithm, then, the output of first order restriction Boltzmann machine is as the input of second level restriction Boltzmann machine, same initiation parameter collection;
Obtain by pre-training after the parameter set of degree of depth belief network, four layers of BP-degree of depth belief network are carried out to accurate adjustment as four layers of common neural net.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310631099.0A CN103888285A (en) | 2013-12-02 | 2013-12-02 | Cognitive network resource intelligent management method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310631099.0A CN103888285A (en) | 2013-12-02 | 2013-12-02 | Cognitive network resource intelligent management method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN103888285A true CN103888285A (en) | 2014-06-25 |
Family
ID=50957012
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310631099.0A Pending CN103888285A (en) | 2013-12-02 | 2013-12-02 | Cognitive network resource intelligent management method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103888285A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2018176768A1 (en) * | 2017-03-27 | 2018-10-04 | 烽火通信科技股份有限公司 | Network architecture of humanoid network and implementation method |
WO2020073870A1 (en) * | 2018-10-12 | 2020-04-16 | 中兴通讯股份有限公司 | Mobile network self-optimization method, system, terminal and computer readable storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070079355A1 (en) * | 2005-09-30 | 2007-04-05 | Rockwell Automation Technologies, Inc. | Data perspectives in controller system and production management systems |
CN102025539A (en) * | 2010-12-07 | 2011-04-20 | 南京邮电大学 | Agent-based available resource assessment method for cognitive network |
CN102075944A (en) * | 2010-12-07 | 2011-05-25 | 南京邮电大学 | Multi-Agent cooperation-based available resource management system and method in cognitive network |
-
2013
- 2013-12-02 CN CN201310631099.0A patent/CN103888285A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070079355A1 (en) * | 2005-09-30 | 2007-04-05 | Rockwell Automation Technologies, Inc. | Data perspectives in controller system and production management systems |
CN102025539A (en) * | 2010-12-07 | 2011-04-20 | 南京邮电大学 | Agent-based available resource assessment method for cognitive network |
CN102075944A (en) * | 2010-12-07 | 2011-05-25 | 南京邮电大学 | Multi-Agent cooperation-based available resource management system and method in cognitive network |
Non-Patent Citations (1)
Title |
---|
李雷: "《认知无线网络自适应信任管理机制研究》", 《万方数据企业知识服务平台》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2018176768A1 (en) * | 2017-03-27 | 2018-10-04 | 烽火通信科技股份有限公司 | Network architecture of humanoid network and implementation method |
WO2020073870A1 (en) * | 2018-10-12 | 2020-04-16 | 中兴通讯股份有限公司 | Mobile network self-optimization method, system, terminal and computer readable storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107871164B (en) | Fog computing environment personalized deep learning method | |
Huang et al. | Software-defined QoS provisioning for fog computing advanced wireless sensor networks | |
CN104299053A (en) | Method for selecting optimal path of power distribution and utilization communication network based on genetic algorithm | |
CN110516065A (en) | A method of the attention neural network based on multichannel dynamic mask | |
CN101951324A (en) | Be used to use the system and method for the self-organizing network of dynamic strategy and situation semantics | |
CN111010341B (en) | Overlay network routing decision method based on deep learning | |
CN103326946A (en) | SVC streaming media transmission optimization method based on OpenFlow | |
CN101552934A (en) | Optical internet cross-domain reliable route calculating method based on PCE backtracking recursion | |
CN101616074B (en) | Multicast routing optimization method based on quantum evolution | |
CN110351188A (en) | A kind of message forwarding method, data processing method, device and network system | |
CN116579417A (en) | Layered personalized federal learning method, device and medium in edge computing network | |
CN103824127B (en) | Service self-adaptive combinatorial optimization method under cloud computing environment | |
Xia et al. | Generative AI for semantic communication: Architecture, challenges, and outlook | |
CN115580582A (en) | Network flow centralized control system applied to computing power network | |
CN103888285A (en) | Cognitive network resource intelligent management method | |
Yuan et al. | Graphcomm: Efficient graph convolutional communication for multiagent cooperation | |
CN102868757A (en) | Dynamic Web service combination method based on quality of service (QoC) indexes | |
CN103581329A (en) | Construction method for topological structure based on clustered peer-to-peer network streaming media direct broadcast system | |
Zervas et al. | Service-oriented multigranular optical network architecture for clouds | |
CN103577899A (en) | Service composition method based on reliability prediction combined with QoS | |
Luo et al. | Network selection algorithm based on group decision making for heterogeneous wireless networks | |
CN109450587A (en) | Spectral integration processing method, device and storage medium | |
CN104732278A (en) | Deep neural network training method based on sea-cloud collaboration framework | |
CN104009943B (en) | LLDP message transmitting methods and DCB equipment | |
CN102710431A (en) | Method for automatically designing topology of business-driven transmission network |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20140625 |
|
WD01 | Invention patent application deemed withdrawn after publication |