CN105069010A - Resource polymerization method based on Agent - Google Patents
Resource polymerization method based on Agent Download PDFInfo
- Publication number
- CN105069010A CN105069010A CN201510394107.3A CN201510394107A CN105069010A CN 105069010 A CN105069010 A CN 105069010A CN 201510394107 A CN201510394107 A CN 201510394107A CN 105069010 A CN105069010 A CN 105069010A
- Authority
- CN
- China
- Prior art keywords
- agent
- task
- service
- resource
- bid
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
Abstract
The invention discloses a resource polymerization method based on Agent; in computer and network technology wide application, resources scattered in a complex heterogeneous network become more and more, and the method can effectively gather the resources scattering in the distributed environment; the method comprises the following steps: configuring the resource polymerization method based on Agent in a distributed system; a task agent resolves and describes according to a user inputted task demand; a decision agent distributes the task to a corresponding service agent according to task description information; the service agent selects a resource agent in an Agent alliance to execute the task, a task execution result is returned to a user Agent, and displayed in a user interface. Communication between the service agent and the resource agent complies with an expansion contract net protocol, thus reducing communication load between Agents, and fast responding to the user task demand. The resource Agent can joint or withdraw from the Agent alliance, so dynamic characteristic is provided, thus enhancing resource polymerization method flexibility and scalability.
Description
Technical field
The invention belongs to Computer Applied Technology field, specifically a kind of aggregation of resources method based on multi-Agent, can be used for exploitation and the realization of the aggregation of resources service in distributed system.
Background technology
Aggregation of resources refers in dynamic, open environment, according to user's request, the resource organizations of dispersion is got up, provides and rationally, easily serve, and realizes sharing and comprehensive utilization of resource.In recent years, along with develop rapidly and the widespread use of computer technology and network technology, the resource be dispersed in complicated heterogeneous network gets more and more, and how these resources to be effectively polymerized in distributed environment, to cause interest and the great attention of Chinese scholars.
In order to realize effective utilization of resource, needing to provide rational resource sharing and polymerization, at present, mainly containing based on the resource share method of SOA framework, the resource restructuring method of sing on web Service and the polymerization based on Agent.
Enterprise SOA SOA is issued by service, find and call and support one to serve to carry out alternately with distribution end-user application in a network or other services, and permission is combined basic service and created more complicated composite services or value-added service.SOA is by the message be defined as alternately between service consumer and ISP between ageng.Service consumer is the ageng of request service, and it finds service, describes and call service; ISP is the ageng performing service, and its issuing service describes and provides service.But SOA also has weak point, the service provided lacks initiative, does not have independence, adaptive ability, lacks collaborative between each service.
The patent " the dynamic integrity development platform system and method based on Agent " (number of patent application: 200910219441.X, publication No. CN101710281A) of Xian Electronics Science and Technology University's application discloses a kind of system dynamic integrity development platform system and method based on Agent.The method is by analyzing the interactive relation between each module, and the hierarchical relationship in certainty annuity between Agent, according to the integrated demand of system, write integrated script, utilize control integration instrument to explain integrated script, spanning set becomes rule, selects corresponding Agent to execute the task.The deficiency of the method is: system needs to generate a large amount of script file, and the modules in system is all be solidificated in platform, and the execution Agent quantity in system is many and when being present in distributed environment, be difficult to select suitable acceptance of the bid Agent.
The information aggregation method based on Agent that Zhejiang University proposes have employed network information crawl technology and webservice technology, its core concept is: whole system can be divided into client, service discovery, data source three parts, client receives the solicited message of user and submits to service discovery Agent, and service discovery Agent selects suitable data message to feed back to user by certain mode according to certain rule in data source (being generally Web page).The weak point of the method: system is generally only applicable to the information fusion in Web application system, and range of application is narrower, the result of polymerization depends on the operation-interface that third party provides, and is difficult to ensure aggregate quality.
The platform that these methods are supplied to user only can represent independently various resource mostly, lacks interactivity, collaborative and initiative between various data.The maintainability of system, extensibility is not enough, and range of application is narrower.
Summary of the invention
The object of the invention is to propose a kind of aggregation of resources method based on multi-Agent, by Agent technology, in distributed environment, the demand according to user finds and combined resource, performs the task that user submits to, realizes the dynamic aggregation of resource service.The present invention effectively improves the dynamic of aggregation of resources with dirigibility, is applicable to the aggregation of resources of distributed system.Technical scheme of the present invention is achieved in that
The present invention is a kind of aggregation of resources method based on multi-Agent, it is characterized in that, comprise aggregation of resources model and extended Contract Net agreement, aggregation of resources model is top-down includes five constituting layers, client layer respectively, task layer, decision-making level, service layer and resource layer, every layer of Agent being solidified with respective level, service layer comprises multiple service agent, each service agent is equal coordination, resource layer comprises multiple resource agents storehouse, each resource agents storehouse is equal coordination, the corresponding resource agents storehouse of each service agent forms an Agent alliance, two-way information interaction is had between service agent and resource agents storehouse, capability service storehouse is provided with between task layer and decision-making level, service ability storehouse and have information interaction respectively between task layer and decision-making level, capability service storehouse two-layerly provides service ability consulting service to above-mentioned.User task demand information is passed to task layer by the User interface Agent of client layer, after task-resource graph receives the task of User interface Agent transmission, Task-decomposing information is delivered to decision-making level by inquiry service ability base, task description distribution of information is passed to Agent alliance by the decision agent of decision-making level, N number of service agent is comprised in service layer, N number of resource agents storehouse is comprised in resource layer, service agent and resource agents storehouse one_to_one corresponding form respective Agent alliance, Agent alliance is exactly ISP, resource agents storehouse comprises one or more resource agents, each service agent is connected with respective public information plate, respective Agent alliance served by respective public information plate, the intrinsic bidding documents Buffer Pool of each resource agents, service agent safeguards a public information plate separately, service agent receives creation task information on bidding after mission bit stream, create bid bidding documents, place it on public information plate, and upgrade common indicium, indicate that new bidding documents arrives, determine whether submit a tender according to self-condition after resource agents in the Agent alliance at service agent place perceives information on bidding, after bid, bid information is sent to service agent, bid information is fed back to by bid information the decision agent of decision-making level by service agent, decision-making level's information is fixed with inquiry case library, after inquiry, bid result is fed back to resource agents by service agent by the decision agent of decision-making level, middle target resource agents is executed the task, service agent in Agent alliance and the message exchange between resource agents are according to extended Contract Net agreement, after task completes, task action result is fed back to user, and display in the user interface.Extended Contract Net agreement comprises: based on the Bidding Strategy of public information plate (PIB), based on the autonomous bidding strategy of bid and the acceptance of the bid strategy in Design case based storehouse.
The resource agents of service agent and its management in the present invention forms Agent alliance, and this alliance is managed by service agent, externally presents specific service function, in the present invention also referred to as ISP.Service agent coordinates the resource agents with administrative institute in alliance by extended Contract Net agreement, realizes the cooperation between service agent and resource agents.
Realization of the present invention is also: initiate to submit a tender in respective Agent alliance after the service agent of service layer receives task, and bidding documents is placed on public information plate PIB, each resource agents in Agent alliance perceives information on bidding, bidding documents is also fed back to decision-making level by service agent by autonomous bid, the decision agent of decision-making level receives the bid information of N number of resource agents, suitable resource agents and successful bidder is selected to execute the task by inquiry case library, present invention reduces traffic when cooperating between service agent and resource agents, improve negotiation efficiency, thus the mission requirements of user can be responded fast, each Agent alliance externally provides the service of certain type, service agent carries out distribution and the coordination of task, resource agents can be selected to add and exit Agent alliance.Aggregation of resources is possessed dynamic perfromance.
In the present invention:
(1) User interface Agent described in, refers to the Agent of completing user and system interaction.
(2) task-resource graph described in, has referred to the Agent that submitting to user of task is decomposed and described.
(3) decision agent described in, refers to the Agent carrying out task distribution and resource selection.
(4) service agent described in, refers to the Agent managed resource agents.
(5) resource agents described in, refers to the Agent specifically executed the task.
(6) the Agent alliance described in, refers to the alliance formed by service agent and its resource agents managed, also referred to as ISP.
Based on the Bidding Strategy of public information plate in the present invention, its main contents are: service layer Agent safeguards a public information plate separately, after service agent receives the task that decision agent sends, create bid bidding documents, place it on public information plate, resource agents obtains new bid bidding documents by public information plate.
The present invention is a kind of aggregation of resources method based on multi-Agent, it is characterized in that, apply based in the aggregation of resources system of multi-Agent above-mentioned, aggregation of resources includes following steps:
(1) in distributed system environment, the resource based on multi-Agent is configured, comprises the startup optimization of Agent respective in each constituting layer and the foundation of Agent alliance.
(2) according to the mission requirements of user's input, mission requirements information is sent to task layer by User interface Agent, and the task-resource graph in task layer decomposes mission requirements and describes, and task description information is transferred to decision-making level as a result.
(3) task description distribution of information is given corresponding Agent alliance by the decision agent of decision-making level.
(4) in decision-making level, carry out between service layer and resource layer calling for bid, submit a tender and getting the bid, the service agent of service layer is after receiving task description information, call for bid in respective Agent alliance, by service agent, bid information is fed back to the decision agent of decision-making level after resource agents perceives information on bidding, bid result is fed back to resource agents by service agent by decision agent, middle target resource agents is executed the task, according to extended Contract Net protocol communication and cooperation between the service agent of service layer and the resource agents of resource layer.
(5) execution result of task is successively upwards fed back to the User interface Agent of client layer by resource agents, and by the display of aggregation of resources result in the user interface.
The present invention compared with prior art tool has the following advantages:
1) Agent technology is applied in aggregation of resources by the present invention, the intelligent characteristic of Agent is utilized to solve resource dispersion in resource service polymerization in process, resource Heterogeneity, and the mission requirements of common completing user are carried out by the cooperation between multiple Agent, enhance the dirigibility of aggregation of resources method, extensibility, and be easy to the conservation of resources in distributed environment.
2) service agent and resource agents are built into different Agent alliances by the present invention, each Agent alliance externally provides the service of certain type, service agent carries out distribution and the coordination of task, resource agents can add and exit Agent alliance, aggregation of resources is possessed dynamic perfromance.
3) cooperation between extended Contract Net protocol realization Agent is used in the present invention, this agreement is improved on the basis of traditional contracts fidonetFido, the Bidding Strategy based on public information plate (PIB) is have employed in bidding process, the acceptance of the bid strategy in Design case based storehouse is have employed in acceptance of the bid process, reduce traffic when cooperating between service agent and resource agents, improve negotiation efficiency, thus the mission requirements of user can be responded fast.
Accompanying drawing explanation
Fig. 1 is System's composition schematic diagram of the present invention;
Fig. 2 is the system emulation pie graph of embodiments of the invention;
Fig. 3 is aggregation of resources method realization flow figure of the present invention;
Fig. 4 is that structural drawing is implemented in the deployment be polymerized based on the resource service of multi-Agent in the present invention;
Fig. 5 is decision agent in the present invention, interactive relation figure between service agent and resource agents;
Fig. 6 is the process flow diagram of extended Contract Net agreement in the present invention, is also that the present invention calls for bid, submits a tender, the schematic flow sheet of process of getting the bid.
Embodiment
Describe the present invention below in conjunction with accompanying drawing.
Embodiment 1
The present invention is a kind of aggregation of resources method based on multi-Agent, see figures.1.and.2 and comprise aggregation of resources model and extended Contract Net agreement, aggregation of resources model is top-down includes five constituting layers, client layer, task layer, decision-making level, service layer and resource layer respectively, every layer of Agent being solidified with respective level.Resource object in this example is the decentralized resource in distributed system, described service layer comprises multiple service agent, each service agent is equal coordination, described resource layer comprises multiple resource agents storehouse, each resource agents storehouse is equal coordination, the corresponding resource agents storehouse of each service agent, two-way information interaction is had between mutual service agent and resource agents storehouse, capability service storehouse is provided with between task layer and decision-making level, service ability storehouse and have information interaction respectively between task layer and decision-making level, capability service storehouse two-layerly provides service ability consulting service to above-mentioned.User task demand information is passed to task layer by the User interface Agent of the client layer in the present invention, after task-resource graph receives the task of User interface Agent transmission, Task-decomposing information is delivered to decision-making level by inquiry service ability base, first inquiry service ability base, if there is service agent can complete separately this task, represent that this task is simple task, do not need to decompose, direct generation task sequence, if do not have service agent can complete separately this task, then need to decompose task, generate executable task sequence, task sequence is sent to decision-making level, task description distribution of information is passed to Agent alliance by the decision agent of decision-making level, N number of service agent is comprised in service layer, N number of resource agents storehouse is comprised in resource layer, service agent and resource agents storehouse one_to_one corresponding form respective Agent alliance, Agent alliance is exactly ISP, resource agents storehouse comprises one or more resource agents, each service agent is connected with respective public information plate, respective Agent alliance served by respective public information plate, the intrinsic bidding documents Buffer Pool of each resource agents, service agent safeguards a public information plate separately, service agent receives creation task information on bidding after mission bit stream, create bid bidding documents, place it on public information plate, and upgrade common indicium, indicate that new bidding documents arrives, determine whether submit a tender according to self-condition after resource agents in the Agent alliance at service agent place perceives information on bidding, after bid, bid information is sent to service agent, bid information is fed back to by bid information the decision agent of decision-making level by service agent, decision-making level's information is fixed with inquiry case library, after inquiry, decision agent finds out the source case information that these resource agents have from inquiry case library, calculate the similarity of these source cases and problem case, obtain the case reliability of resource agents, weighted mean is made together with the ability parameter of each resource agents, draw the scale value of each resource agents, the scale value of each resource agents last sorts, the resource agents that label taking value is the highest is acceptance of the bid Agent, bid result is fed back to resource agents by service agent by the decision agent of decision-making level, middle target resource agents is executed the task, service agent in Agent alliance and the message exchange between resource agents are according to extended Contract Net agreement, after task completes, task action result is fed back to user, and display in the user interface.
In the present invention, extended Contract Net agreement comprises: based on the Bidding Strategy of public information plate (PIB), based on the autonomous bidding strategy of bid and the acceptance of the bid strategy in Design case based storehouse.
Simulation process needs to use JADE (JavaAgentDevelopmentEnvironment) to develop this resource service aggregation platform, and as shown in the upper left side JADE in Fig. 2, in the present invention, all Agent all survive in JADE.JADE is called container (Container) for the running environment that Agent provides, and a platform can have multiple container, and a container can hold multiple Agent.In a JADE platform, have and only have a container being called primary tank (Main-Container), the common vessel being distributed in many places is managed by primary tank.Primary tank is Agent management system (AgentManagementSystem, and directory service (DirectoryFaciliator AMS), DF) resident container, the RMI used by JADE inside is created when registering, Agent in JADE uses message transmission protocol (MessageTransportProtocol, MTP) to communicate.Due to the importance of primary tank, need for primary tank backs up.
Embodiment 2
Based on the formation of the aggregation of resources method model of multi-Agent and principle of work with embodiment 1, aggregation of resources has independence and dynamic perfromance.Initiate to submit a tender in respective Agent alliance after the service agent of service layer receives task, and bidding documents is placed on public information plate PIB, each resource agents in Agent alliance perceives information on bidding, bidding documents is also fed back to decision-making level by service agent by autonomous bid, the decision agent of decision-making level receives the bid information of N number of resource agents, suitable resource agents and successful bidder is selected to execute the task by inquiry case library, present invention reduces traffic when cooperating between service agent and resource agents, improve negotiation efficiency, thus the mission requirements of user can be responded fast, each Agent alliance externally indicates the service that oneself can provide certain type, therefore decision-making level is when distributed tasks, can indicate with reference to each Agent alliance COS, according to finishing the work, required ability selects suitable Agent alliance from the service ability storehouse of system, task is distributed to service layer.According to extended Contract Net agreement, service agent carries out distribution and the coordination of task, comprise one or N number of resource agents in each Agent alliance, in alliance, resource agents can be selected to add and exit Agent alliance, and therefore aggregation of resources of the present invention has possessed independence and dynamic perfromance.
Embodiment 3
The present invention or a kind of aggregation of resources method based on multi-Agent are the aggregation of resources systems based on multi-Agent, based on the aggregation of resources system of multi-Agent with embodiment 1-2,
With reference to Fig. 3, apply based in the aggregation of resources system of multi-Agent above-mentioned, aggregation of resources includes following steps:
(1) in distributed system environment, the resource based on multi-Agent is configured, comprises the startup optimization of User interface Agent, task-resource graph, decision agent, the foundation of Agent alliance.
(2) according to the mission requirements of user's input, mission requirements information is sent to task layer by User interface Agent, the first inquiry service ability base of task-resource graph, if there is service agent can complete separately this task, represent that this task is simple task, do not need to decompose, direct generation task sequence, if do not have service agent can complete separately this task, then need to decompose task, generate executable task sequence, task sequence is sent to decision-making level, task description information is transferred to decision-making level as a result.
(3) task description distribution of information is given corresponding Agent alliance by decision-making level's decision agent.Each Agent alliance externally indicates the service that oneself can provide certain type, therefore decision-making level is when distributed tasks, can indicate with reference to each Agent alliance COS, according to finishing the work, required ability selects suitable Agent alliance from the service ability storehouse of system, and task is distributed to service layer.
(4) in decision-making level, carry out between service layer and resource layer calling for bid, submitting a tender and middle target process, the service agent of service layer is after receiving task description information, call for bid in respective Agent alliance, by service agent, bid information is fed back to the decision agent of decision-making level after resource agents perceives information on bidding, bid result is fed back to resource agents by service agent by decision agent, middle target resource agents is executed the task, according to extended Contract Net protocol communication and cooperation between the service agent of service layer and the resource agents of resource layer.
(5) execution result of task is successively upwards fed back to the User interface Agent of client layer by resource agents, and by the display of aggregation of resources result in the user interface.
The cooperation between extended Contract Net protocol realization Agent is used in the present invention, this agreement is improved on the basis of traditional contracts fidonetFido, the Bidding Strategy based on public information plate (PIB) is have employed in bidding process, the acceptance of the bid strategy in Design case based storehouse is have employed in acceptance of the bid process, reduce traffic when cooperating between service agent and resource agents, improve negotiation efficiency, thus the mission requirements of user can be responded fast.
Embodiment 4
Based on the aggregation of resources method of multi-Agent with embodiment 3, task-resource graph wherein in task layer described in step (2) decomposes mission requirements and description is the first inquiry service ability base of task-resource graph, if there is service agent can complete separately this task, represent that this task is simple task, do not need to decompose, direct generation task sequence, if do not have service agent can complete separately this task, then need to decompose task, generate executable task sequence, task sequence is sent to decision-making level.
Embodiment 5
Based on the aggregation of resources method of multi-Agent with embodiment 3, wherein the bid described in step (4), bid and acceptance of the bid flow process are as follows:
(4a) utilize based on the Bidding Strategy of public information plate PIB, after service agent receives the task that decision agent sends, creation task bid bidding documents propose, is placed on bidding documents on public information plate PIB, starts bid;
(4b) resource agents in the alliance of service agent place knows the bid bidding documents in public information plate, according to the state of self, determine whether submit a tender, if determine bid propose, then create bid submission document by resource agents, and sent to service agent before closing time;
If (4c) service agent does not receive bidding message within closing time, go to step (4h), otherwise go to step (4d) continuation;
(4d) service agent will receive n the reply of submitting a tender produced by resource agents, send to decision agent, decision agent calculates the resource agents ability parameter participating in submitting a tender, comprise: degree of belief, availability, liveness and case reliability, after being weighted on average, obtaining the scale value of resource agents, the scale value of each resource agents is sorted, the resource agents selecting scale value maximum is the task undertaker, and feeds back to service agent;
(4e) service agent sends acceptance of tenders message accept-proposal to middle target resource agents, notifies that it starts to execute the task, simultaneously to not having middle target resource agents to send refusal bidding message reject-proposal;
(4f) after resource agents receives acceptance of tenders message, start to execute the task, if tasks carrying success, transmit a reply message inform-done or inform-result to service agent; If task expires, then advertise service Agent task time-out, service agent determines whether again to initiate new round bid, how to initiate new round bid, go to step (4b) and continue, if do not initiate new round bid, go to step (4g), if tasks carrying failure, but not yet due, then send tasks carrying to service agent and unsuccessfully reply message failure;
(4g) service agent adjusts degree of belief and the liveness of resource agents in this task cooperative process, belongs in Agent alliance and operates;
(4h) bid of epicycle task terminates.
Embodiment 6
Based on the aggregation of resources method of multi-Agent with embodiment 5, wherein the degree of belief described in step (4d), availability, liveness and case reliability definition are as follows:
(1) degree of belief, refers to that resource agents can complete the credibility of certain task;
(2) availability, what refer to that resource agents performs certain task can producing level;
(3) liveness, refer to resource agents to perform certain task play an active part in degree;
(4) case reliability, refer to that the resource agents recorded in Agent case library performs the degree of reliability of certain task, for the resource agents participating in the bid of certain task, the similarity that this resource agents performs the similar cases of this task is calculated by retrieval case library, and combine retrieve the completeness of case, pass judgment on the degree of reliability that this resource agents performs this kind of task in case library.
Embodiment 7
Based on the aggregation of resources method of multi-Agent with embodiment 3-6, describe the present invention below in conjunction with accompanying drawing.
With reference to Fig. 3, implementation step is:
Step 1, is configured the aggregation of resources method based on multi-Agent, comprises the startup optimization of User interface Agent, task-resource graph, decision agent, the foundation of Agent alliance in distributed system environment.
With reference to Fig. 4, in distributed environment, the present invention is when concrete configuration, User interface Agent, task-resource graph and decision agent operate in primary tank Main-Container as the resident Agent in model, the Agent alliance of service agent and resource agents composition is placed in different non-master container C ontainer according to the network site at the type of executing the task and place, and back up decision agent, and backup is in other primary tank.Wherein, the primary tank that non-primary tank depends on belonging to it runs.
Step 2, according to the mission requirements of user's input, to be decomposed mission requirements by task-resource graph and describes.
Analyze the demand of executing the task of user's input, being decomposed into can by the simple task of single resource Agent complete independently, also referred to as subtask.According to mission requirements, set the dependence of the logical OR time between these subtasks, and to the descriptor of going out on missions, the mission requirements of such user can be participated in jointly by multiple resource agents.
Step 3, according to task description information, decision agent is responsible for task to be distributed to corresponding service agent.
Subtask by step 1 gained, according to service ability storehouse, is distributed to corresponding service agent by decision agent, and wherein service ability storehouse describes the service function that in the distributed environment at place, all Agent alliances can provide.
Step 4, service agent, after receiving new task, selects resource agents to execute the task in its Agent alliance, and by the communication between extended Contract Net protocol realization service agent and resource agents and cooperation.
With reference to Fig. 5, when service agent receives new task, use extended Contract Net agreement in alliance's inner initiation task bid behavior, if the task of this service agent has dependence, except task result being sent to User interface Agent, also task result to be sent to the service agent waiting for this task result.After often wheel task cooperative completes, service agent, to the cooperating process of decision agent circular task, is carried out the renewal work of case library by decision agent.Resource agents accepts the bid behavior that service agent is initiated, and determines whether submit a tender, participate in completing of task according to autonomous bidding strategy of submitting a tender.
Communication between service agent with resource agents and cooperate by extended Contract Net protocol realization, this agreement have employed the Bidding Strategy based on public information plate PIB in bidding process, have employed based on autonomous bidding strategy of submitting a tender in bid, in acceptance of the bid process, add the measurement index of degree of belief, availability, liveness and case reliability, this extended Contract Net agreement idiographic flow is with reference to shown in Fig. 6:
(4a) adopt based on the Bidding Strategy of public information plate PIB at bidding phase, after service agent receives the task that decision agent sends, creation task bid bidding documents propose, is placed on public information plate PIB by bidding documents;
(4b) resource agents in the alliance of service agent place knows the bid bidding documents in public information plate, adopt a kind of based on autonomous bidding strategy of submitting a tender, according to the state of self, determine whether submit a tender, if resource agents determines to submit a tender, then create bid submission document, and sent to service agent before closing time.
The current number of having submitted a tender of setting resource agents is num, and the number can simultaneously submitted a tender at most is Size, and its bidding strategy of independently submitting a tender is specially:
If current bidding documents Buffer Pool is full, i.e. num=Size, this resource agents is not submitted a tender, and num is constant;
If current bidding documents Buffer Pool is less than, i.e. num < Size, this resource agents can be submitted a tender, num=num+1 after bid, by this bid submission document write Buffer Pool;
If when this Agent receives the message informing of refusal bid, delete the bid submission document of this task in bidding documents Buffer Pool, num=num-1;
If this Agent gets the bid, by etc. pending task put into the task queue of Agent, after tasks carrying, then num=num-1, deletes the bid submission document of this task in bidding documents Buffer Pool.
If (4c) service agent does not receive reply within closing time, go to step (4h), otherwise go to step (4d) continuation;
(4d) service agent will receive n the reply of submitting a tender produced by resource agents, send to decision agent, decision agent calculates the resource agents ability parameter participating in submitting a tender respectively, comprise: degree of belief, availability, liveness and case reliability, and carry out the scale value of weighted average calculation resource agents, the resource agents selecting scale value maximum is the task undertaker.Resource agents is designated as a, and the task of execution is designated as t, wherein:
Degree of belief Trust (a, t) is calculated as:
1) if resource agents a is the Agent newly added, then need the initial value setting degree of belief, establishing method is: searched out in alliance the resource agents that can perform with generic task t by the service agent of its place alliance, its total number is designated as count, by the resource agents a searched out
ithe respective quantity participating in performing this generic task is designated as m
i, then according to following formulae discovery.
The Agent making the resource agents newly adding system can perform identical function with other like this has a close historical process.
2) resource agents a is after executing task t, and adjust its degree of belief according to following formula, degree of belief is designated as Trust
0.
Trust(a,t)＝α×Trust
0+β×Completed(a,t)(2)
Wherein, α, β are weights, and alpha+beta=1, Trust
0for the trust angle value before the adjustment of this resource agents, Completed (a, t) is completeness, namely resource agents is completed to the evaluation of result of this task t, and its evaluation method is:
Do not complete value 0, complete the imperfect value 0.3 of result, complete but long value 0.7 consuming time, complete value 1.
Availability Available (a) is calculated as:
Wherein, ReplyNum is the number of the issued bidding message propose of this resource agents, and cost (a) is the current bid expense of resource agents, and its computation rule is as follows:
As ReplyNum=0, cost (a)=cost
0=0.1, namely current bid expense is initial cost cost
0;
As 0 < ReplyNum≤Size, cost (a)=cost
0+ 0.25 (2 × ReplyNum-1) cost
0;
Wherein, Size is the number that resource agents can be submitted a tender at most simultaneously.
Liveness Active (a, t) is calculated as:
Wherein, post
tthe number of times of before representing resource agents a, the type task t being submitted a tender altogether, total
trepresent the number of times that in this resource agents place Agent alliance, all resource agents are submitted a tender altogether for the type task t.
Being calculated as of case reliability:
For the resource agents a that certain participates in bid, decision agent retrieves the similar cases obtained n bar (case
1, case
2..., case
n), the similarity of these cases and problem case is respectively Simi
1, Simi
2..., Simi
n, the completeness of these cases is respectively Completed
1, Completed
2..., Completed
n, then case reliability Relibility (a) of resource agents a draws according to following formula.
Relibility(a)＝Max{Simi
1*Completed
1,Simi
2*Completed
2,....,Simi
n*Completed
n}(5)
Wherein, Completed
nthe completeness of case, this value is provided by case library, Simi
nfor similarity, its computation rule is as follows:
1. the attributes similarity computing method for sign pattern are: the value of conventional letter type attribute in case library is x
i, and in problem case, the value of corresponding attribute is y
iif, x
iwith y
iidentical, similarity is 1, otherwise similarity is 0;
2. the attributes similarity computing method for interval number Value Types are: in case library, in certain case, property value is [x
lo, x
up], the value in problem case is [y
lo, y
up], be calculated as follows:
Wherein, a, b are respectively smallest limit and the maximum upper limit of this attribute.
3. for value type attribute between similarity calculating method as follows:
Determine that the value of several type attribute k in the case of source is x
k, the value in problem case is y
k, then determine to count and determine that distance computing formula between number is according to shown in following formula:
Wherein, a, b are respectively minimum value and the maximal value of this attribute.
(4e) service agent sends acceptance of tenders message accept-proposal to middle target resource agents, notifies that it starts to execute the task, simultaneously to not having middle target resource agents to send refusal bidding message reject-proposal;
(4f) after resource agents receives acceptance of tenders message, start to execute the task, if tasks carrying success, transmit a reply message inform-done or inform-result to service agent; If task expires, then advertise service Agent task time-out, service agent determines whether again to initiate new round bid, how to initiate new round bid, go to step (4b) and continue, if do not initiate new round bid, go to step (4g), if tasks carrying failure, but not yet due, then send tasks carrying to service agent and unsuccessfully reply message failure;
(4g) service agent adjusts degree of belief and the liveness capacity factor of resource agents in this task cooperative process;
(4h) epicycle task cooperative terminates.
Step 5, finally feeds back to User interface Agent by the execution result of task, and display in the user interface.
Prove that decentralized resource is effectively polymerized by the present invention in distributed environment through simulating, verifying.
To sum up, the invention discloses a kind of aggregation of resources method based on multi-Agent, be that in the widespread use for computer and network technology, the resource be dispersed in complicated heterogeneous network gets more and more, solve the technical matters of in distributed environment, decentralized resource being carried out effectively polymerization.Implementation step is: be configured the aggregation of resources method based on multi-Agent in a distributed system; According to the mission requirements of user's input, task-resource graph decomposes it and describes; According to task description information, task is distributed to corresponding service agent by decision agent; Service agent selects resource agents to execute the task in Agent alliance, and the execution result of task feeds back to User interface Agent, and display in the user interface.Extended Contract Net agreement is followed in communication between service agent of the present invention and resource agents, reduces the traffic between Agent, can respond the mission requirements of user fast.Resource agents can be selected to add and exit Agent alliance, possesses dynamic perfromance, enhances the dirigibility of aggregation of resources method, extensibility.
Claims (8)
1. the aggregation of resources method based on multi-Agent, it is characterized in that, comprise aggregation of resources model and extended Contract Net agreement, aggregation of resources model is top-down includes five constituting layers, client layer respectively, task layer, decision-making level, service layer and resource layer, every layer of Agent being solidified with respective level, two-way information interaction is had between service agent and resource agents storehouse, capability service storehouse is provided with between task layer and decision-making level, service ability storehouse and have information interaction respectively between task layer and decision-making level, user task demand information is passed to task layer by the User interface Agent of client layer, after task-resource graph receives task, Task-decomposing information is delivered to decision-making level by inquiry service ability base, task description distribution of information is passed to Agent alliance by decision-making level, N number of service agent is comprised in service layer, N number of resource agents storehouse is comprised in resource layer, service agent and resource agents storehouse one_to_one corresponding form respective Agent alliance, resource agents storehouse comprises one or more resource agents, each service agent is connected with respective public information plate, respective Agent alliance served by respective public information plate, the intrinsic bidding documents Buffer Pool of each resource agents, bid information is fed back to decision-making level by service agent after perceiving information on bidding by resource agents, decision-making level is fixed with inquiry case library, after inquiry, bid result is fed back to resource agents by service agent by decision-making level, middle target resource agents is executed the task, service agent in Agent alliance and the message exchange between resource agents are according to extended Contract Net agreement, after task completes, task action result is fed back to user, and display in the user interface, extended Contract Net agreement comprises: based on the Bidding Strategy of public information plate (PIB), based on the autonomous bidding strategy of bid and the acceptance of the bid strategy in Design case based storehouse.
2. the aggregation of resources method based on multi-Agent according to claim 1, it is characterized in that, aggregation of resources has independence and dynamic, initiate to submit a tender in respective Agent alliance after service layer receives task, and bidding documents is placed on public information plate PIB, each resource agents in Agent alliance perceives information on bidding, bidding documents is also fed back to decision-making level by service agent by autonomous bid, decision-making level receives the bid information of N number of resource agents, suitable resource agents and successful bidder is selected to execute the task by inquiry case library, each Agent alliance externally indicates the service that oneself can provide certain type, according to extended Contract Net agreement, service agent carries out distribution and the coordination of task, resource agents can be selected to add and exit Agent alliance.
3. the aggregation of resources method based on multi-Agent according to claim 1, is characterized in that:
(1) User interface Agent described in, refers to the Agent of completing user and system interaction;
(2) task-resource graph described in, has referred to the Agent that submitting to user of task is decomposed and described;
(3) decision agent described in, refers to the Agent carrying out task distribution and resource selection;
(4) service agent described in, refers to the Agent managed resource agents;
(5) resource agents described in, refers to the Agent specifically executed the task;
(6) the Agent alliance described in, refers to the alliance formed by service agent and its resource agents managed, also referred to as ISP.
4. the aggregation of resources method based on multi-Agent according to claim 1, it is characterized in that, the described Bidding Strategy based on public information plate, its main contents are: service layer Agent safeguards a public information plate separately, after service agent receives the task that decision agent sends, create bid bidding documents, place it on public information plate, resource agents obtains new bid bidding documents by public information plate.
5. the aggregation of resources method based on multi-Agent according to claim 1, it is characterized in that, aggregation of resources includes following steps:
(1) in distributed system environment, the resource based on Agent is configured, comprises the startup optimization of Agent respective in each constituting layer and the foundation of Agent alliance;
(2) according to the mission requirements of user's input, mission requirements information is sent to task layer by User interface Agent, and the task-resource graph in task layer decomposes mission requirements and describes, and task description information is transferred to decision-making level as a result;
(3) task description distribution of information is given corresponding Agent alliance by the decision agent of decision-making level;
(4) in decision-making level, carry out between service layer and resource layer calling for bid, submit a tender and getting the bid, the service agent of service layer is after receiving task description information, call for bid in respective Agent alliance, by service agent, bid information is fed back to the decision agent of decision-making level after resource agents perceives information on bidding, bid result is fed back to resource agents by service agent by decision agent, middle target resource agents is executed the task, according to extended Contract Net protocol communication and cooperation between the service agent of service layer and the resource agents of resource layer;
(5) execution result of task is successively upwards fed back to the User interface Agent of client layer by resource agents, and by the display of aggregation of resources result in the user interface.
6. the aggregation of resources method based on multi-Agent according to claim 5, it is characterized in that: the task-resource graph in task layer described in step (2) decomposes mission requirements and description is the first inquiry service ability base of task-resource graph, if there is service agent can complete separately this task, represent that this task is simple task, do not need to decompose, direct generation task sequence, if do not have service agent can complete separately this task, then need to decompose task, generate executable task sequence, task sequence is sent to decision-making level.
7. the aggregation of resources method based on multi-Agent according to claim 5, is characterized in that: the bid described in step (4), bid and its flow process of acceptance of the bid are as follows:
(4a) utilize based on the Bidding Strategy of public information plate PIB, after service agent receives the task that decision agent sends, creation task bid bidding documents propose, is placed on bidding documents on public information plate PIB, starts bid;
(4b) resource agents in the alliance of service agent place knows the bid bidding documents in public information plate, according to the state of self, determine whether submit a tender, if determine bid propose, then create bid submission document by resource agents, and sent to service agent before closing time;
If (4c) service agent does not receive bidding message within closing time, go to step (4h), otherwise go to step (4d) continuation;
(4d) service agent will receive n the reply of submitting a tender produced by resource agents, send to decision agent, decision agent calculates the resource agents ability parameter participating in submitting a tender, comprise: degree of belief, availability, liveness and case reliability, after being weighted on average, obtaining the scale value of resource agents, the scale value of each resource agents is sorted, the resource agents selecting scale value maximum is the task undertaker, and feeds back to service agent;
(4e) service agent sends acceptance of tenders message accept-proposal to middle target resource agents, notifies that it starts to execute the task, simultaneously to not having middle target resource agents to send refusal bidding message reject-proposal;
(4f) after resource agents receives acceptance of tenders message, start to execute the task, if tasks carrying success, transmit a reply message inform-done or inform-result to service agent; If task expires, then advertise service Agent task time-out, service agent determines whether again to initiate new round bid, how to initiate new round bid, go to step (4b) and continue, if do not initiate new round bid, go to step (4g), if tasks carrying failure, but not yet due, then send tasks carrying to service agent and unsuccessfully reply message failure;
(4g) service agent adjusts degree of belief and the liveness of resource agents in this task cooperative process, belongs in Agent alliance and operates;
(4h) bid of epicycle task terminates.
8. the aggregation of resources method based on multi-Agent according to claim 7, is characterized in that: wherein the degree of belief described in step (4d), availability, liveness and case reliability definition are as follows:
(1) degree of belief, refers to that resource agents can complete the credibility of certain task;
(2) availability, what refer to that resource agents performs certain task can producing level;
(3) liveness, refer to resource agents to perform certain task play an active part in degree;
(4) case reliability, refer to that the resource agents recorded in Agent case library performs the degree of reliability of certain task, for the resource agents participating in the bid of certain task, the similarity that this resource agents performs the similar cases of this task is calculated by retrieval case library, and combine retrieve the completeness of case, pass judgment on the degree of reliability that this resource agents performs this kind of task in case library.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510394107.3A CN105069010B (en) | 2015-07-07 | 2015-07-07 | Resource polymerization method based on multi-Agent |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510394107.3A CN105069010B (en) | 2015-07-07 | 2015-07-07 | Resource polymerization method based on multi-Agent |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105069010A true CN105069010A (en) | 2015-11-18 |
| CN105069010B CN105069010B (en) | 2018-04-17 |
Family
ID=54498383
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510394107.3A Active CN105069010B (en) | 2015-07-07 | 2015-07-07 | Resource polymerization method based on multi-Agent |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105069010B (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105975332A (en) * | 2016-05-03 | 2016-09-28 | 北京理工大学 | Method for forming multi-agent distributed union |
| CN106550028A (en) * | 2016-10-25 | 2017-03-29 | 广东科海信息科技股份有限公司 | A kind of Multi-Agent Negotiation model of Service-Oriented Architecture Based |
| CN106845790A (en) * | 2016-12-27 | 2017-06-13 | 合肥城市云数据中心股份有限公司 | A kind of local service system and its local service access method based on multi-Agent technology in single operation system |
| CN106873562A (en) * | 2017-04-13 | 2017-06-20 | 冶金自动化研究设计院 | Energy management Agent system implementation method based on JADE platforms |
| CN108073699A (en) * | 2017-12-12 | 2018-05-25 | 中国联合网络通信集团有限公司 | Big data polymerization analysis method and device |
| CN108182116A (en) * | 2018-01-23 | 2018-06-19 | 江苏国泰新点软件有限公司 | A kind of bidding documents analysis method, device, equipment and storage medium |
| CN110162400A (en) * | 2019-05-21 | 2019-08-23 | 湖南大学 | The method and system of intelligent body cooperation in MAS system is realized under complex network environment |
| CN111160525A (en) * | 2019-12-17 | 2020-05-15 | 天津大学 | Task unloading intelligent decision method based on unmanned aerial vehicle group in edge computing environment |
| CN111199359A (en) * | 2020-01-08 | 2020-05-26 | 中国电子科技集团公司第五十四研究所 | Multi-agent task allocation method under network resource constraint |
| CN111898908A (en) * | 2020-07-30 | 2020-11-06 | 华中科技大学 | Production line scheduling system and method based on multiple wisdom bodies |
| CN112306684A (en) * | 2020-10-29 | 2021-02-02 | 天津蓝鳍海洋工程有限公司 | Deep water space data integration-oriented LQAFCN-based task control method |
| CN113256094A (en) * | 2021-05-17 | 2021-08-13 | 安徽帅尔信息科技有限公司 | Service resource allocation method based on improved particle swarm optimization |
| CN113315812A (en) * | 2021-04-30 | 2021-08-27 | 桂林理工大学 | Agent-based trust management system in cloud environment |
| CN113961726A (en) * | 2021-12-20 | 2022-01-21 | 中国人民解放军战略支援部队航天工程大学士官学校 | Command task matching method and system |
| CN116468229A (en) * | 2023-04-03 | 2023-07-21 | 四川大学 | Distributed combination contract net warhead firepower distribution method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2002080055A3 (en) * | 2001-03-29 | 2003-04-03 | British Telecomm | Work allocation system |
| CN101710281A (en) * | 2009-12-11 | 2010-05-19 | 西安电子科技大学 | Dynamic integrated system and method of development platform based on Agent |
| CN101815326A (en) * | 2010-01-11 | 2010-08-25 | 北京邮电大学 | Method for allocating tasks in wireless sensor network based on negotiation |
-
2015
- 2015-07-07 CN CN201510394107.3A patent/CN105069010B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2002080055A3 (en) * | 2001-03-29 | 2003-04-03 | British Telecomm | Work allocation system |
| CN101710281A (en) * | 2009-12-11 | 2010-05-19 | 西安电子科技大学 | Dynamic integrated system and method of development platform based on Agent |
| CN101815326A (en) * | 2010-01-11 | 2010-08-25 | 北京邮电大学 | Method for allocating tasks in wireless sensor network based on negotiation |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105975332A (en) * | 2016-05-03 | 2016-09-28 | 北京理工大学 | Method for forming multi-agent distributed union |
| CN106550028A (en) * | 2016-10-25 | 2017-03-29 | 广东科海信息科技股份有限公司 | A kind of Multi-Agent Negotiation model of Service-Oriented Architecture Based |
| CN106845790A (en) * | 2016-12-27 | 2017-06-13 | 合肥城市云数据中心股份有限公司 | A kind of local service system and its local service access method based on multi-Agent technology in single operation system |
| CN106873562A (en) * | 2017-04-13 | 2017-06-20 | 冶金自动化研究设计院 | Energy management Agent system implementation method based on JADE platforms |
| CN108073699A (en) * | 2017-12-12 | 2018-05-25 | 中国联合网络通信集团有限公司 | Big data polymerization analysis method and device |
| CN108073699B (en) * | 2017-12-12 | 2020-06-16 | 中国联合网络通信集团有限公司 | Big data aggregation analysis method and device |
| CN108182116A (en) * | 2018-01-23 | 2018-06-19 | 江苏国泰新点软件有限公司 | A kind of bidding documents analysis method, device, equipment and storage medium |
| CN110162400A (en) * | 2019-05-21 | 2019-08-23 | 湖南大学 | The method and system of intelligent body cooperation in MAS system is realized under complex network environment |
| CN111160525A (en) * | 2019-12-17 | 2020-05-15 | 天津大学 | Task unloading intelligent decision method based on unmanned aerial vehicle group in edge computing environment |
| CN111199359A (en) * | 2020-01-08 | 2020-05-26 | 中国电子科技集团公司第五十四研究所 | Multi-agent task allocation method under network resource constraint |
| CN111898908A (en) * | 2020-07-30 | 2020-11-06 | 华中科技大学 | Production line scheduling system and method based on multiple wisdom bodies |
| CN111898908B (en) * | 2020-07-30 | 2023-06-16 | 华中科技大学 | Production line scheduling system and method based on multiple intelligent objects |
| CN112306684A (en) * | 2020-10-29 | 2021-02-02 | 天津蓝鳍海洋工程有限公司 | Deep water space data integration-oriented LQAFCN-based task control method |
| CN112306684B (en) * | 2020-10-29 | 2023-02-17 | 天津蓝鳍海洋工程有限公司 | Deep water space data integration-oriented LQAFCN-based task control method |
| CN113315812A (en) * | 2021-04-30 | 2021-08-27 | 桂林理工大学 | Agent-based trust management system in cloud environment |
| CN113256094A (en) * | 2021-05-17 | 2021-08-13 | 安徽帅尔信息科技有限公司 | Service resource allocation method based on improved particle swarm optimization |
| CN113256094B (en) * | 2021-05-17 | 2022-09-13 | 安徽帅尔信息科技有限公司 | Service resource allocation method based on improved particle swarm optimization |
| CN113961726A (en) * | 2021-12-20 | 2022-01-21 | 中国人民解放军战略支援部队航天工程大学士官学校 | Command task matching method and system |
| CN113961726B (en) * | 2021-12-20 | 2022-03-01 | 中国人民解放军战略支援部队航天工程大学士官学校 | Command task matching method and system |
| CN116468229A (en) * | 2023-04-03 | 2023-07-21 | 四川大学 | Distributed combination contract net warhead firepower distribution method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105069010B (en) | 2018-04-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105069010A (en) | Resource polymerization method based on Agent | |
| CN111679905B (en) | Calculation network fusion network model system | |
| Lin et al. | A time-driven data placement strategy for a scientific workflow combining edge computing and cloud computing | |
| Yin et al. | An efficient collaboration and incentive mechanism for Internet of Vehicles (IoV) with secured information exchange based on blockchains | |
| CN104461740B (en) | A kind of cross-domain PC cluster resource polymerization and the method for distribution | |
| CN113364831B (en) | Multi-domain heterogeneous computing network resource credible cooperation method based on block chain | |
| Zhao et al. | Edge-MapReduce-based intelligent information-centric IoV: Cognitive route planning | |
| Pechoucek et al. | A knowledge-based approach to coalition formation | |
| Liao et al. | Securing collaborative environment monitoring in smart cities using blockchain enabled software-defined internet of drones | |
| CN101086778A (en) | System and method for providing policy hierarchy in an enterprise data processing system | |
| Asensio et al. | Designing an efficient clustering strategy for combined Fog-to-Cloud scenarios | |
| CN103701894A (en) | Method and system for dispatching dynamic resource | |
| Wang et al. | BC-mobile device cloud: A blockchain-based decentralized truthful framework for mobile device cloud | |
| CN109343945A (en) | A kind of multitask dynamic allocation method based on contract net algorithm | |
| CN115115451A (en) | Block chain service management system for bulk commodity transaction supervision | |
| Li | Green technology innovation path based on blockchain algorithm | |
| CN101783768A (en) | Quantity assurance method of grid service based on resource reservation | |
| CN106095591A (en) | A kind of virtual machine two-stage optimizing management and running platform based on cloud computing | |
| CN103051730B (en) | Multi-source information service-resource allocating system and IA-Min allocating method in cloud-computing business environment | |
| Liu et al. | 5G network education system based on multi-trip scheduling optimization model and artificial intelligence | |
| Hao et al. | Evaluation of nine heuristic algorithms with data‐intensive jobs and computing‐intensive jobs in a dynamic environment | |
| CN102223385A (en) | Multi-agent-based grid geographic information system (GIS) resource management system | |
| Patra et al. | Enforcing social laws in goal directed communication | |
| CN103617084B (en) | Emergency command cloud service implementation method and system based on microkernel cluster | |
| Xhafa et al. | Jxta-Overlay: An interface for efficient peer selection in P2P JXTA-based systems |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |