CN101166342B - Independent radio resource management method, system and device under multi-operator scenario - Google Patents

Independent radio resource management method, system and device under multi-operator scenario Download PDF

Info

Publication number
CN101166342B
CN101166342B CN2007101754970A CN200710175497A CN101166342B CN 101166342 B CN101166342 B CN 101166342B CN 2007101754970 A CN2007101754970 A CN 2007101754970A CN 200710175497 A CN200710175497 A CN 200710175497A CN 101166342 B CN101166342 B CN 101166342B
Authority
CN
China
Prior art keywords
network
information
radio resource
resource management
management method
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.)
Expired - Fee Related
Application number
CN2007101754970A
Other languages
Chinese (zh)
Other versions
CN101166342A (en
Inventor
冯志勇
张平
张永靖
曾宪
薛圆
苗丹
张奇勋
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing University of Posts and Telecommunications
Original Assignee
Beijing University of Posts and Telecommunications
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing University of Posts and Telecommunications filed Critical Beijing University of Posts and Telecommunications
Priority to CN2007101754970A priority Critical patent/CN101166342B/en
Publication of CN101166342A publication Critical patent/CN101166342A/en
Application granted granted Critical
Publication of CN101166342B publication Critical patent/CN101166342B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Abstract

After terminal of user in overlapping covered area of networks of multiple operation managers sends out session request, networks of operation managers analyze current network states as well as calculate yield situations of each network in the overlapping covered area brought by different radio resource management methods under current network states. Optimal radio resource management method is selected based on synthetical profit level of each operation manager. The invention solves competitive management with reasonableness among multiple operation managers, and meets requirement of autonomy from system.

Description

Independent radio resource management method, system and equipment under the multi-operator scenario
Technical field
The present invention relates to the mobile wireless network technical field, be applied in independent radio resource management method under the multi-operator scenario and system in the heterogeneous network with a kind of more precisely
Background technology
The coexistence of heterogeneous wireless network (comprising universal mobile telecommunications system (UMTS), wireless local access network (WLAN) etc.) is the main feature of following B3G environment.Various wireless networks have overlapping coverage, multiple business demand and complementary technical characteristic, in order to obtain higher systematic function and abundanter user experience, must carry out mutual and collaborative between the wireless network.In recent years, progressively developed based on the end-to-end reconfiguration technology of software and radio technique, more helping RRM provides dynamic network selecting ability, is fitted to the ability of required wireless network for terminal and network.
Although present radio resource management method can successfully be finished the work, do not embody the management independence.Yet the increase along with the various technology and equipments that satisfy user and operator's demand makes system complexity constantly increase, thereby independence just becomes more and more important.Consider that an operator runs several access networks with different radio access technologies in a urban area, jointly controlling between the wireless access technology must be from different traffic flow requirements management and suitable, do not have the planning and the maintenance cost of many human interventions.
And under the scene of multi-operator, operator not only will consider problems such as the interests, network quality of self, also will weigh the relevant issues of other operators.Often exist a kind of relation of competition between the operator, can not simply solve with the method for handling a plurality of networks under the operator.There is not at present relevant method to solve this problem yet.
Summary of the invention
In view of above problem, the present invention proposes the free associating wireless resource management method, system and the equipment that are used under the multi-operator scenario, properly solve the competition management between the multi-operator, and further satisfied the requirement of system independence.
Independent radio resource management method based under above-mentioned purpose a kind of multi-operator scenario provided by the invention comprises:
After the user terminal that is in the network overlapped overlay area of multi-operator sends conversation request, the network analysis current network state of operator;
Calculating is adopted the situation of Profit of each network of the different overlapping coverings that radio resource management method brought under the current network situation;
Comprehensive profit level according to each operator is selected optimum radio resource management method.
The described analysis current network state of this method comprises: collect the characteristic and the network load information of this conversation request, and be translated into the usable network information of expression current network state.
Characteristic and load information that the described collection of this method arrives session mainly comprise: position, the classification of service and the concrete network parameter of location network of requiring;
Described conversion process transforms out usable network information for the characteristic and the load information that will arrive session by data-measuring and the corresponding means of parameter.
The described available network information of this method comprises: this user's required service is actual to require the actual loading percentage of parameters such as bandwidth that network provides, service quality and location network.
The described calculating situation of Profit of this method comprises: the network state information that obtains is undertaken extensive by neural net, calculate under the current network situation, adopt the situation of Profit of different radio resource management method brought self network.
The optimum radio resource management method process of the described selection of this method adopts the autonomous learning method.
The described autonomous learning method of this method is the Q learning algorithm.
Each operator of this method safeguards the Q value table of oneself separately, share the Q value information between them, the Q value of each operator table is formed a matrix, and then the relevant Q linear programming of employing, obtain a comprehensive evaluation criteria, each operator implements according to this Standard Selection appropriate wireless resources management method.
The optimum radio resource management method of the described selection of this method comprises: the pairing income of different behaviors under the network condition of the different operators that produces is carried out game, obtain a comprehensive criterion;
According to the synthetic determination standard that draws, select a kind of radio resource management method to preserve selection result, and this method is transferred to user terminal.
This method is described obtain a comprehensive criterion after, calculate the state of each network again, be used for upgrading the financial value in the history
This method is described after distributing the appropriate wireless resources management method for user terminal, also comprise: obtain the income of present networks, and by with communicating by letter of other carrier networks obtain overlapping covering other networks income and selected radio resource management method and preserve.
Also comprise before the described analysis current network state of this method: set initial parameter;
Finish at every turn and also comprise after optimum radio resource management method is selected: the parameter to each module is carried out convergence process.
The described initiation parameter of this method comprises: the length of the preference of neural net and training formation, the discount factor of autonomous learning part, initial exploration probability, initial learn rate and the initial ranging probability in the method selection algorithm.
After the described convergence process of this method, also comprise: judge according to the parameter after upgrading whether study has arrived a stable state; If then advance to stop the process that current iteration is learnt; Otherwise proceed the iterative learning process of next round.
The process of the described judgement stable state of this method comprises: when learning rate is reduced to 0 or predefined thresholding, judge that this process reaches stable state, otherwise, think not reach stable state as yet.
Based on above-mentioned purpose, the present invention also provides the management system of the independent radio resource under a kind of multi-operator scenario, include wireless network control apparatus, also comprise: be connected to the joint radio resource management unit on the wireless network control apparatus, be used for after the user terminal of the network overlapped overlay area that is in multi-operator sends conversation request, analyzing current network state; Calculating is adopted the situation of Profit of each network of the different overlapping coverings that radio resource management method brought under the current network situation; Comprehensive profit level according to each operator is selected optimum radio resource management method.
The wireless network control apparatus of this described operator of system, be used for providing the traffic type information of network state and user's request, and the radio resource management method that described joint radio resource management unit is produced is handed down to described user terminal to described joint radio resource management unit.
This system described joint radio resource management unit produces radio resource management method by the process of constantly study, and sends to described user terminal.
Inside, this system described joint radio resource management unit comprises:
Perception and information translation module are used for to the perception of wireless environment with to the conversion of information;
The autonomous learning module, be connected with the information translation module with perception, be used to put in order data that perception and information translation module transmit as present network state, according to calculating the income level that gets, and carry out the extensive and Knowledge Storage of corresponding data by the extensive and Knowledge Storage module of data, further take all factors into consideration the avail information of other operators and select a kind of appropriate wireless resources management method, send to described wireless network control apparatus, send described user terminal then to;
Extensive and the knowledge store module of data is connected with the autonomous learning module, and it is extensive to use neural net to do data, is used to reduce required memory space, and historical experience is in the past stored.
Described perception of this system and information translation module, also having described perceptional function to wireless environment is the specifying information that obtains in the network environment.
Specifying information in the described network environment of this system comprises: position, the classification of service and the concrete network parameter of location network of requiring.
Described perception of this system and information translation module comprise that also the information that information conversion function is specially collecting quantizes, and is converted to operable digital information.
In the described autonomous learning module of this system, the described income of taking all factors into consideration other operators adopts the game mode, is met the radio resource management method of multi-operator competitive relation from game.
Also provide a kind of joint radio resource management equipment at last based on above-mentioned purpose the present invention, having comprised:
Perception and information translation module are used for to the perception of wireless environment with to the conversion of information;
The autonomous learning module, be connected with the information translation module with perception, be used to put in order data that perception and information translation module transmit as present network state, according to calculating the income level that gets, and carry out the extensive and Knowledge Storage of corresponding data by the extensive and Knowledge Storage module of data, further take all factors into consideration the avail information of other operators and select a kind of appropriate wireless resources management method, send to described wireless network control apparatus, send described user terminal then to;
Extensive and the knowledge store module of data is connected with the autonomous learning module, and it is extensive to use neural net to do data, is used to reduce required memory space, and historical experience is in the past stored.
Described perception of this equipment and information translation module, also having described perceptional function to wireless environment is the specifying information that obtains in the network environment.
Specifying information in the described network environment of this equipment comprises: the relevant information of network quality, traffic carrying capacity, type of service and other carrier networks.
Described perception of this equipment and information translation module comprise that also the information that information conversion function is specially collecting quantizes, and is converted to operable digital information.
In the described autonomous learning module of this equipment, the described income of taking all factors into consideration other operators adopts the game mode, is met the radio resource management method of multi-operator competitive relation from game.
From above as can be seen, independent radio resource management method under the multi-operator scenario provided by the present invention, realized the autonomous learning characteristic of the joint radio resource management in the heterogeneous wireless network, can constantly revise its control strategy according to practical operation situation, with the planning of minimizing manpower participation and the cost of maintenance.Considered the competitive relation between the operator under the multi-operator scenario, the change centralized management is distributed management, and has adopted the method for game to solve balance of interest problem between the operator, reduces network blocking probability effectively and improves the income of operator.
Description of drawings
Fig. 1 is the schematic network structure that the embodiment of the invention adopted;
Fig. 2 is the joint radio resource management unit internal structure schematic diagram of the embodiment of the invention;
Fig. 3 is the schematic flow sheet of embodiment of the invention radio resource management method.
Embodiment
Be to solve problems of the prior art, major programme of the present invention comprises: after the user terminal that is in the network overlapped overlay area of multi-operator sends conversation request, and the network analysis current network state of operator; Calculating is adopted the situation of Profit of each network of the different overlapping coverings that radio resource management method brought under the current network situation; Comprehensive profit level according to each operator is selected optimum radio resource management method.
With reference to the accompanying drawings the present invention is described more fully, exemplary embodiment of the present invention wherein is described.
The network architecture figure at the inventive method place is mainly concerned with multimode terminal 11, this carrier network 12 and other carrier networks 13 as shown in Figure 1.Wherein, this carrier network 12 relates generally to this carrier network control appliance 121 and this operator joint radio resource management control unit 120; Other carrier networks 13 relate generally to other carrier network control appliance 131 and other operator's joint radio resource management control units 130.
Suppose that several operators run several heterogeneous wireless networks respectively in a densely populated zone, and the region overlapping that covers.
The user carries multimode terminal 11 and is in overlapping coveredly, by buying the business of any one Radio Access Network, enjoys various wireless applications.During multimode terminal 11 access networks, can finish software automatically according to the radio resource management method that wireless network control apparatus issues and install and network settings, not need the user manually to carry out.
In whole network system shown in Figure 1, the joint radio resource management unit 120 (or be connected to other operator's joint radio resource management control units 130 of other carrier network control appliances 131) that is connected to wireless network management equipment 121 is the core cell that is mainly used to finish this patent.Its internal structure mainly comprises perception and information translation module 210, autonomous learning module 211 and data are extensive and the knowledge store module as shown in Figure 2.Cooperatively interact between them, and rely on from the joint radio resource management unit 23 of wireless network control apparatus 22 and other carrier networks and obtain information needed, finish the independent radio resource under the multi-operator scenario is managed.
Wherein, perception and information translation module 210 link to each other with joint radio resource management unit 23, the autonomous learning module 211 of Local wireless network control appliance 22, other operators respectively.Mainly comprise perception part and information translation part to wireless environment.The perception part can obtain the specifying information (generally comprising: the characteristic and the load information that arrive session) in the network environment, for example relevant information of network quality, traffic carrying capacity, type of service and other carrier networks.The information quantization that information translation is partly then partly collected perception is converted to operable digital information.This module 210 is unique modules that exchange information from the external world in the whole unit, and the needed information in every other module operating room all needs it to obtain from network.
Autonomous learning module 211, respectively with joint radio resource management unit 23, perception and the information translation module 210 of Local wireless network control appliance 22, other operators, data are extensive links to each other with knowledge store module 212.It is for the scene that solves multi-operator in the RRM and autonomous problem and a custom-designed module.The data preparation that it at first transmits perception and information translation module 210 adopts some to represent present network state.Calculate the income of under current state, using certain radio resource management method then.Income level is input to the extensive and Knowledge Storage module 212 of data carries out the extensive and Knowledge Storage of corresponding data.This autonomous learning module 211 is selected a kind of appropriate wireless resources management method according to last avail information at last, sends to wireless network management equipment 22, sends user terminal 11 then to.The purpose of whole joint radio resource management unit 120 is exactly to select certain network of certain operator to come for the user provides required service, so the main target of autonomous learning module 211 also is to concentrate on by study constantly to optimize this choice mechanism.Optimizing main is Actual Return according to network after selecting each time according to condition, after recognizing the actual gain of all different operators networks each time, just can select a kind of method that is more suitable for to carry out Radio Resource resource allocation under the multi-operator scenario according to the process that situation of Profit is carried out a game.This autonomous learning module 211 can be finished whole RRM process with last decision information notice terminal at last.
Because the continuity of the actual avail information of network, inevitable requirement has very big memory space.It is extensive to use neural net to do data in the extensive and Knowledge Storage module 212 of data, reduces required memory space.And, because learning process needs the accumulation of a large amount of in the past experiences, be necessary to design separately the purpose that a module is finished knowledge store.Mainly used neural network method here, because the characteristic of neural net is easy to just can finish the storage of required knowledge.
On the basis that above several modules cooperatively interact, concrete method is independent execution the in the distributed associating radio resource managing unit.The detailed operation flow process is the process of a continuous iteration, and is as shown in Figure 3, specific as follows:
Step 31 parameter initialization: need be some initial parameters of each several part module settings when workflow begins.
Specifically comprise: in the extensive and Knowledge Storage module 212 of data, the parameter that neural net need be provided with comprises the preference of neural net and the length of training formation etc.; In autonomous learning module 211, need to be provided with discount factor (parameter of the autonomous learning of discount factor service in fact in other words) in the autonomous learning, the initial exploration probability in the method selection algorithm, initial learn rate and initial ranging probability etc.
Step 32 information gathering and conversion: when a new conversation request arrives, collect the specifying information in the network environment that arrives session, and be translated into usable network information and give study module, represent current network state.
Wherein, the specifying information in the described network environment comprises characteristic and the load information that arrives session, and the characteristic and the load information that arrive session mainly comprise: position, the classification of service and the concrete network parameter of location network of requiring.Transform out the available network information by means such as data-measuring and parameter correspondences then.These available network informations mainly refer to actual actual loading percentage that requires parameters such as bandwidth that network provides, service quality and location network of this user's required service etc.
The acquisition of step 33 network profit: the usable network information that step 32 is obtained is undertaken extensive by neural net, and calculates under the current network situation, adopts the situation of Profit of different each networks of radio resource management method.
Neural net itself possesses the function approximation function, utilizes this ability, the back-propagation algorithm that minimizes mean square error is attached in the iterative process of autonomous learning, can handle the challenge with continuous or huge state/motion space effectively.In fact be exactly to make in learning process, can be according to known limited, discrete state/action experience, by the nonlinear fitting of neural net, to other the unknown (experience) but approximate evaluation is carried out in the evaluation of the state/action of " similar ", thereby realize extensive function.
The algorithm of supposing the autonomous learning of employing is the Q learning algorithm, and the Q value in the Q learning algorithm can be used for representing the income of each network.Neural net is by the above-mentioned functions memory and express the Q value.Can use the back-propagation algorithm that minimizes mean square error to adjust the Q value.With the usable network information input neural network that step 32 obtained, the output of neural net is promptly corresponding to the Q value of each radio resource management method.
The game of step 34 multi-operator: the pairing income of different behaviors is taken all factors into consideration under the particular network situation that step 33 produces, and carries out game, obtains a comprehensive criterion, calculates the state of each network again, is used for upgrading the financial value in the history.
Continuation is an example with the Q learning algorithm: each operator safeguards the Q value table of oneself separately.Suppose to share between them the Q value information.Can form a matrix to the Q value of each operator table, and then adopt relevant Q linear programming to solve problem of game between this matrix, obtain a comprehensive evaluation criteria.So just be equivalent to aspect the market segmentation, can reach certain agreement between the operator, thereby can make full use of the advantage of heterogeneous wireless access technology, reach the achievement of doulbe-sides' victory.
Step 35 method decision: according to the comprehensive criterion that step 34 provides, select a kind of radio resource management method, and the identification information of this method is transferred to the entrained multimode terminal of user, and be recorded for use in later learning process.
Continuation is an example with the Q learning algorithm: based on the Q value vector that obtains, and the radio resource management method that autonomous learning module 211 adopts the method decision of similar ξ greedy algorithm to take.Carry out back feedback enhanced signal according to this method, the Q value obtains upgrading, and the weights of neural net are that training set is adjusted with the back-propagation algorithm that minimizes mean square error with the Q value of upgrading then.For improving algorithm stability, the Q value of renewal can be cached to earlier in the training formation, gives neural net with batch fashion again.
Step 36 financial value upgrades: after distributing the appropriate wireless resources management method for user terminal, the joint radio resource management unit can obtain the income of present networks this time; And by obtaining the income of other networks of overlapping covering and the radio resource management method of selection, the result is put into the extensive and knowledge store module of data preserve, with convenient later matrix games process with communicating by letter of other joint radio resource management modules.
(began to carry out) when next time, new service request arrived from step 32, just can obtain next state, new state value just obtains by step 34, and current income also obtains, so just upgraded network profit, and the result has been put into the extensive and knowledge store module preservation of data.
It is to be noted that the acquisition of Q value of step 33 refers to the information of directly obtaining current time deposited out from Q value table, this information can help to carry out Action Selection once, and step 36 is to calculate the income of this action according to action this time, upgrades the pairing Q value in the Q value table.
Continuation is an example with the Q learning algorithm: the computing formula of income can be expressed as Q t + 1 ( s , a ) = ( 1 - α ) Q t ( s , a ) + α ( r t + γ max a ′ Q t ( s ′ , a ′ ) ) ;
Wherein, Q (s, a) stored before being in corresponding state that step 33 obtains and the Q value under the action, α is a learning rate, γ is the discount factor in the autonomous learning, its size is just represented the attention degree to long-range income,
Figure S2007101754970D00092
The NextState issuable maximum Q value of action down, the financial value of this that r is exactly are adopted after this method in expression.Briefly, exactly should be according to this income and the Q value of the NextState Q value of removing to upgrade this state, take into account the Q value of not carrying out before this action again.
Step 37 parameter update: after the e-learning and judging process finished once, change the parameter of each several part module, carry out convergence process, make learning process have convergence, carry out later study and judging process again such as parameter to each module.
Learning rate in the autonomous learning module 211 and exploration probability all need to upgrade.For example they can be according to the negative exponent rule along with the process of learning be reduced to 0 gradually, to satisfy the convergence requirement of study.
Above step 32 will constantly be carried out iteration to the process of step 37 and be carried out, and will carry out step 38 when each iterative process of taking turns is finished.
Step 38 autonomous learning module judges according to the parameter after upgrading whether study has reached a stable state, if stable, then enters step 39, stops the process of such iterative learning.That is to say that the scheme when later radio resource management method stops according to learning process is carried out.If not, then to proceed the iterative learning process of next round.
Wherein, describedly judge whether to reach stable state, can adopt following method: when learning rate is reduced to 0 or one predefined thresholding, judge that this process reaches stable state, otherwise, think not reach stable state as yet, return step 302, continue next conversation request is carried out the processing of step 32 to 38.
Step 39 stops the current iteration learning process.
Description of the invention provides for example with for the purpose of describing, and is not exhaustively or limit the invention to disclosed form.Many modifications and variations are obvious for the ordinary skill in the art.Selecting and describing embodiment is for better explanation principle of the present invention and practical application, thereby and makes those of ordinary skill in the art can understand the various embodiment that have various modifications that the present invention's design is suitable for special-purpose.

Claims (26)

1. the independent radio resource management method under the multi-operator scenario is characterized in that, comprising:
After the user terminal that is in the network overlapped overlay area of multi-operator sends conversation request, the network analysis current network state of operator;
The network state information that obtains is undertaken extensive by neural net, calculate under the current network situation, adopt the situation of Profit of each network of the different overlapping coverings that radio resource management method brought;
According to the comprehensive profit level of each operator, adopt the autonomous learning method to select optimum radio resource management method.
2. method according to claim 1 is characterized in that, described analysis current network state comprises: collect the characteristic and the network load information of this conversation request, and be translated into the usable network information of expression current network state.
3. method according to claim 2 is characterized in that, the characteristic and the load information of the conversation request of described collection mainly comprise: position, the classification of service and the concrete network parameter of location network of requiring;
Described conversion process transforms out usable network information for the characteristic and the load information that will arrive session by data-measuring and the corresponding means of parameter.
4. method according to claim 3 is characterized in that, the described available network information comprises: this user's required service is actual to require the actual loading percentage of bandwidth, service quality and location network that network provides.
5. method according to claim 1 is characterized in that, described autonomous learning method is the Q learning algorithm.
6. method according to claim 5, it is characterized in that, each operator safeguards the Q value table of oneself separately, share the Q value information between them, the Q value of each operator table is formed a matrix, and then adopt relevant Q linear programming, and obtaining a comprehensive evaluation criteria, each operator implements according to this Standard Selection appropriate wireless resources management method.
7. method according to claim 1 is characterized in that, the optimum radio resource management method of described selection comprises: the pairing income of different behaviors under the network condition of the different operators that produces is carried out game, obtain a comprehensive criterion;
According to the synthetic determination standard that draws, select a kind of radio resource management method to preserve selection result, and this method is transferred to user terminal.
8. method according to claim 7 is characterized in that, described obtain a comprehensive criterion after, calculate the state of each network again, be used for upgrading the financial value in the history.
9. method according to claim 7, it is characterized in that, a kind of radio resource management method of described selection is preserved selection result, and this method is transferred to after the user terminal, also comprise: obtain the income of present networks, and by with communicating by letter of other carrier networks obtain overlapping covering other networks income and selected radio resource management method and preserve.
10. method according to claim 9 is characterized in that, also comprises before the described analysis current network state: set initial parameter;
Finish at every turn and also comprise after optimum radio resource management method is selected: the parameter to each module is carried out convergence process.
11. method according to claim 10, it is characterized in that described initial parameter comprises: the length of the preference of neural net and training formation, the discount factor of autonomous learning part, initial exploration probability, initial learn rate and the initial ranging probability in the method selection algorithm.
12. method according to claim 10 is characterized in that, after the described convergence process, also comprises: judge according to the parameter after upgrading whether study has arrived a stable state; If then advance to stop the process that current iteration is learnt; Otherwise proceed the iterative learning process of next round.
13. method according to claim 12 is characterized in that, the process of described judgement stable state comprises: when learning rate is reduced to or during a predefined thresholding, judges that this process reaches stable state, otherwise, think not reach stable state as yet.
14. the independent radio resource management system under the multi-operator scenario, include wireless network control apparatus, it is characterized in that, also comprise: be connected to the joint radio resource management unit on the wireless network control apparatus, be used for after the user terminal of the network overlapped overlay area that is in multi-operator sends conversation request, analyzing current network state; The network state information that obtains is undertaken extensive by neural net, calculate under the current network situation, adopt the situation of Profit of each network of the different overlapping coverings that radio resource management method brought; According to the comprehensive profit level of each operator, adopt the autonomous learning method to select optimum radio resource management method.
15. system according to claim 14, it is characterized in that, the wireless network control apparatus of described operator, be used for providing the traffic type information of network state and user's request, and the radio resource management method that described joint radio resource management unit is produced is handed down to described user terminal to described joint radio resource management unit.
16. system according to claim 14 is characterized in that, described joint radio resource management unit produces radio resource management method by the process of constantly study, and sends to described user terminal.
17. system according to claim 14 is characterized in that, inside, described joint radio resource management unit comprises:
Perception and information translation module are used for to the perception of wireless environment with to the conversion of information;
The autonomous learning module, be connected with the information translation module with perception, be used to put in order data that perception and information translation module transmit as present network state, according to calculating the income level that gets, and carry out the extensive and Knowledge Storage of corresponding data by the extensive and Knowledge Storage module of data, further take all factors into consideration the avail information of other operators and select a kind of appropriate wireless resources management method, send to described wireless network control apparatus, send described user terminal then to;
Extensive and the knowledge store module of data is connected with the autonomous learning module, and it is extensive to use neural net to do data, is used to reduce required memory space, and historical experience is in the past stored.
18. system according to claim 17 is characterized in that, described perception and information translation module are specially the specifying information that obtains in the network environment to described perceptional function to wireless environment.
19. system according to claim 18 is characterized in that, the specifying information in the described network environment comprises: position, the classification of service and the concrete network parameter of location network of requiring.
20., it is characterized in that described perception and information translation module comprise that also the information that information conversion function is specially collecting quantizes, and is converted to operable digital information according to any described system of claim 17 to 19.
21. system according to claim 17 is characterized in that, in the described autonomous learning module, the described income of taking all factors into consideration other operators adopts the game mode, is met the radio resource management method of multi-operator competitive relation from game.
22. a joint radio resource management equipment is characterized in that, comprising:
Perception and information translation module are used for to the perception of wireless environment with to the conversion of information;
The autonomous learning module, be connected with the information translation module with perception, be used to put in order data that perception and information translation module transmit as present network state, according to calculating the income level that gets, and carry out the extensive and Knowledge Storage of corresponding data by the extensive and Knowledge Storage module of data, further take all factors into consideration the avail information of other operators and select a kind of appropriate wireless resources management method, send to wireless network control apparatus, send user terminal then to;
Extensive and the knowledge store module of data is connected with the autonomous learning module, and it is extensive to use neural net to do data, is used to reduce required memory space, and historical experience is in the past stored.
23. equipment according to claim 22 is characterized in that, described perception and information translation module are specially the specifying information that obtains in the network environment to described perceptional function to wireless environment.
24. equipment according to claim 23 is characterized in that, the specifying information in the described network environment comprises: the relevant information of network quality, traffic carrying capacity, type of service and other carrier networks.
25., it is characterized in that described perception and information translation module comprise that also the information that information conversion function is specially collecting quantizes, and is converted to operable digital information according to any described equipment of claim 22 to 24.
26. equipment according to claim 22 is characterized in that, in the described autonomous learning module, the described income of taking all factors into consideration other operators adopts the game mode, is met the radio resource management method of multi-operator competitive relation from game.
CN2007101754970A 2007-09-29 2007-09-29 Independent radio resource management method, system and device under multi-operator scenario Expired - Fee Related CN101166342B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2007101754970A CN101166342B (en) 2007-09-29 2007-09-29 Independent radio resource management method, system and device under multi-operator scenario

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2007101754970A CN101166342B (en) 2007-09-29 2007-09-29 Independent radio resource management method, system and device under multi-operator scenario

Publications (2)

Publication Number Publication Date
CN101166342A CN101166342A (en) 2008-04-23
CN101166342B true CN101166342B (en) 2011-09-07

Family

ID=39334753

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2007101754970A Expired - Fee Related CN101166342B (en) 2007-09-29 2007-09-29 Independent radio resource management method, system and device under multi-operator scenario

Country Status (1)

Country Link
CN (1) CN101166342B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110029318A1 (en) * 2008-06-16 2011-02-03 Anna Pucar Rimhagen Global operator license prioritizing and distribution
CN102238590B (en) * 2011-07-07 2014-04-02 曲桦 Cross-layer optimization method for radio network resource management based on joint game
CN102572987B (en) * 2011-12-31 2015-04-15 华中科技大学 Network selection method orienting to heterogeneous wireless network environment
CN105139227B (en) * 2015-08-20 2018-06-26 上海华为技术有限公司 A kind of method for computing data and device

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1588946A (en) * 2004-09-14 2005-03-02 北京邮电大学 Managing system for providing service quality assuranced on internet and its realizing method
CN101001439A (en) * 2006-01-10 2007-07-18 华为技术有限公司 Method for terminal switching between heterogeneous network

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1588946A (en) * 2004-09-14 2005-03-02 北京邮电大学 Managing system for providing service quality assuranced on internet and its realizing method
CN101001439A (en) * 2006-01-10 2007-07-18 华为技术有限公司 Method for terminal switching between heterogeneous network

Also Published As

Publication number Publication date
CN101166342A (en) 2008-04-23

Similar Documents

Publication Publication Date Title
CN1327668C (en) System and method for user scheduling in a communication network
CN109947545A (en) A kind of decision-making technique of task unloading and migration based on user mobility
CN101599854B (en) Method, equipment and system for reconfiguring protocol stack based on assembly
CN101516093B (en) Method and device for judging comprehensive priority of mobile communication system
CN101132363B (en) Free associating wireless resource management system and method based on reinforcement learning
CN101198128B (en) Multi-module equipment and its service stream switching method
CN110809306A (en) Terminal access selection method based on deep reinforcement learning
CN110049566B (en) Downlink power distribution method based on multi-unmanned-aerial-vehicle auxiliary communication network
CN101166342B (en) Independent radio resource management method, system and device under multi-operator scenario
CN111726833A (en) Network load balancing method, device and storage medium
CN105446817B (en) A kind of federated resource configuration of reservations algorithm based on robust optimization in mobile cloud computing
CN102427595A (en) Method and device for controlling adaptive optimization switching for heterogeneous converged wireless network
CN105992252A (en) Processing method and apparatus for context of UE
CN102098684A (en) System and method for allocating cross-layer resources in cognitive radio network
CN108601074A (en) A kind of network resource allocation method and device based on isomery joint caching
CN114422349B (en) Cloud-edge-end-collaboration-based deep learning model training and reasoning architecture deployment method
CN104022859A (en) OWN (Open Wireless Network)-based adaptive flow load sharing technology
CN102364893A (en) Reconfiguration management plane based on cognitive network and reconfiguration method
CN102077526B (en) Method, apparatus and computer program product for distributed information management
CN111885551B (en) Selection and allocation mechanism of high-influence users in multi-mobile social network based on edge cloud collaborative mode
CN113473449A (en) Intelligent connection system based on Internet of things terminal
CN107333273A (en) Multi net voting cut-in method based on grey correlation analysis
CN102711178B (en) Group association accepting control method based on user requirement and network load balancing
CN106304307A (en) A kind of resource allocation methods under heterogeneous network converged
CN102752805A (en) Radio resource distributing method and system based on business satisfaction degree

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20110907

Termination date: 20140929

EXPY Termination of patent right or utility model