CN109409773A - A kind of earth observation resource dynamic programming method based on Contract Net Mechanism - Google Patents
A kind of earth observation resource dynamic programming method based on Contract Net Mechanism Download PDFInfo
- Publication number
- CN109409773A CN109409773A CN201811351029.9A CN201811351029A CN109409773A CN 109409773 A CN109409773 A CN 109409773A CN 201811351029 A CN201811351029 A CN 201811351029A CN 109409773 A CN109409773 A CN 109409773A
- Authority
- CN
- China
- Prior art keywords
- task
- resource
- planning
- contract net
- distributed
- 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
- 238000000034 method Methods 0.000 title claims abstract description 168
- 230000007246 mechanism Effects 0.000 title claims abstract description 31
- 238000013439 planning Methods 0.000 claims abstract description 98
- 230000008569 process Effects 0.000 claims abstract description 32
- 238000009826 distribution Methods 0.000 claims abstract description 28
- 238000004891 communication Methods 0.000 claims description 19
- 230000006854 communication Effects 0.000 claims description 17
- 238000007726 management method Methods 0.000 claims description 8
- 238000010845 search algorithm Methods 0.000 claims description 8
- 238000007667 floating Methods 0.000 claims description 7
- 230000008901 benefit Effects 0.000 abstract description 8
- 230000001360 synchronised effect Effects 0.000 abstract description 4
- 230000011218 segmentation Effects 0.000 abstract description 2
- 238000004422 calculation algorithm Methods 0.000 description 39
- 230000008859 change Effects 0.000 description 16
- 238000010586 diagram Methods 0.000 description 15
- 230000000052 comparative effect Effects 0.000 description 8
- 238000005457 optimization Methods 0.000 description 8
- 230000000694 effects Effects 0.000 description 7
- 101100005071 Drosophila melanogaster CanB gene Proteins 0.000 description 6
- 238000002474 experimental method Methods 0.000 description 6
- 230000010354 integration Effects 0.000 description 6
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 238000003780 insertion Methods 0.000 description 4
- 230000037431 insertion Effects 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 239000000203 mixture Substances 0.000 description 4
- GOLXNESZZPUPJE-UHFFFAOYSA-N spiromesifen Chemical compound CC1=CC(C)=CC(C)=C1C(C(O1)=O)=C(OC(=O)CC(C)(C)C)C11CCCC1 GOLXNESZZPUPJE-UHFFFAOYSA-N 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 230000007423 decrease Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 230000002195 synergetic effect Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 230000007812 deficiency Effects 0.000 description 2
- 238000005265 energy consumption Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000010921 in-depth analysis Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 230000002045 lasting effect Effects 0.000 description 2
- 230000033001 locomotion Effects 0.000 description 2
- 238000005295 random walk Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 229920006395 saturated elastomer Polymers 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000013468 resource allocation Methods 0.000 description 1
- 239000000523 sample Substances 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06312—Adjustment or analysis of established resource schedule, e.g. resource or task levelling, or dynamic rescheduling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/08—Auctions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
- G06Q50/265—Personal security, identity or safety
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Theoretical Computer Science (AREA)
- Tourism & Hospitality (AREA)
- General Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Educational Administration (AREA)
- Finance (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Security & Cryptography (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to satellite remote sensing fields, disclose a kind of earth observation resource dynamic programming method based on Contract Net Mechanism, using distributed collaboration planning framework from bottom to top and the collaborative planning process based on contract net；Large-scale concurrent task is dynamically distributed by taking turns incomplete combination distribution method more.The present invention is on the basis of analyzing existing planning system with the resource method of operation from bottom architecture, break through the mindset of intrinsic Planning Model from top to bottom, in conjunction with contract net distributed computing advantage towards the Dynamic Programming problem of empty world heterogeneous resource, it proposes a kind of distributed contract net collaborative planning frame from bottom to top and provides planning process, to give full play to the calculating advantage of distributed resource and then improve task allocative efficiency.On this basis, the incomplete combination distribution method of multiclass towards extensive task is proposed by using three kinds of combined task segmentation, the synchronous distribution of multitask collection, multi-level matching strategies, can quickly distribute a large amount of concurrent tasks.
Description
Technical field
The present invention relates to a kind of earth observation resource dynamic programming method based on Contract Net Mechanism.
Background technique
Space-air-ground integration earth observation including the observation resource such as satellite, unmanned plane, ground transaucer is in environment
Many aspects such as monitoring, Disaster Assessment, city analysis and national security play key effect.Different application fields are to observation
Resource spatial resolution, time window, in terms of have different needs.With industrial production and observation method
It improves, earth observation increasingly develops to fining, to meet the different earth observation demand in each field.However, single class money
Due to the limitation of its intrinsic observed pattern and observing capacity, single resource is difficult to meet numerous isomery observation requirements in source, so
The collaborative planning of heterogeneous resource is inevitable development trend.In addition, external environment is dynamic with height during execution task
State property and uncertainty, task may change at any time, and observation resource also faces damage and the danger with communication chain fracture at any time.Institute
To need to carry out dynamic adjustment to original observation program at any time.
The collaborative planning research of earth observation resource mostly concentrates on the scheduling of single class resource, multiclass isomery earth observation money
The collaborative planning method in source is still at an early stage.The programming dispatching technology of the list class resource such as satellite, unmanned plane, ground installation is
Gradually tend to mature, all kinds of Heuristic Models, intelligent optimization algorithm are widely applied.Due to single class resource observed pattern, energy
The observation income of the inherent limitation of power, isolated planning has reached bottleneck, faces precision, diversified observation requirements, and multiclass is seen
The overall planning method for surveying resource gradually attracts attention.
In terms of the collaborative planning of multiclass earth observation resource, existing method usually proposes a kind of hierarchical from top to bottom first
Heterogeneous resource collaborative planning framework, integrate it is different observation one loose couplings of resource composition earth observation systems.Its framework
Include 4 different levels: input layer, collaboration layer, mission planning layer and observation resource layer.Input layer is responsible for receiving and manages user
The task of submission, collaboration layer possess the global view of planning system, comprising a planning center and manage all second level as a whole
Planning center.Mission planning layer includes multiple subplan centers, and each subplan center is responsible for managing its single classification having under its command
Observe resource.Observation resource layer includes numerous heterogeneous resources and is responsible for execution task.Based on this, this scheme is by multiclass resource
Collaborative planning process regard task assignment procedure as, pass through analysis different resource observation income, observation airplane meeting and resource capability
Etc. conditions construct 3 observation airplane meeting, conflict degree and degree of resource consumption heuristic criterion, and then propose a kind of towards central plan
Task is distributed to multiple second level subplans center by planning center with this by heuristic Task Assignment Model, subplan center into
And combine existing single class resource planning mode assignments observation mission.
Existing multiclass earth observation resource coordinating planing method is suitable for the resource planning under static environment.However in reality
During the task execution on border, external environment and user's subjective intention have highly dynamic property and uncertainty, and this requires associations
With planing method there is high-volume task quickly to plan and draws ability with weight-normality.
The shortcomings of the prior art specifically includes that the resource planning frame of 1) its formula from top to bottom used includes one
Planning center, thus the downward distribution task of centre punch one, the mode of this similar centralization planning will lead to system with weaker
Robustness and low allocative efficiency, and require planning center have powerful computing capability and well stable communication loop
Border；2) existing collaborative planning technology has ignored the highly dynamic property and uncertainty of task execution environment, it is difficult to dynamically part
Adjust original programme, it is difficult to quickly weight planning tasks allocation result；3) prior art frame does not give full play to observation money
The distributed computation ability of source alone, it is difficult to cope with the planning of extensive pop-up mission.
Each field earth observation task has diversity, complexity and a dynamic, and typical extensive task and
Feature is sent out, this proposes the abilities such as accuracy, duration, timeliness and the strain rate effect of empty world synergistic observation system higher
Requirement.And in current observation system, all kinds of observation platforms mostly independent operating, this isolated mission planning operating mode
Far from the observation requirements met in the case of Multi-task Concurrency, becomes and restrict the bottleneck that emergency disaster task is quickly handled.In addition,
The research of a small amount of isomery observation resource coordinating planning is still limited to static external environment, it is difficult to agree with dynamic actual task
Planning demand.Therefore, space-air-ground integration earth observation resource is effectively organized, efficient, dynamic synergistic observation network is formed,
It is comprehensive to meet calamity emergency demand with the calamity emergency task for supporting multiple types concurrent, it has also become space-air-ground integration is over the ground
The significant challenge that observation application faces.The essence of heterogeneous resource dynamic programming problems is task and money based on original observation program
The dynamic Rapid matching in source.
Heterogeneous resource dynamic cooperation planning process is primarily present following difficult point and requires: firstly, empty world observation resource is not
Only spatially discrete distribution, using constraint, maneuverability, load performance, observing capacity, it is motor-driven in terms of all deposit
In very big difference.There is relatively independent planning system at each planning center, it is difficult to construct a distributed system frame and plan as a whole to adjust
Spend all observation resources.Secondly, observing environment has height uncertain, observation resource is necessary during mission planning can dynamic
Access system, and in the case where resource capability constantly changes can dynamic high-efficiency handle the observation mission that occurs at random.
Finally, observation mission of meeting an urgent need has random concurrency, planning process must be on the basis of maximizing overall observation income dynamically
Distribute large-scale concurrent tasks.
Currently, each planning center possesses relatively independent observation resource in earth observation field, planning system is relatively only
It is vertical.It as shown in Figure 1, upload and data receiver center as instruction, may be implemented to communicate between each planning center, and belong to same
Has communication condition between the partial interior resource at one planning center, the resource for being not belonging to same planning center, which is then difficult to realize, leads to
Letter.When have emergency Disaster Event occur when, monitoring task usually by user according to observation resource distributing position be submitted near
Planning center, this planning center further according in set heuristic Rule set to resource node distribute task.It is traditional from upper and
The distribution frame of following formula generallys use this with different levels structure, from high-level Global motion planning center to the part money of secondary level
Task is gradually distributed in source.However the planning environment of Dynamic Uncertain is difficult to ensure its desired strong computing capability and well stable
Communication environment.
Summary of the invention
The object of the present invention is to provide a kind of the earth observation resource dynamic programming method based on Contract Net Mechanism, emphasis solution
The certainly distributed integration of isomery earth observation resource and the quick planning problem under dynamic uncertain environments.
To achieve the goals above, the present invention provides a kind of earth observation resource Dynamic Programming side based on Contract Net Mechanism
Method, using distributed collaboration planning framework from bottom to top and the collaborative planning process based on contract net；By taking turns not exclusively more
Combination distribution method dynamically distributes large-scale concurrent task.
Further, the distributed collaboration planning framework from bottom to top includes following four level:
Task management layer is responsible for receiving the task that user submits, and extensive task is divided according to resource observing capacity
At different task-sets, and then task-set is distributed to the respective resources of next level；
Resource coordinating layer, including numerous heterogeneous resources, observe resource receive task-set to be done and with need weight
The task of planning is distributed to through consultation together can be with the neighborhood resource set of this source communications；On this basis, resource node will
Unassigned task is committed to next level；
Plan that central interior cooperates with layer, according to resource observing capacity by unallocated task assignment on the basis of inside is negotiated
The internal resource administered to this planning center；
Planning center cooperates with layer, and planning center then continues the unfinished task of residue to be committed to the bid of other planning centers.
Further, the collaborative planning process based on contract net is carried out with the active bidding pattern of earth observation resource
Task distribution.
Further, the active bidding pattern includes following five stages:
Phase 1: resource riReceive the task sequence T=(t that need to be inserted into1,t2,…tn), when j ∈ n is instructed, according to certainly
Body caching task to be done and T actively discriminate whether to have the ability to complete T；
Phase 2: if riTask T cannot be completed, then riRole switch to issue of bidding documents person, the resource that can be communicated to itself
Group RSi-LIssue information on bidding；According to RSi-LReturn scale value and complete information judgement can complete task T；
Phase 3: if all resources cannot all complete T, riMission bit stream is fed back in the subplan of upper level
Heart Pk；PkResource issue of bidding documents as from issue of bidding documents person to planning central interior, riRole bidder is switched to by issue of bidding documents person；PkAccording to returning
It returns scale value and completion task subset judges whether there is resource and can complete T；
Phase 4: if completing T, P without resourcekCenter is planned to the neighborhood that can be communicated as issue of bidding documents person again
Set PSk-LIssue bidding documents；Each planning center [P1,P2,…Pk-1,Pk+1…Ppn] respectively specified that and can be completed according to internal resource
Task subset and feed back；PkFinal task allocation plan is evaluated according to feedback result；
Phase 5: if PkNeighborhood planning center cannot complete T then according to Neighbourhood set PSk-LCommunication proximity continue
Expand issue of bidding documents range, until can complete this task T or without remaining planning center.
Further, more incomplete combination distribution method of wheel is respectively successively on three levels from bottom to top to appointing
Business collection is allocated:
In first level, based on contract network method is respectively by each resource as bid its communication proximity resource set of direction
Bidding documents is issued, combined task allocation plan is determined by resource negotiation；
In second level, each planning center receives the task of not completing respectively, and specified planning central interior resource is complete
At task；
In third level, all planning centers joint consultation determines abortive allocation plan.
Further, heterogeneous resource collaborative planning process is solved using the local search algorithm of mark mechanism is selected based on floating
The determining problem of middle contract net victory mark.
Through the above technical solutions, following beneficial technical effect may be implemented:
The present invention is based on based on the collaborative planning of isomery earth observation resource to existing empty world planning system and dynamic
The in-depth analysis of task matching problem under environment, the innate advantage in conjunction with Contract Net Mechanism in distributed planning field, for different
The distributed collaboration and dynamic high-efficiency of structure resource plan two key problems, propose distributed collaboration planning framework from bottom to top
And the extensive task dynamic allocation method of contract net driving, to realize between the isomery observation resource towards dynamic task allocation
Loose couplings, play the potential computing capabilitys of all distributed observation resources to the maximum extent, improve space-air-ground integration pair
The observation benefit of resource is observed on ground, breaks through the technical bottleneck that extensive task is quickly planned under dynamic uncertain environments, final quasi-
Really, calamity emergency department quickly, reliably, is neatly directly served in, disaster assistance decision is instructed.
The present invention, from bottom architecture, is broken through from upper on the basis of analyzing existing planning system with the resource method of operation
And the mindset of lower intrinsic Planning Model, in conjunction with contract net distributed computing advantage towards the dynamic of empty world heterogeneous resource
It plans problem, innovatively propose a kind of distributed contract net collaborative planning frame from bottom to top and provides planning process, with
It gives full play to the calculating advantage of distributed resource and then improves task allocative efficiency.On this basis, by using combined task
Segmentation, multitask collection, which synchronize, to be distributed, matches the three kinds of multiclass of strategy proposition towards extensive task at many levels and not exclusively combine and divide
Method of completing the square can quickly distribute a large amount of concurrent tasks.
The other feature and advantage of the embodiment of the present invention will the following detailed description will be given in the detailed implementation section.
Detailed description of the invention
Attached drawing is to further understand for providing to the embodiment of the present invention, and constitute part of specification, under
The specific embodiment in face is used to explain the present invention embodiment together, but does not constitute the limitation to the embodiment of the present invention.Attached
In figure:
Fig. 1 is empty world resource land observation system schematic diagram of the invention；
Fig. 2 is the distributed contract net collaborative planning configuration diagram of the present invention from bottom to top；
Fig. 3 is the flow diagram of collaborative framework mission planning from bottom to top of the invention；
Fig. 4 is earth observation systems contract net bidding and tendering process schematic diagram of the present invention；
Fig. 5 is that the present invention takes turns incomplete combined task allocation flow schematic diagram more, wherein (a) is according to resource
Voronoi diagram divides extensive task, (b) for partial task be based on contract net be assigned to can communication proximity resource, (c) for successively
The planning center that remaining task is distributed to planning central interior resource and can communicated；
Fig. 6 is distinct methods task completion rate Comparative result schematic diagram of the present invention；
Fig. 7 is that distinct methods of the present invention solve time Comparative result schematic diagram；
Fig. 8 is that the method for the present invention solves the time by the decomposition diagram of three levels, wherein (a) is resource coordinating layer algorithm
Runing time (b) cooperates with layer Riming time of algorithm for planning central interior, (c) cooperates with layer Riming time of algorithm for planning center；
Fig. 9 is distinct methods programme quality versus's schematic diagram of the present invention；
Figure 10 is the task performance Comparative result schematic diagram of distinct methods dynamic replanning of the present invention；
Figure 11 is the solution time Comparative result schematic diagram of distinct methods dynamic replanning of the present invention；
Figure 12 is the scheme rate of change Comparative result schematic diagram of distinct methods dynamic replanning of the present invention.
Specific embodiment
It is described in detail below in conjunction with specific embodiment of the attached drawing to the embodiment of the present invention.It should be understood that this
Locate described specific embodiment and be merely to illustrate and explain the present invention embodiment, is not intended to restrict the invention embodiment.
Earth observation resource dynamic programming method proposed by the present invention based on Contract Net Mechanism is specifically described as follows:
1), distributed contract net collaborative planning framework from bottom to top and process
A, distributed collaboration from bottom to top plans framework
The collaborative planning frame of empty world isomery observation resource has to meet the needs of three aspects.First, it is necessary to meet
The management structure and communication mechanism of existing all kinds of observation resources.Second, frame must have good extended capability, to ensure to provide
Source dynamically can be accessed and release, and realize that the real-time insertion of high-volume observation mission is changed with demand.Third, in order to reach
The target of Dynamic Programming, frame must have higher planning efficiency.The present invention fully considers the above principle, proposes from bottom to top
Distributed contract net co-architecture.Compared to top-down frame, this framework more emphasizes personal resource to uncertain environment
Dynamic response, focus more on Self-organizing Coordinated strategy of the personal resource under role pressure reaction, be a kind of from individual part
The allocation strategy of interaction response promotion global optimum.
As shown in Fig. 2, this framework is divided into four levels from bottom to top.Task management layer is responsible for receiving times that user submits
Business, and extensive task is divided into different task-sets according to resource observing capacity.And then task-set is distributed into next layer
Secondary respective resources.Resource coordinating layer includes numerous heterogeneous resource, observation resource receive task-set to be done and with need
Task of weight-normality being wanted to draw is distributed to through consultation together can be with the neighborhood resource set of this source communications.On this basis, resource section
Unassigned task is committed to next level by point.Plan that central interior collaboration layer is negotiated according to resource observing capacity in inside
On the basis of the internal resource of being administered unallocated task assignment to this planning center.Unfinished task then continues to be committed to down
One planning center cooperates with layer.Finally, legacy tasks are distributed to and can be led to according to the observing capacity at adjacent planning center by planning center
The neighborhood of letter plans center.This frame does not change the operational mode of existing resource.In order to reach the target quickly planned, using from
Resource gradually distributes task to this level from bottom to top in planning center, preferentially assigns the task to neighborhood observation resource, then
Task is gradually diffused by the task assignment procedures taken turns from bottom to top more other planning centers until task can complete be
Only.In addition, dynamic and scalability in order to realize collaborative planning system, this frame sees contract net protocol mechanism and the empty world
The mission planning process for surveying resource blends, and CNP (Contract Net Protocol, contract net protocol) can be each round
Task assignment procedure provides programme.
B, the collaborative planning process based on contract net
As a kind of communication interaction agreement of high level, CNP supports the cooperation and competition of extensive resource in dynamic scene.
The active that the centralized distribution of task can be changed to resource by " bidding " mechanism that CNP uses for reference in economic behaviour is submitted a tender.Using each
The distributed computation ability of observation resource can greatly improve the efficiency of task distribution, meanwhile, active Bid Mechanism is conducive to task
The resource dynamic expansion of planning system.
Collaborative planning process is considered as communication process of the task between resource.In distributed contract net association from bottom to top
With in framework, the resource of numerous low levels is administered as high level manager in planning center.Each resource can be seen
Make the different bidder of ability, with the progress of task assignment procedure, their role converts between bidder and issue of bidding documents person.
Only can just it be in communication with each other between planning center or resource with communication path.Its process is as shown in figure 3, when certain observation resource
When monitoring the task of dynamic insertion in need, point 5 stages are gradually completing mission planning scheme really based on contract net
It is vertical:
Phase 1: resource riReceive the task sequence T=(t that need to be inserted into1,t2,…tn), when j ∈ n is instructed, according to certainly
Body caching task to be done and T actively discriminate whether to have the ability to complete T；
Phase 2: if riTask T cannot be completed, then riRole switch to issue of bidding documents person, the resource that can be communicated to itself
Group RSi-LIt issues information on bidding (Fig. 3 a)；According to RSi-LReturn scale value and complete information judgement can complete task T；
Phase 3: if all resources cannot all complete T, riMission bit stream is fed back in the subplan of upper level
Heart Pk。PkResource issue of bidding documents as from issue of bidding documents person to planning central interior, riRole bidder (Fig. 3 b) is switched to by issue of bidding documents person；Pk
It judges whether there is resource according to return scale value and completion task subset and can complete T；
Phase 4: if completing T, P without resourcekCenter is planned to the neighborhood that can be communicated as issue of bidding documents person again
Set PSk-LIssue bidding documents；Each planning center [P1,P2,…Pk-1,Pk+1…Ppn] respectively specified that and can be completed according to internal resource
Task subset and feed back；PkFinal task allocation plan is evaluated according to feedback result；
Phase 5: if PkNeighborhood planning center cannot complete T then according to Neighbourhood set PSk-LCommunication proximity continue
Expand issue of bidding documents range, until can complete this task T or without remaining planning center.
Each stage in planning process determines Resource Allocation Formula using Contract Net Mechanism.Empty world earth observation is closed
There are three classes participation role: issue of bidding documents person, bidder and winning bidder with net mechanism.As shown in figure 4, the operation bidding and tendering process of contract net
Following four step can be divided into: bidding documents publication, resource are submitted a tender, victory mark is preferred, contract is signed and executed.
In bidding documents issuing process, task promoter is public as the potential neighborhood resource publication bid that bid direction can communicate
It accuses.In empty world earth observation contract net, the triggering of task publication is mainly derived from two aspects: since resource is by outside
The reasons such as interference or self-ability deficiency cause this resource to have no ability to continue to execute the task in reservation task buffer pool；Currently
Resource itself monitors that needs monitor but has no ability to the new task completed in itself.The message format of the call for tender is as follows:
TaskDocument=< ContractID, ContractType, TaskInfo, TaskRequirement, TaskGrade,
ExpireTime, QuoteRequirement >, wherein ContractID, ContractType respectively indicate unique knowledge of contract
Other ID and type.TaskInfo is spatial position where the detailed description to observation mission, including task, executes time, task
Title etc..TaskRequirement is extra demand and constraint, including spatial resolution, spectral band of task etc..
The important level of TaskGrade expression task.ExpireTime is deadline for submission of tenders.QuoteRequirement is bid amounts
Range.
In the bidding period, empty world resource can be according to contractual requirement peace treaty after receiving the call for tender as candidate bidder
Execution income, cost and the influence to existing surveillance program of this task of beam Conditions Evaluation.Then bid task and quotation are determined
And tender document is fed back to tenderer at the appointed time.Resource must satisfy resource current location, pending in tendering process
The constraint conditions such as task sequence, remaining cruising ability.So the tendering process of resource is considered as constraint satisfaction problemx
(CSP).Tender documents format are as follows: BidDoucument=< ContractID, Bid, ExecutionScheme, BidPrice,
TaskSequences,IndicatorsStatus>.Wherein, Bid is a Boolean variable for indicating whether to submit a tender.
ExecutionScheme indicates that bidder carries into execution a plan to the task of completion, including the information such as deadline, resolution ratio, wave band.
BidPrice indicates the bid value of feedback.TaskSequences is current bid person by all set of tasks of execution and its
Execute sequence.Can IndicatorsStatus be the state description that complete to all task index and constraint condition.
The mark stage is being selected, after having successfully received more parts of tender documents, tenderer is by contract optimization algorithm from all biddings documents
Filter out it is optimal carry into execution a plan, and contract is authorized to selected bidder.It is true that the selection of optimal case is considered as victor
Determine problem (WDP).
In tasks execution phases, tenderer and winning bidder complete Contract Signing, and by successfully signing message informing to all
Bidder.Signing task is inserted into task buffer pool by the resource that success is contracted, and is adjusted according to itself task execution strategy dynamic
The execution order of task in whole buffer pool.Other, which are not received, awards the bidder of about information and then continues to execute its existing task sequence
Column.
2), the incomplete combination distribution method of more wheels towards extensive task
During empty world resource execution task, once emergency event occurs, it will while triggering a large amount of observation mission.
The allocative efficiency of these tasks is the key that task execution success or not.Connected based on traditional improved single task of Contract-Net Model
Continuous allocation plan is difficult meet the needs of high-timeliness in Dynamic Programming.Therefore, the present invention is on the basis of frame from bottom to top
It is proposed more incomplete combined task distribution methods (MICA) of wheel, the method by mostly wheel matching can rapidly mass distributed it is a large amount of
Task supports multitask, the bid of more winning bidders, more bouts.Wherein, the determination of each round bid-winning scheme is that task is quickly distributed
Core.For this purpose, the present invention, which is regarded as victor, determines problem, and propose that one kind selects target local search algorithm based on floating
(FLS) the choosing mark of every wheel is solved the problems, such as.The characteristics of distribution of FLS algorithm combination task and resource are submitted a tender, using taboo and preferentially
Strategy improves optimal solution convergence rate.Furthermore select mark mechanism that algorithm is avoided to fall into local optimum using probability parameter and floating.
A, take turns incomplete combination distribution method more
According to frame from bottom to top, the present invention proposes that incomplete combined task distribution (MICA) method of more wheels is come rapidly
Match large-scale concurrent task.MICA method improves the allocative efficiency of task: task combination distribution, more using three kinds of strategies
Business collection is synchronous to be distributed, is multi-level resource matched.As shown in figure 5, solving difficult and efficiency to solve extensive task bring
Low problem, this method construct resource Voronoi diagram according to resource current location and its observing capacity first.Then by task
Multiple small set of tasks (Fig. 5 a) are divided into according to this Voronoi diagram and are synchronized to distribute to different resources.Due to resource
With distributed computation ability, so the task in multiple tasks set can distribute parallel.Secondly, according to collaborative planning frame
Frame, MICA method are successively allocated task-set on three levels from bottom to top respectively.In first level, according to conjunction
With network method respectively by each resource as bid its communication proximity resource set of direction publication bidding documents (Fig. 5 b).Pass through resource negotiation
Determine combined task allocation plan.In second level, each planning center receives the task of not completing respectively, and in specified planning
Interior portion resource completes task (Fig. 5 c).In third level, all planning centers joint consultation determines abortive point
With scheme (Fig. 5 c).Shown in its detailed algorithm is described as follows, and referring to table 1-2.
If calling for bid quotient Tenderer, bid resource set Bidder；Planning center collection P=[P1,P2,…,Ppn],k∈pn；Pk
Manage resource set Rk=(r1,r2,…rm),i∈m；PSk-LFor Pk, can communicate neighborhood planning center collection；Resource riIt receives wait advise
Draw task-set T=(t1,t2,…tn),j∈n；RSi-LFor resource riThe neighborhood resource set that can be communicated；B=< VR-T,GT> indicate
Bidder returns to the bid set of Tenderer, VR-TIndicate Bidder to set of tasks GTOffer by tender set, GTIt indicates
Bid set of tasks,Expression schemeIn include set of tasks T in task.
In the first level, if resource riThe neighborhood resource set RS that can be communicatedi-L≠ φ, then by resource riFor set of tasks
T is preferentially to RSi-LIssue the call for tender, i.e. Tenderer=ri, Bidder=RSi-L.It is logical to receive bid for each resource in Bidder
Know, the set of tasks G completed is determined to according to task to be done and own resource loading conditionTWith offer values VR-T, and return
Submit a tender set B=< VR-T,GT>.Then, resource riOptimal bidding plans set is selected according to choosing mark algorithm (Algorithm 3)Due to the limitation of resource capability, optimal case tends not to complete whole tasks of initiating task set, i.e.,Enter the second level at this time, as preplanning center PkThe specified all resource R that can be managedkComplete remaining
Business setIf all tasks in task-set T cannot be completed, bid range is expanded to neighborhood planning center collection
Close PSk-L.This is third level.After each planning center arranges task observation scheme, bid result is fed back according to algorithm 2.If still
So there is unfinished task, then continues at the neighborhood that can be communicated planning center as Tenderer to expand bid range, until complete
At task or using until all resources.Finally according to optimal case CbestComplete the signing of respective resource and task.
1 MICA algorithm frame of table
Planning central task allocation algorithm frame of the table 2 based on CNP
B, based on the local search algorithm for selecting mark mechanism of floating
Victor's decision problem of empty world resource is a kind of combinatorial optimization problem.In the negotiation distribution of each round of contract net
In the process, all bidder will return to set of submitting a tender to tenderer.If task-set T=(t to be allocated1,t2,…tn),j∈n；
The bid item B=(B of all resources1,B2,…Bi…Bm) constitute Candidate Set CanB；V=(V1,V2,…Vm),i∈m,ViIt indicates to throw
Mark item BiOffer by tender set；G=(G1,G2,…Gm),i∈m,GiIndicate bid item BiBid set of tasks.If one
Bid item BiBeing selected becomes winning bidder, then BiOne of composition item as solution C.Victor's decision problem (WDP) is from Candidate Set
The sum of the offer by tender that a subset makes as feasible solution C in feasible solution maximum is selected in CanB.Solution can be by boolean set x table
Show, xi=1 indicates bid item BiIt is selected.If S is the binary matrix of a m*n, if task tj∈Gi, then Sij=1, otherwise
Sij=0.Thus objective function can indicate are as follows:
Its constraint condition are as follows:
xi∈{0,1} (2)
Constraining (2) indicates that bid item is only selected and without selected two states；Constraining (3) indicates each task only
Can be selected primary, i.e., do not conflict mutually between selected task.
Definition conflict bid: for two bidder bdiAnd bdkBid set of tasks GiAnd Gk, If at least one task is present in two set, i.e. { Gi∩Gk≠ φ }, then claim BiAnd BkConflict each other
It submits a tender, GiAnd GkConflict bid task each other.Otherwise claim two bids compatible.Conflict bid can be constructed according to conflict relationship two-by-two
Matrix Mcon(symmetrical matrix).
For the WDP problem in empty world resource Contract Net Mechanism, the present invention propose it is a kind of based on float select mark mechanism
Local search algorithm (FLS), algorithm frame is shown in Table 3.The algorithm is swum at random using probability parameter and the choosing mark mechanism control that floats
The diversity for walking to enhance solution, greatly improves the accuracy of solution.It is prevented using taboo and preference strategy to candidate disaggregation space
Repeated retrieval, improve optimal solution convergence rate.
FLS algorithm is the process of a successive ignition and Stepwise optimization.The number of iterations y can taking human as determine or find optimal
Until solution.In search process, if each iteration of algorithm all searches for all candidate solution set of submitting a tender, slow convergence speed will certainly be dragged
Degree.In fact, the first search weight of candidate disaggregation is different.So in order to accelerate search speed, algorithm is devised preferentially
Search bid set QBWith tabu search bid set HB。QBIt is the bid set compatible with current optimal solution set, HBIt is and candidate
The bid set of optimal solution set C conflict.To improve solution efficiency, start in each iteration, algorithm first search emphasis, which is submitted a tender, to be collected
Close QB, exclude taboo bid set HB.In each iteration finally, algorithm is according to conflict bid matrix MconSuccessively update QBAnd HB。
Continuous searching process is easy that algorithm is made to fall into local optimum.In order to avoid this problem and reduce solving complexity,
This algorithm devises a probability parameter ρ and income fluctuation area σ.Algorithm executes greedy search with probability ρ, is held with probability 1- ρ
Row random walk.As shown in formula (4), according to optimal solution C and introduce taboo list HBCandidate disaggregation CanB can be obtained.Algorithm or with
Probability ρ determines maximum offer by tender value V from candidate tasks collection CanBmaxBid BcanOptimal solution C or probability 1- ρ is added
Select scale value B at random from B-CcanOptimal solution C is added.However, only choosing has the bid of maximum quotation excessively greedy in algorithm.
This still is not enough to enhance the diversity of solution.So algorithm is selected from Candidate Set CanB offers with maximum as shown in formula (5)
Value VmaxDiffer the bid set F in fluctuation area σB.And from float set FBOne bid B of middle random selectioncanIt is added optimal
Solve C.In this way, algorithm can jump out locally optimal solution as much as possible.Finally, algorithm is defeated by the optimal solution C for updating each iteration
Globally optimal solution C outbest。
CanB=B-C-HB (4)
FB={ Bi||Vmax-Vi|≤σ},Bi∈CanB (5)
Table 3 selects the local search algorithm frame of mark mechanism based on floating
By the in-depth analysis to the world empty under dynamic environment resource planning problem, we given up it is traditional from top to bottom
Planning framework, innovatively propose a kind of distributed collaboration planning framework from bottom to top, and devise from bottom frame to
A whole set of of upper layer algorithm copes with the solution of empty world resource dynamic programming problems.This framework from bottom to top more accords with
Close the dynamic and uncertain feature in planning process.In addition, the present invention is by task assignment procedure and contract net protocol depth integration,
Give full play to the distributed computation ability of all observation resource nodes.
Aiming at the problem that the quick weight-normality of large-scale concurrent task is drawn, the present invention proposes a kind of more wheels towards dynamic environment not
Complete combination method for allocating tasks.We improve task allocative efficiency using three kinds of strategies.Firstly, extensive task is divided
At multiple tasks set to achieve the purpose that the synchronous distribution of multiple tasks set.Secondly, suitably bundling set according to resource capability
In multiple tasks combine distribution, thus compared to continuous single task distribute it is more efficient.Finally, according to collaboration from bottom to top
Frame, task are successively assigned to resource into the system at planning center, this is equally to decompose extensive task to improve distribution
The means of efficiency.
The present invention proposes that a kind of local search algorithm determines to solve contract net victory mark during heterogeneous resource collaborative planning
The problem of.Algorithm, to enhance the diversity of solution, is greatly improved using probability parameter and choosing mark mechanism control random walk of floating
The accuracy of solution.The repeated retrieval to candidate disaggregation space is prevented using taboo and preference strategy, improves optimal solution convergence rate.
It is prominent in extensive task in order to test distributed contract net co-architecture and MICA method from bottom to top of the invention
Collaborative planning performance under heat condition and dynamic environment, the present invention are provided with two groups of comparative tests.When urgent thing is answered on earthquake, landslide etc.
After part occurs, the case where in order to obtain devastated, the observation side according to contingency tasks rapid development earth observation resource is needed
Case.According to the feature of emergency event, concurrent and uncertain environment the Dynamic Programming that continues of extensive task is rapid development rule
Two main problems that the scheme of drawing is faced.So first group of experiment of the present invention is concurrent in extensive contingency tasks for verifying
In the case where, MICA method can effectively rapid development resource observation program.Second group of experiment is used to verify dynamic not true
Determine in environment, can MICA method quickly make reasonable observation program for original programme.
1), under extensive task complications different planing methods contrast verification
In this comparative experiments, we are provided with one and include the imitative of the different observation resource of satellite, unmanned plane, dirigible three classes
True scene.Three classes are observed resource and managed as a whole respectively by four planning centers: 1 satellite plans 2 earth observations of centre management
Satellite；2 unmanned plane planning centers manage 25,28 unmanned planes respectively；1 dirigible plans 9 dirigibles of centre management.In order to test
Can MICA method of the present invention be demonstrate,proved cope with the pressure of extensive task, our tasks provided with 600 spatially random distributions
(in order to eliminate experiment accidental error, we generate 5 groups of same amount of data in same space at random).The present invention demonstrates
The solution formulation effect of collaborative planning method of the task quantity from 1 to 600 under all situations, and with four kinds of common collaborative plannings
Method compares.First method is the continuous auction system of individual event (SSA) based on contract net, in this method one it is virtual
Planning center manage all resources as issue of bidding documents person, and all resources feed back tender offer and observation side as bidder
Case.Individual task is successively auctioned at this virtual planning center and the highest resource that selects to bid is as winning bidder, and this task is divided
This acceptance of the bid resource of dispensing.Virtual planning centre wheel abortive lot sells all tasks until all tasks are all assigned or cannot be observed
Only.Second method is to be sequentially allocated task (AUS) according to the sequence of dirigible, unmanned plane, satellite, has article to point out such sequence
Highest task completion rate (Wu, 2016) can be obtained in single task continuous dispensing method.And in same category resource, this
Method preferentially assigns the task to the observation biggish resource of income, and (observation income is directly proportional to task weight, with resource to task
Distance be inversely proportional).SSA and AUS method is all the method to the distribution of the iteration of individual event task, they all do not account for resource it
Between cooperation with cooperate with.And method three (MCP) and method four (BCP) they are all the collaborative planning methods of centralization, they are from the overall situation
Visual angle consider observation resource how to be performed in unison with task.Method three select Mosek optimization tool as solver solve with
The upper Task Assignment Model, it can select suitable optimization algorithm to obtain more accurate programme.Method four is selected
It selects branch-bound algorithm (Branch and bound) and solves problem above, BnB algorithm has higher optimization efficiency, so can
Comparatively quickly obtain allocation result.Integer programming model shown in formula (1) is the abstract modeling to collaborative planning problem,
For MCP and BCP method all using it as Task Assignment Model, optimization aim is the maximization of global gain.In order to verify the present invention
Performance of the method under extensive role pressure, we compared three important indicators of five kinds of methods respectively: task completion rate,
Method operational efficiency and programme reasonability.As shown in formula (6), using the average energy consumption of observation mission as evaluation side
The rational foundation of case, smaller then this scheme of average energy consumption are more reasonable.
In formula, Si, k represent the resource ri in the mission planning center Pk for completing to be assigned and need mobile distance total
With.Ni, k indicate the task quantity that the resource ri in the Pk of planning center can be completed.
Experimental result is as shown in Figure 6, Figure 7, MCP and BCP method is all the collaborative planning method of centralization.Due to considering
Cooperation between resource with cooperate with, they can complete more tasks than the method for individual event task continuous dispensing.Wherein, due to
Advantage of the Mosek optimizer in terms of solving accuracy, the MCP method based on global collaborative have reached highest task and have completed
Rate, and keep in available result 100% completion rate (since calculator memory overflow error cannot test all appoint
Business amount).However, the calculation of this centralization causes the computational efficiency of MCP and BCP method extremely inefficient, so that being difficult
Cope with the emergency case of extensive task.Solution time being continuously increased in explosive growth with task of MCP method, it is in office
When business amount reaches 400 or so, even tens hours of needs time are solved, computer is caused to bear not living memory pressure.This
Also having confirmed Mosek solver can obtain relatively that more preferably solution but has that computational efficiency is low.According to experimental result,
It is reasonable that relative to the very high completion rate of MCP method, its solution efficiency is more difficult to receive in the case where emergency management and rescue.
Further, since the characteristics of branch and bound method fast search, BCP method ratio MCP method possesses higher solution efficiency (Fig. 7),
But the curve fluctuation of task completion rate is as can be seen that BCP method has certain randomness in terms of solving accuracy from Fig. 6
And hardly result in optimal solution.
Since SSA and AUS method is all the iteration assigning process to individual task, so in task completion rate and calculating effect
Rate etc. has very strong similitude.Both methods only has an iteration process to all tasks, and iteration is only divided every time
With a task, so its computational efficiency is significantly higher than MCP and BCP method (Fig. 7), elapsed time and task quantity are being calculated by force just
Correlation, i.e. Tmethod∝Ntask, wherein TmethodRepresentation method solves spent time, NtaskIndicate the task quantity for needing to plan.
However, the task completion rate index of SSA and AUS method is obviously in a disadvantageous position (Fig. 6) due to the synergistic mechanism between being deficient in resources.
SSA method selects optimal solution according to the highest bid of all resources, and AUS method only considers the income of single classification resource, institute
In a little higher than AUS method of SSA method task completion rate, but the operation efficiency of AUS method is more excellent.
Comprehensive all planing methods it can be found that the completion rate and solution efficiency of programme can not often get both, and this
Method achieves balance between the two.The task completion rate of MICA method and MCP method are very close, or even big at 500
The task completion rate for still having reached 95% or so under the pressure of scale task is significantly better than other three kinds of planing methods (Fig. 6).?
In MICA method, when task quantity reaches 318 or so, resource capability gradually tends to be saturated, and task completion rate starts slowly to decline
(Fig. 6), and Riming time of algorithm also switchs to rapidly propradation (Fig. 7) by steadily increasing.When in order to probe into the operation of MICA method
Between zooming internal mechanism, we by the runing time of MICA method according to distributed contract net co-architecture from bottom to top
Three levels decomposed (Fig. 8).In resource coordinating layer, task is assigned to the neighborhood resource that can be communicated, by using three
Kind optimisation strategy and choosing mark algorithm make level allocative efficiency with higher.And with the increase of task amount, resource is gradually
Expand bid range, becomes larger so as to cause the communication number scale in contract net bidding and tendering process, and then cause operational efficiency fast
Speed decline.When the observing capacity of neighborhood resource reaches saturation, unfinished task is passed into planning central interior collaboration layer (figure
8(b)).In this level, plan that internally resource calls for bid at center, as shown in Fig. 8 (b), task quantity is between 250 to 350
When, data point only has small part task than sparse explanation and needs to be allocated in the second level in figure.And with task quantity
Increase, figure midpoint more and more intensively illustrates that more tasks need to be allocated in the second layer even third layer.So as to cause
The increase of time is solved, but is still below first layer.Cooperate with layer at planning center, planning center by the unfinished task of residue to its
He plans that center is called for bid, since most of task is allocated at first two layers, so the solution efficiency highest of third layer.It is overall
From the point of view of, due to using distributed computing and local search algorithm, MICA method is significantly better than it in terms of algorithm solution efficiency
His control methods (Fig. 7).
Programme reasonability index can reflect the quality of programme, and numerical value is smaller to illustrate that completing individual task puts down
The energy consumed is smaller.The rational result of scheme has very high correlation with task completion rate, and task completion rate gets over Gao Ze
The quality of programme is relatively higher (Fig. 9).In addition, resource capability obtains more rationally and tight with the increase of task quantity
The utilization gathered, the average stage length for executing task have the tendency that becoming smaller.
2), the quick weight-normality of different planing methods draws contrast verification under dynamic environment
We devise one group of experiment to verify the effect of MICA method dynamic replanning.It is seen over the ground when resource is carrying out
When survey task, the contingency tasks that need to be observed occur suddenly, this often means that the deficiency of resource observing capacity, needs to comprehensively consider
New task place and original programme determine whether to abandon some original tasks and formulate new programme.In order to verify
Dynamic Programming effect of the MICA method under resource capability saturated conditions, we are provided with the number of tasks greater than resource observing capacity
Amount.Simulating scenes are comprising 1 satellite, 18 sortie unmanned planes and 3 dirigibles and respectively by 4 planning centre managements.They are initial
It is required to execute the task of 40 spatially random distributions.Observe resource be carrying out task during can it is continuous 6 take turns with
Dynamically there is 30-50 emergency observation mission in machine, and observation resource need to be according to different planning algorithms on the basis of original scheme
Constantly formulate new observation program.MCP, BCP and AUS method all will first be not carried out task and recycles and divide together with new addition task
It is fitted on all resources.SSA and MICA method all calls for bid using new task as candidate tasks towards all resources, resource according to
Current location and remaining observing capacity feed back financial value, and determine to observe which new task and abandon which original task.
During dynamic replanning, task completion rate is not the index being uniquely concerned about.The process of resource execution task
In environment in lasting dynamic change, time and observation program rate of change (times to change in original scheme that weight-normality is drawn
The ratio for the quantity and all task quantity of original scheme of being engaged in) it is the key factor for determining weight planing method superiority and inferiority.Weight-normality draw when
Between it is shorter, resource is waiting the energy spent by interval also fewer, and can acquisition earlier monitor region remotely-sensed data.This
Outside, on the basis of guaranteeing to observe income, the rate of change of existing scheme is smaller, and the movement adjustment for observing resource is also fewer, rule
The scheme of drawing is also more excellent.Therefore, the present invention compared task completion rate, weight-normality stroke of the distinct methods in lasting insertion new task
Time and scheme rate of change index.
According to experimental result as shown in table 4, during multiple dynamic insertion new task, different planing method institutes are complete
At task quantity can all continue to increase, but task completion rate all has a declining tendency.Similar, the MCP method with experiment (1) result
Task completion rate be generally better than other methods, and the method (AUS and SSA method) based on individual event task continuous dispensing is in task
Still in disadvantage (Figure 10) in terms of completion rate.However, task completion rate is not dynamic programming process sole indicator of interest.It is dynamic
Program scheme that not only will be as fast as possible in state uncertain environment, will also be guaranteeing maintain as far as possible on the basis of observing income
Original programme.As shown in figure 11, the heavy planning time of distinct methods all can draw the increase of number with weight-normality and rise, and
With task total amount correlation, wherein the computational efficiency of MCP method is significantly less than other planing methods.And divide due to using
Cloth calculation simultaneously optimizes choosing mark algorithm, and MICA method has a clear superiority in terms of solution efficiency, and continuous 6 wheel dynamic is advised
The algorithm elapsed time drawn is in steady ascendant trend, and maintains reduced levels always.
Scheme rate of change can reflect the influence caused by original programme of newly added task, different planing methods
Influence size to initial planning scheme is also different.As shown in figure 12, scheme rate of change can with the increase of Dynamic Programming number and
Gradually decrease, in fact, this is because the radix scale of task is excessive in a upper scheme and the quantity of new task causes very little
's.Therefore, we are indirect with task sequential growth rate (ratio of total task number amount in new addition task quantity and a upper scheme)
Ground reflects influence of the task radix to scheme rate of change.If the mark point of scheme rate of change is in the inside of histogram, table
Show that the method changes the task less than new task quantity.Conversely, the method changes the task more than new task quantity.Example
Such as, the 3rd wheel Dynamic Programming in, the scheme rate of change mark point of most of planing method all in histogram outside, indicate this
The new observation program of a little methods all changes 33 or more in original scheme tasks.The scheme rate of change of MICA method of the present invention
It remains lower than task sequential growth rate, and is better than other planing methods.So regardless of whether ignore task radix influence,
The scheme rate of change of MICA method is all more excellent than other methods, is more in line with the demand of dynamic replanning.
The dynamic replanning Comparative result of 4 distinct methods of table
Note: NT indicates the task quantity that kainogenesis need to be observed；AT indicates all task quantity.TCR indicates task completion rate
Index；RPT indicates that weight-normality draws time index；RSC indicates original scheme rate of change index.
The optional embodiment of the embodiment of the present invention is described in detail in conjunction with attached drawing above, still, the embodiment of the present invention is simultaneously
The detail being not limited in above embodiment can be to of the invention real in the range of the technology design of the embodiment of the present invention
The technical solution for applying example carries out a variety of simple variants, these simple variants belong to the protection scope of the embodiment of the present invention.
It is further to note that specific technical features described in the above specific embodiments, in not lance
In the case where shield, it can be combined in any appropriate way.In order to avoid unnecessary repetition, the embodiment of the present invention pair
No further explanation will be given for various combinations of possible ways.
In addition, any combination can also be carried out between a variety of different embodiments of the embodiment of the present invention, as long as it is not
The thought of the embodiment of the present invention is violated, equally should be considered as disclosure of that of the embodiment of the present invention.
Claims (6)
1. a kind of earth observation resource dynamic programming method based on Contract Net Mechanism, which is characterized in that use from bottom to top
Distributed collaboration plans framework and the collaborative planning process based on contract net；By taking turns incomplete combination distribution method to big rule more
Mould concurrent tasks are dynamically distributed.
2. the earth observation resource dynamic programming method according to claim 1 based on Contract Net Mechanism, which is characterized in that
The distributed collaboration planning framework from bottom to top includes following four level:
Task management layer is responsible for receiving the task that user submits, and extensive task is divided into not according to resource observing capacity
With task-set, and then task-set is distributed to the respective resources of next level；
Resource coordinating layer observes resource and receives task-set to be done and draw with weight-normality is needed including numerous heterogeneous resources
Task distribute to through consultation together can be with the neighborhood resource set of this source communications；On this basis, resource node will not by
The task of distribution is committed to next level；
It plans that central interior cooperates with layer, gives unallocated task assignment to this on the basis of inside is negotiated according to resource observing capacity
The internal resource that planning center is administered；
Planning center cooperates with layer, and planning center then continues the unfinished task of residue to be committed to the bid of other planning centers.
3. the earth observation resource dynamic programming method according to claim 2 based on Contract Net Mechanism, which is characterized in that
The collaborative planning process based on contract net carries out task distribution with the active bidding pattern of earth observation resource.
4. the earth observation resource dynamic programming method according to claim 3 based on Contract Net Mechanism, which is characterized in that
The active bidding pattern includes following five stages:
Phase 1: resource riReceive the task sequence T=(t that need to be inserted into1,t2,…tn), when j ∈ n is instructed, waited for according to itself
At caching task and T actively discriminate whether have the ability complete T；
Phase 2: if riTask T cannot be completed, then riRole switch to issue of bidding documents person, the group of resources that can be communicated to itself
RSi-LIssue information on bidding；According to RSi-LReturn scale value and complete information judgement can complete task T；
Phase 3: if all resources cannot all complete T, riMission bit stream is fed back to the subplan center P of upper levelk；
PkResource issue of bidding documents as from issue of bidding documents person to planning central interior, riRole bidder is switched to by issue of bidding documents person；PkAccording to return scale value
T can be completed by judging whether there is resource with completion task subset；
Phase 4: if completing T, P without resourcekCentralization is planned to the neighborhood that can be communicated as issue of bidding documents person again
PSk-LIssue bidding documents；Each planning center [P1,P2,…Pk-1,Pk+1…Ppn] respectively specify that can be completed appoints according to internal resource
Business subset is simultaneously fed back；PkFinal task allocation plan is evaluated according to feedback result；
Phase 5: if PkNeighborhood planning center cannot complete T then according to Neighbourhood set PSk-LCommunication proximity continue to expand
Issue of bidding documents range, until can complete this task T or without remaining planning center.
5. the earth observation resource dynamic programming method according to claim 4 based on Contract Net Mechanism, which is characterized in that
More incomplete combination distribution method of wheel is successively allocated task-set on three levels from bottom to top respectively:
In first level, based on contract network method is respectively by each resource as bid its communication proximity resource set of direction publication
Bidding documents determines combined task allocation plan by resource negotiation；
In second level, each planning center receives the task of not completing respectively, and specified planning central interior resource is completed to appoint
Business；
In third level, all planning centers joint consultation determines abortive allocation plan.
6. the earth observation resource dynamic programming method according to claim 1-5 based on Contract Net Mechanism,
It is characterized in that, solves contract net during heterogeneous resource collaborative planning using the local search algorithm of mark mechanism is selected based on floating
The determining problem of victory mark.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811351029.9A CN109409773B (en) | 2018-11-14 | 2018-11-14 | Dynamic planning method for earth observation resources based on contract network mechanism |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811351029.9A CN109409773B (en) | 2018-11-14 | 2018-11-14 | Dynamic planning method for earth observation resources based on contract network mechanism |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109409773A true CN109409773A (en) | 2019-03-01 |
| CN109409773B CN109409773B (en) | 2022-04-08 |
Family
ID=65472973
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811351029.9A Active CN109409773B (en) | 2018-11-14 | 2018-11-14 | Dynamic planning method for earth observation resources based on contract network mechanism |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109409773B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111124637A (en) * | 2019-12-09 | 2020-05-08 | 中国空间技术研究院 | Dynamic task planning method for distributed satellite system |
| CN112101747A (en) * | 2020-08-28 | 2020-12-18 | 上海交通大学 | Unmanned ship interception task allocation method based on tabu consensus auction algorithm |
| CN112181658A (en) * | 2020-09-30 | 2021-01-05 | 南京工程学院 | Computing task allocation method for maximizing network benefits in heterogeneous network |
| CN112991745A (en) * | 2021-04-30 | 2021-06-18 | 中南大学 | Traffic flow dynamic cooperative allocation method under distributed framework |
| CN113240324A (en) * | 2021-06-02 | 2021-08-10 | 中国电子科技集团公司第五十四研究所 | Air and space resource collaborative planning method considering communication mechanism |
| CN113313414A (en) * | 2021-06-21 | 2021-08-27 | 哈尔滨工程大学 | Task collaborative planning method for multi-class heterogeneous remote sensing platform |
| CN113487220A (en) * | 2021-07-23 | 2021-10-08 | 中南大学 | Static target observation-oriented space-sky heterogeneous earth observation resource cooperative scheduling method |
| CN113537782A (en) * | 2021-07-19 | 2021-10-22 | 福州大学 | Contract network-based multi-satellite situation awareness system distributed task planning method |
| CN115616902A (en) * | 2022-11-07 | 2023-01-17 | 中国人民解放军国防科技大学 | Cluster spacecraft task allocation method and device based on robust auction algorithm |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102903028A (en) * | 2012-09-28 | 2013-01-30 | 中国人民解放军国防科学技术大学 | Emergency-orientated co-cooperation allocation method for multi-airship earth observation tasks |
| CN107180309A (en) * | 2017-05-31 | 2017-09-19 | 中南大学 | The collaborative planning method of resource is observed in a kind of empty world |
| CN108256671A (en) * | 2017-12-26 | 2018-07-06 | 佛山科学技术学院 | A kind of more resources of multitask based on learning-oriented genetic algorithm roll distribution method |
| US10068242B1 (en) * | 2001-03-15 | 2018-09-04 | Versata Development Group, Inc. | Method and system for managing distributor information |
-
2018
- 2018-11-14 CN CN201811351029.9A patent/CN109409773B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10068242B1 (en) * | 2001-03-15 | 2018-09-04 | Versata Development Group, Inc. | Method and system for managing distributor information |
| CN102903028A (en) * | 2012-09-28 | 2013-01-30 | 中国人民解放军国防科学技术大学 | Emergency-orientated co-cooperation allocation method for multi-airship earth observation tasks |
| CN107180309A (en) * | 2017-05-31 | 2017-09-19 | 中南大学 | The collaborative planning method of resource is observed in a kind of empty world |
| CN108256671A (en) * | 2017-12-26 | 2018-07-06 | 佛山科学技术学院 | A kind of more resources of multitask based on learning-oriented genetic algorithm roll distribution method |
Non-Patent Citations (4)
| Title |
|---|
| JIANG WEI等: "multi-agent based dynamic mission planning for satellite data transmission", 《2016 12TH INTERNATIONAL CONFERENCE ON NATURAL COMPUTATION, FUZZY SYSTEMS AND KNOWLEDGE DISCOVERY(ICNC-FSKD)》 * |
| 伍国华: "基于勘探和开采策略控制的智能优化算法及其应用研究", 《中国博士学位论文全文数据库工程科技I辑》 * |
| 王慧林,伍国华等: "多类异构对地观测平台协同任务规划方法", 《航空学报》 * |
| 祝周鹏: "面向任务的卫星平台载荷配置与应急规划技术", 《中国优秀博硕士学位论文全文数据库(硕士)基础科学辑》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111124637B (en) * | 2019-12-09 | 2023-08-29 | 中国空间技术研究院 | Dynamic task planning method for distributed satellite system |
| CN111124637A (en) * | 2019-12-09 | 2020-05-08 | 中国空间技术研究院 | Dynamic task planning method for distributed satellite system |
| CN112101747A (en) * | 2020-08-28 | 2020-12-18 | 上海交通大学 | Unmanned ship interception task allocation method based on tabu consensus auction algorithm |
| CN112101747B (en) * | 2020-08-28 | 2023-11-03 | 上海交通大学 | Unmanned ship interception task allocation method based on tabu consensus auction algorithm |
| CN112181658A (en) * | 2020-09-30 | 2021-01-05 | 南京工程学院 | Computing task allocation method for maximizing network benefits in heterogeneous network |
| CN112181658B (en) * | 2020-09-30 | 2024-04-05 | 南京工程学院 | Calculation task allocation method for maximizing network benefits in heterogeneous network |
| CN112991745A (en) * | 2021-04-30 | 2021-06-18 | 中南大学 | Traffic flow dynamic cooperative allocation method under distributed framework |
| CN113240324A (en) * | 2021-06-02 | 2021-08-10 | 中国电子科技集团公司第五十四研究所 | Air and space resource collaborative planning method considering communication mechanism |
| CN113313414A (en) * | 2021-06-21 | 2021-08-27 | 哈尔滨工程大学 | Task collaborative planning method for multi-class heterogeneous remote sensing platform |
| CN113313414B (en) * | 2021-06-21 | 2024-05-14 | 哈尔滨工程大学 | Task collaborative planning method for multi-class heterogeneous remote sensing platform |
| CN113537782B (en) * | 2021-07-19 | 2023-08-18 | 福州大学 | Multi-satellite situation awareness system distributed task planning method based on contract network |
| CN113537782A (en) * | 2021-07-19 | 2021-10-22 | 福州大学 | Contract network-based multi-satellite situation awareness system distributed task planning method |
| CN113487220B (en) * | 2021-07-23 | 2022-07-15 | 中南大学 | Static target observation-oriented space-sky heterogeneous earth observation resource cooperative scheduling method |
| CN113487220A (en) * | 2021-07-23 | 2021-10-08 | 中南大学 | Static target observation-oriented space-sky heterogeneous earth observation resource cooperative scheduling method |
| CN115616902A (en) * | 2022-11-07 | 2023-01-17 | 中国人民解放军国防科技大学 | Cluster spacecraft task allocation method and device based on robust auction algorithm |
| CN115616902B (en) * | 2022-11-07 | 2023-03-07 | 中国人民解放军国防科技大学 | Cluster spacecraft task allocation method and device based on robust auction algorithm |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109409773B (en) | 2022-04-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109409773A (en) | A kind of earth observation resource dynamic programming method based on Contract Net Mechanism | |
| CN112016812B (en) | Multi-unmanned aerial vehicle task scheduling method, system and storage medium | |
| Ding et al. | Improved particle swarm optimization algorithm based novel encoding and decoding schemes for flexible job shop scheduling problem | |
| Abdel-Basset et al. | A comprehensive review of quadratic assignment problem: variants, hybrids and applications | |
| CN105069010B (en) | Resource polymerization method based on multi-Agent | |
| CN105159762B (en) | Heuristic cloud computing method for scheduling task based on Greedy strategy | |
| CN108880663A (en) | Incorporate network resource allocation method based on improved adaptive GA-IAGA | |
| CN112766813A (en) | Air-space cooperative observation complex task scheduling method and system | |
| CN108256671A (en) | A kind of more resources of multitask based on learning-oriented genetic algorithm roll distribution method | |
| Durgadevi et al. | Resource allocation in cloud computing using SFLA and cuckoo search hybridization | |
| Mahato et al. | On scheduling transaction in grid computing using cuckoo search-ant colony optimization considering load | |
| CN106155791A (en) | A kind of workflow task dispatching method under distributed environment | |
| CN111784211B (en) | Cluster-based group multitask allocation method and storage medium | |
| Mahato et al. | On scheduling transactions in a grid processing system considering load through ant colony optimization | |
| CN108989098A (en) | A kind of mixing scientific workflow data layout method of the cloud environment towards time delay optimization | |
| CN109343945A (en) | A kind of multitask dynamic allocation method based on contract net algorithm | |
| CN114326827B (en) | Unmanned aerial vehicle cluster multitasking dynamic allocation method and system | |
| CN108647771A (en) | The layout method of research-on-research flow data under a kind of mixing cloud environment | |
| CN109615143A (en) | Wide-area Measurement Information management system task schedule ant colony optimization algorithm based on multi-QoS constraint | |
| CN109740870A (en) | The resource dynamic dispatching method that Web is applied under cloud computing environment | |
| CN114461368A (en) | Multi-target cloud workflow scheduling method based on cooperative fruit fly algorithm | |
| CN104506576A (en) | Wireless sensor network and node task migration method thereof | |
| Zhang et al. | Multi-task deep reinforcement learning for scalable parallel task scheduling | |
| Mahato | Cuckoo search-ant colony optimization based scheduling in grid computing | |
| CN110070287A (en) | A kind of dynamic task allocation method based on similitude clustering and average thought |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |