CN109087023A - A kind of more stars observation layering dispatching method and system based on divide-and-conquer strategy - Google Patents
A kind of more stars observation layering dispatching method and system based on divide-and-conquer strategy Download PDFInfo
- Publication number
- CN109087023A CN109087023A CN201810967096.7A CN201810967096A CN109087023A CN 109087023 A CN109087023 A CN 109087023A CN 201810967096 A CN201810967096 A CN 201810967096A CN 109087023 A CN109087023 A CN 109087023A
- Authority
- CN
- China
- Prior art keywords
- task
- scheduled
- schedule sequences
- circle time
- time
- 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
- 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
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Educational Administration (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Position Fixing By Use Of Radio Waves (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention provides a kind of more stars observation layering dispatching method and system based on divide-and-conquer strategy, specifically includes the following steps: S1, using allocation algorithm task is distributed to each track circle time, forms the task-set of each track circle time；S2, solved using approximate optimum algorithm each track circle time take up an official post business collection schedule sequences；S3, allocation algorithm update the allocation plan of allocation algorithm according to not the generating schedule sequences of the task that each track circle time schedule sequences are fed back again, and then form the new task-set of new each track circle time；S4, step S1, S2, S3 are repeated until reaching the termination condition of more star observation layering scheduling.Simplified by the way that a complicated combinatorial optimization problem is carried out effective decompose, become a dual layer resist, effectively reduced the complexity of problem solving, brilliant performance is especially shown when solving extensive more star observation scheduling problems.The present invention is applied to satellite dispatching technique field.
Description
Technical field
The present invention relates to satellite dispatching technique field more particularly to a kind of more stars observation layering scheduling based on divide-and-conquer strategy
Method and system.
Background technique
Earth observation satellite (Earth Observing Satellites, EOSs) orbits the earth, and can obtain the earth
Surface specifies the image information in region to meet the observation requirements of user, plays pass in fields such as environmental monitoring, intelligence reconnaissances
Bond angle color.In order to make full use of rare satellite resource, efficient observation scheduling method is studied to improve satellite application level, tool
There is important meaning.Since the earth observation demand of user is usually more than the observing capacity of existing satellite resource, satellite is seen
Surveying scheduling problem (Satellite Observation Scheduling Problem, SOSP) is a kind of excessively subscription (over
Subscribed) problem.
Early stage, research relevant to SOSP laid particular emphasis on the scheduling of single satellite.Vasquez and Hao asks moonscope scheduling
Topic is converted into classical knapsack model, proposes a kind of Tabu-Search Algorithm model.In another paper, Vasquez
The method based on decomposition is proposed with Hao to obtain optimization problem supremum, is defended obtained by Tabu-Search Algorithm with assessing
The quality of star observation scheduling scheme.The scheduling of Spot5 satellite is converted constraint satisfaction problemx by Bensana etc., and using accurately
Or approximate algorithm respectively solves the problem.Gabrel and Vanderpooten proposes a kind of acyclic graph model to describe
The scheduling problem of Spot5 satellite firstly generates a plurality of active path, then therefrom selectes a best path.Lin etc. is used
Mathematic programming methods obtain near-optimization moonscope scheduling scheme.Baek etc. is using a kind of new genetic algorithm analog satellite
Actual observation scheduling problem.Miguel etc. is defended based on the effective inequality being packaged in node with 3- canonical autonomous system for Spot5
The daily Imaging Scheduling of star provides programme.
Moonscope scheduling is to determine every satellite every under the conditions of meeting a series of observation mission demands and constraint
The maximization of observation income is realized in observation activity in a track circle time, is the key that improve satellite service efficiency.Traditional is more
Using every satellite as thread, but if scheduling time section is larger, a satellite can orbit the earth star dispatching algorithm
Multiple track circle times (such as imaging reconnaissance satellite is generally daily around 14 track circle times of ground flight or so), different satellites may be seen
Identical target is measured, and same satellite may can observe identical target in its different track circle time, cause
Observation conflict is more complicated (commonly assuming that a target only need to observe once), especially in extensive more star observation scheduling problems
Solution degree it is more complicated.
Summary of the invention
In view of the deficienciess of the prior art, the object of the present invention is to provide a kind of, more stars observation based on divide-and-conquer strategy divides
Layer scheduling method and system, this method effectively reduce the complexity of problem solving, are especially solving extensive more star observation tune
Brilliant performance is shown when degree problem.
Satellite is decomposed track circle time first, each rail ring when carrying out more star observation layering scheduling by the present invention
It is secondary to regard the resource for having earth observation ability as, i.e., the scheduling of multi-track circle time is converted by more star multi-track circle time scheduling problems
Problem.
If there is earth observation ability track circle time resource, O={ o comprising general in set Oj| j=1,2 ...,
M }, it is the quantity of scheduling problem middle orbit circle time resource that M, which is the base of set,.If set T is target collection to be observed, packet
Band target after including point target and decomposing, T={ ti| i=1,2 ..., N }, N indicates task quantity.0-1 variable xijFor table
Show the task t in scheduling schemeiWhether in track circle time ojUpper completion: task tiIn track circle time ojX when upper completionij=1, it is no
Then xij=0.The target of more star observation schedulings is to maximize observation income, for the sum of the income of all completion tasks, it may be assumed that
In formula, piIt is completion task tiObtainable income.
Need to meet complicated constraint condition, including the constraint of unique constraints, energy constraint, memory capacity in scheduling process
Time-constrain is put with surveying, specific as follows:
Unique constraints: each task only needs at most to complete primary, it may be assumed that
Energy constraint: track circle time ojOn, the energy of observation activity and sensor side-sway activity consumption cannot be more than and should be somebody's turn to do
Allow the ceiling capacity consumed in track circle time:
In formula, eojIt indicates in track circle time ojThe rate of upper satellite and sensor observation activity consumption energy；teiIt indicates to appoint
Be engaged in tiEnd time；tsiExpression task tiAt the beginning of；yihIt is 1-0 variable, indicates tiWhether prior to thIt completes, prior to then
1 is taken, otherwise takes 0；esjIt indicates in track circle time ojThe rate of upper sensor side-sway activity consumption energy；θiIt indicates in track ojOn
Observation mission tiWhen sensor needed for survey pendulum angle；θhIt indicates in track ojUpper observation mission thWhen sensor needed for survey pendulum
Angle；EjIndicate track circle time ojAllow consume maximum energy storage capacity.
Memory capacity constraint: in track circle time ojOn, observation activity, which obtains earth observation data, goes forward side by side and stores on planet, but
It is the storage resource of consumption no more than maximum storage capacity allowed in the track circle time:
In formula, wjIndicate track circle time ojThe rate of upper observation activity consumption storage resource；WjIndicate track circle time ojPermit
Perhaps maximum storage capacity is consumed.
Side-sway time-constrain: satellite needs to turn back on when observing different target, simultaneously because opposite space geometry
Relationship changes, and needs to adjust the side-sway angle of sensor to be directed at observed object.Therefore adjacent in the same track circle time
Two observation missions between, need the sufficient time that satellite sensor is allowed to boot up and calibrate and (adjust side-sway angle):
tsh-tei≥aj+|θh-θi|/vj, i, h=1,2 ..., N, yih=1, j=1,2 ... M
In formula, tshExpression task thAt the beginning of；ajIt indicates in track ojOn before being observed to some target
The sensor available machine time.
It is assumed that star scheduling problem more than one considers N number of task, M track circle time, then variable number is NM, constraint condition
Quantity is N+3M.Generally all problem scale is very big for the satellite scheduling problem of real scene, considers the scheduling problem of general scale: example
If number of targets to be observed (including the observation band after decomposing) is 500,5 satellites are considered, dispatching cycle is one day, every satellite
Around 14 circle of ground flight, then about 70 track circle times are shared, then the scale of variable is 500x70=35000, the rule of constraint condition
Mould is 500+3x70=710.It is seen that more star observation scheduling problems are a complicated discrete optimization problems of device, it is how quickly high
Solution this problem of effect, needs to design efficient scheduling frame and algorithm.
To sum up, the technical solution adopted by the present invention is that: it is a kind of based on divide-and-conquer strategy more stars observation layering dispatching method, packet
Include following steps:
S1, task is distributed to each track circle time using allocation algorithm, forms the task-set of each track circle time；
S2, solved using approximate optimum algorithm each track circle time take up an official post business collection schedule sequences；
Not the generating schedule sequences of the task that S3, allocation algorithm are fed back according to each track circle time schedule sequences updates again divides
Allocation plan with algorithm, and then form the new task-set of new each track circle time；
S4, step S1, S2, S3 are repeated until reaching the termination condition of more star observation layering scheduling.
As a further improvement of the above technical scheme, in step S2, the approximate optimum algorithm is simulated annealing.
As a further improvement of the above technical scheme, the simulated annealing obtains the initial tune of current orbit circle time
The step of degree series includes:
S211, the scheduler task to be planned in monorail circle time task-set is arranged successively from big to small according to weight, shape
The initiating task collection U not being scheduled at monorail circle time；
S212, the maximum task t of weight is selected from Uk, using being inserted into and deleting neighbour structure transformation, tkIt is added to office
New local solution S ' is generated in portion solution S, if the fitness of S ' is higher than S, S is updated, S=S ' is enabled, otherwise S not carried out
It updates, while by tkIt removes from U to UjIn, wherein UjIndicate the initiating task collection not being scheduled in track circle time；
S213, step S212 is repeated up to all traversal completions of the task in U, obtain the schedule sequences S on current orbit
With the task-set U not being scheduled on current orbitj。
As a further improvement of the above technical scheme, the field searching structure of the simulated annealing includes based on greedy
The neighborhood search structure of greedy criterion and the neighborhood search structure based on probabilistic search, the simulated annealing includes neighborhood search
Structure dynamics selection strategy: it according to optimization performance of two kinds of neighborhood search structures in allocation plan before, determines at subsequent point
Selection with field searching structure in scheme.
As a further improvement of the above technical scheme, the simulated annealing of the neighborhood search structure based on greedy criterion
It specifically includes:
The task-set U that S221, the schedule sequences S for obtaining current orbit circle time and current orbit circle time are not scheduledj；
S222, in UjThe maximum task t of one weight not yet dispatched of middle taking-upk, enable Uj=Uj-tk；
S223, by tkIt is inserted into the schedule sequences S of current orbit circle time, judges whether to meet side-sway time-constrain,
New schedule sequences S " is directly formed if meeting, if being unsatisfactory for deleting and tkNew schedule sequences are formed after the task of conflict
S ", wherein the task deposit S deleteduIn, SuIt is the set of tasks that all track circle times are not scheduled；
S224, judge whether new schedule sequences meet energy constraint and memory capacity constraint:
If satisfied, then repeating step S221, S222, S223, S224 until UjIn all task all traverse completion, obtain
Final schedule sequences S " on the current orbit and task that current orbit is not scheduled is stored in set SuIn；
It is repeated after meeting energy constraint and memory capacity constraint if not satisfied, then successively deleting the smallest task of weight
Step S221, S222, S223, S224 are until UjIn all task all traverse completion, obtain the final scheduling on current orbit
Sequence S " and the task that current orbit is not scheduled deposit set SuIn, wherein SuIt is that all track circle times are not scheduled to appoint
Business set.
As a further improvement of the above technical scheme, the simulated annealing of the neighborhood search structure based on probabilistic search
It specifically includes:
The task-set U that S231, the schedule sequences S for obtaining current orbit circle time and current orbit circle time are not scheduledj；
S232, in UjIt is middle to take out any task tkProbability beWherein indexkFor task tkPriority
Index enables Uj=Uj-tk；
S233, by tkIt is inserted into the schedule sequences S of current orbit circle time, judges whether to meet side-sway time-constrain,
New schedule sequences S " is directly formed if meeting, if being unsatisfactory for deleting and tkNew schedule sequences are formed after the task of conflict
S ", wherein the task deposit S deleteduIn, SuIt is the set of tasks that all track circle times are not scheduled；
S234, judge whether new schedule sequences meet energy constraint and memory capacity constraint:
If satisfied, then repeating step S231, S232, S233, S234 until UjIn all task all traverse completion, obtain
Final schedule sequences S " on the current orbit and task that current orbit is not scheduled is stored in set SuIn；
If not satisfied, then deleting task tjProbability beWherein indexjFor task tjIt is excellent
First grade index repeats step S231, S232, S233, S234 until U after meeting energy constraint and memory capacity constraintjIn
All tasks all traverse completion, obtaining the final schedule sequences S " on current orbit and current orbit not being scheduled for task
It is stored in set SuIn, wherein SuIt is the set of tasks that all track circle times are not scheduled.
As a further improvement of the above technical scheme, the allocation algorithm is ant colony optimization algorithm.
As a further improvement of the above technical scheme, it is specifically included in step S3:
S31, the set of tasks S that all track circle times are not scheduled is obtained according to each track circle time schedule sequencesu；
S32, from SuOne highest task of weight of middle selection inserts it into the conflict the smallest track circle time of weight,
If the conflict weight of the task is less than the weight of the task itself, it is inserted into successfully and is updated allocation plan, is not otherwise updated point
With scheme；
S33, step S32 is repeated until SuIn all not scheduled tasks all traverse completion.
As a further improvement of the above technical scheme, in step S4, the termination condition of more star observation layering scheduling
To calculate, the time reaches time preset value or the number of iterations reaches number preset value.
The invention also discloses a kind of more stars observation layering scheduling system based on divide-and-conquer strategy, the technical solution used
It is:
A kind of more stars observation layering scheduling system based on divide-and-conquer strategy, including memory and processor, the memory
It is stored with computer program, when the processor executes the computer program the step of realization above method.
Advantageous effects of the invention:
Task is assigned to different track circle times up by allocation algorithm by the present invention, and each track circle time corresponding one is appointed
Business collection, this is the task Distribution Layer of this method, then obtains the observation of satellite in each track circle time by approximate optimum algorithm
Scheme, this is the task scheduling layer of this method, finally updates allocation algorithm again according to the feedback of each track circle time schedule sequences
Allocation plan, and then form the new task-set of new each track circle time.It iterates until meet algorithm termination condition,
One complicated combinatorial optimization problem carries out effective decompose and simplifies, and becomes a dual layer resist, effectively reduces problem solving
Complexity, brilliant performance is especially shown when solving extensive more star observation scheduling problems.
Detailed description of the invention
Fig. 1 is the frame construction drawing of the present embodiment.
Specific embodiment
For the ease of implementation of the invention, it is further described below with reference to specific example.
A kind of more stars observation layering dispatching method based on divide-and-conquer strategy as shown in Figure 1, comprising the following steps:
S1, task is distributed to each track circle time using allocation algorithm, forms the task-set of each track circle time；
S2, solved using approximate optimum algorithm each track circle time take up an official post business collection schedule sequences；
Not the generating schedule sequences of the task that S3, allocation algorithm are fed back according to each track circle time schedule sequences updates again divides
Allocation plan with algorithm, and then form the new task-set of new each track circle time；
S4, step S1, S2, S3 are repeated until reaching the termination condition of more star observation layering scheduling, more star observation layerings are adjusted
The termination condition of degree is that the calculating time reaches time preset value or the number of iterations reaches number preset value.
In step S2, approximate optimum algorithm is simulated annealing, and simulated annealing obtains the first of current orbit circle time
The step of beginning schedule sequences includes:
S211, the scheduler task to be planned in monorail circle time task-set is arranged successively from big to small according to weight, shape
The initiating task collection U not being scheduled at monorail circle time；
S212, the maximum task t of weight is selected from Uk, using being inserted into and deleting neighbour structure transformation, tkIt is added to office
New local solution S ' is generated in portion solution S, if the fitness of S ' is higher than S, S is updated, S=S ' is enabled, otherwise S not carried out
It updates, while by tkIt removes from U to UjIn, wherein UjThe initiating task collection not being scheduled in track circle time is indicated, wherein inserting
Entering and deleting neighborhood transformation indicates first to be inserted into task, then deletes a kind of neighborhood search mode with newly insertion task conflict task,
Local solution refer to it is potential continue plus task solution, fitness measure solution quality, here solution i.e. needed for seek it is initial
Schedule sequences；
S213, step S212 is repeated up to all traversal completions of the task in U, obtain the schedule sequences of current orbit circle time
The task-set U that S and current orbit circle time are not scheduledj。
The field searching structure of simulated annealing in the present embodiment include neighborhood search structure based on greedy criterion with
Neighborhood search structure based on probabilistic search, simulated annealing include neighborhood search structure dynamics selection strategy: according to two kinds
Optimization performance of the neighborhood search structure in allocation plan before determines the choosing of the field searching structure in subsequent allocations scheme
It selects:
During to problem Optimization Solution, different neighbour structure search strategies may have different advantages, therefore
The present embodiment proposes that a kind of adaptation mechanism realizes the dynamic adjustment of two kinds of different neighborhood search structures, according to every kind of neighborhood search
Optimization performance of the structure in allocation plan before gives the optimization performance in the number of iterations before, determines them subsequent
The probability selected in iterative process, this is a kind of thought based on intensified learning.It is assumed that pro1And pro2It is expressed as selecting
The probability of neighborhood search structure based on greedy criterion and the neighborhood search structure based on probabilistic search, in algorithm initialization,
Set proi=0.5, i=1,2 pass through the selection of each neighbour structure of following Policy Updates every certain the number of iterations Itr
Probability:
proi=proi′/∑I=1,2proi′
In formula, proi' indicate intermediate variable；η be the inertia weight factor, indicate neighbour structure before select probability shared by
Specific gravity；(1- η) indicates current newest historical search experience to the weight for updating select probability；SeliFor nearest Itr iteration
In the process, the selected number of i-th of neighbour structure；SuciIt indicates to produce higher-quality solution using i-th of neighbour structure
Number.
Finally the select probability of different neighbour structures is standardized.The update rule of above-mentioned neighbour structure select probability
Show that the neighbour structure more preferably solved can be generated, it will obtain bigger select probability.It was run to realize in optimization algorithm
The dynamic adaptive selection of neighbour structure in journey.
In neighbour structure search, insertion/deletion neighborhood is a kind of effective neighbour structure of classics, and is commonly known as handed over
Change neighborhood.The maximum task of weight is selected in each set of tasks for being never scheduled and not yet traversing, is inserted into current sight
It surveys in schedule sequences, if there is the task with insertion task conflict in schedule sequences, Conflict Tasks is all deleted, because
The simulated annealing of this neighborhood search structure based on greedy criterion specifically includes:
The task-set U that S221, the schedule sequences S for obtaining current orbit circle time and current orbit circle time are not scheduledj；
S222, in UjThe maximum task t of one weight not yet dispatched of middle taking-upk, tkIt is not included in taboo list, enables Uj
=Uj-tk, wherein tkBeing not included in taboo list indicates that a kind of algorithm mechanism, that is, being placed in taboo list for task are not involved in neighborhood
Search structure；
S223, by tkIt is inserted into the schedule sequences S of current orbit circle time, judges whether to meet side-sway time-constrain,
New schedule sequences S " is directly formed if meeting, if being unsatisfactory for deleting and tkNew schedule sequences are formed after the task of conflict
S ", wherein the task deposit S deleteduIn, SuIt is the set of tasks that all track circle times are not scheduled；
S224, judge whether new schedule sequences meet energy constraint and memory capacity constraint:
If satisfied, then repeating step S221, S222, S223, S224 until UjIn all task all traverse completion, obtain
Final schedule sequences S " on the current orbit and task that current orbit is not scheduled is stored in set SuIn；
It is repeated after meeting energy constraint and memory capacity constraint if not satisfied, then successively deleting the smallest task of weight
Step S221, S222, S223, S224 are until UjIn all task all traverse completion, obtain the final scheduling on current orbit
Sequence S " and the task that current orbit is not scheduled deposit set SuIn, wherein SuIt is that all track circle times are not scheduled to appoint
Business set.
Neighborhood search structure based on greedy criterion is conducive to local search.In order to improve the diversity of solution, have also been devised
A kind of neighborhood search structure based on probabilistic search not only considers the weight of task in the neighbour structure, it is also contemplated that completes
Each task needs the side-sway resource and time window resource of potential consumption.Comprehensive weight, side-sway angle and time window, are being inserted into
It is each task t in replacement neighbour structure conversion processkCalculate a priority indicator indexk, weight is calculated separately first
Index, time window index and side-sway angle index:
Weight index are as follows:
In formula, iWkFor weight index；wkIt is the weight of k-th of task；wiIt is the weight of i-th of task；njIt is UjNumber of tasks
Amount.
Time window index are as follows:
In formula, iTkFor time window index；spankFor tkTime window length；spaniFor tiTime window length.
Side-sway angle index are as follows:
In formula, iTkFor side-sway angle index；θiIn track ojUpper observation mission tiWhen sensor needed for survey pendulum angle；θk
In track ojUpper observation mission tkWhen sensor needed for survey pendulum angle.
Priority indicator indexkIt is ultimately expressed as:
indexk=iWk α·iTk β·iθk γ
In formula, α, beta, gamma respectively indicates the impact factor of different elements, is specifically arranged by user for particular problem.
The simulated annealing of neighborhood search structure based on probabilistic search specifically includes:
The task-set U that S231, the schedule sequences S for obtaining current orbit circle time and current orbit circle time are not scheduledj；
S232, in UjIt is middle to take out any task tkProbability beWherein indexkFor task tkPreferentially
Grade index, tkIt is not included in taboo list, enables Uj=Uj-tk；
S233, by tkIt is inserted into the schedule sequences S of current orbit circle time, judges whether to meet side-sway time-constrain,
New schedule sequences S " is directly formed if meeting, if being unsatisfactory for deleting and tkNew schedule sequences are formed after the task of conflict
S ", wherein the task deposit S deleteduIn, SuIt is the set of tasks that all track circle times are not scheduled；
S234, judge whether new schedule sequences meet energy constraint and memory capacity constraint:
If satisfied, then repeating step S231, S232, S233, S234 until UjIn all task all traverse completion, obtain
Final schedule sequences S " on the current orbit and task that current orbit is not scheduled is stored in set SuIn；
If not satisfied, then deleting task tjProbability beWherein indexjFor task tjIt is excellent
First grade index repeats step S231, S232, S233, S234 until U after meeting energy constraint and memory capacity constraintjIn
All tasks all traverse completion, obtaining the final schedule sequences S " on current orbit and current orbit not being scheduled for task
It is stored in set SuIn, wherein SuIt is the set of tasks that all track circle times are not scheduled.
Allocation algorithm is ant colony optimization algorithm, is specifically included in step S3:
S31, the set of tasks S that all track circle times are not scheduled is obtained according to each track circle time schedule sequencesu；
S32, from SuOne highest task of weight of middle selection inserts it into the conflict the smallest track circle time of weight,
If the conflict weight of the task is less than the weight of the task itself, it is inserted into successfully and is updated allocation plan, is not otherwise updated point
With scheme；
S33, step S32 is repeated until SuIn all not scheduled tasks all traverse completion.
Wherein, when task being distributed to each track circle time for the first time using ant colony optimization algorithm in step sl, ant colony
Algorithm can carry out original allocation according to parameter setting, directly form the allocation plan of first time.
As a further improvement of the above technical scheme, in step S4, algorithm termination condition is the calculating time to reach the time
Preset value or the number of iterations reach number preset value.
The present embodiment also discloses a kind of more stars observation layering scheduling system based on divide-and-conquer strategy, the technical side used
Case is:
A kind of more stars observation layering scheduling system based on divide-and-conquer strategy, including memory and processor, memory storage
There is the step of computer program, processor realizes the above method when executing computer program.
In order to assess the performance of the base ACO-SA algorithm of divide-and-conquer strategy in this present embodiment, by itself and TABU search (TS),
Genetic algorithm (GA), simulated annealing (SA) highest priority Priority-driven Scheduling Algorithm (HPFS) and the ant group optimization with local search are calculated
Method compares (ACO-LS).ACO-LS in conjunction with local searching strategy, generates more star programming dispatchings using ant colony optimization algorithm
The solution of problem, but the strategy divided and rule is not used in the algorithm, but scheduling problem is regarded as an entirety and is carried out
It solves.
For the performance of more comprehensive assessment algorithm comprehensively, the present embodiment proposes six simulating scenes.It uses and defends
Star includes two groups, and first group is 8 resources and point's series reconnaissance satellite, and second group is 16 reconnaissance satellites.Every satellite around
Time about 100 minutes of one circle of earth operation, the 14 track circle times of operation in one day.The cross side swinging range of sensor be [-
330,330].Dispatching cycle is 24 hours.SEE time window and the side of satellite and target are calculated with professional software STK
Swinging.Six scenes are respectively: observation of first scene, 8 satellites to Daxinganling forest area；Second scenario is 8 and defends
Observation of the star to Forest in Changbai Mountain Forest Region；Third scene is observation of 8 satellites to Daxinganling forest area and Forest in Changbai Mountain Forest Region；4th
Observation of 16 satellites of a scene to Daxinganling forest area；5th scene is observation of 16 satellites to Forest in Changbai Mountain Forest Region；The
Six scenes are observation of 16 satellites to Daxinganling forest area and Forest in Changbai Mountain Forest Region.
Using C Plus Plus program realize the algorithm, and Intel Core (TM) i7-4810MQ@2.8GHZ CPU,
Running experiment on the computer of 7 system of Windows of 16.0GB RAM.
First group of emulation experiment scene is to the calculated result such as table 1 of the 6th group of emulation experiment scene to the data institute in table 6
Show.By the observation to experimental data, it is seen that following several phenomenons:
1.ACO-SA steadily shows optimum performance in each group of experiment, shows and integrates ant using divide-and-conquer strategy
Group's algorithm and simulated annealing can be with the more star observation scheduling problems of effective solution.Especially ACO-SA algorithm always compares ACO-
LS generates better observation scheduling scheme, shows divide-and-conquer strategy for solving extensive moonscope scheduling problem with significant excellent
Gesture；
2. under the simulating scenes of 8 satellites, to significantly less than 16 satellite simulation fields of observation coverage rate of wood land
To the observation coverage rate of wood land under scape, show that suitably increasing moonscope resource is to improve the weight of forest observation covering power
Want means；
3. by the data of third group emulation experiment it can be found that 8 are difficult to efficiently accomplish to Daxing in satellite 24 hours
Pacify effective observation covering of ridge and Changbaishan area, the most efficient ACO-SA algorithm of use also can only achieve general 50% covering
Rate.16 then can effectively observe the forest reserves of covering Changbaishan area in satellite 24 hours substantially, and observation completion rate reaches
97% or more；
4. the timeliness of the observation programming dispatching of star more than sees that the timeliness of ACO-SA is in by-level, and ACO-SA generation is defended
The time of star observation program increases with the growth of number of tasks and number of satellite scale in approximately linear, in maximum-norm the (the 6th
Group simulating scenes) under, the programming dispatching scheme of high quality can be obtained in 900 seconds, show the applicability of ACO-SA algorithm.
Calculated result of the different moonscope dispatching algorithms of table 1 in first group of test scene
| Algorithm | CT | NT | Cov (%) | Time(s) |
| TS | 1862 | 1215 | 65.22% | 321.52 |
| GA | 1862 | 1244 | 66.52% | 425.62 |
| SA | 1862 | 1198 | 62.12% | 564.24 |
| ACO-LS | 1862 | 1254 | 67.25% | 368.45 |
| ACO-SA | 1862 | 1351 | 72.65% | 388.78 |
Calculated result of the different moonscope dispatching algorithms of table 2 in second group of test scene
| Algorithm | CT | NT | Cov (%) | Time(s) |
| TS | 1456 | 1025 | 70.40% | 280.25 |
| GA | 1456 | 1085 | 74.52% | 355.32 |
| SA | 1456 | 987 | 67.79% | 451.52 |
| ACO-LS | 1456 | 1102 | 75.69% | 298.12 |
| ACO-SA | 1456 | 1203 | 83.24% | 326.45 |
Calculated result of the different moonscope dispatching algorithms of table 3 in third group test scene
| Algorithm | CT | NT | Cov (%) | Time(s) |
| TS | 3318 | 1425 | 42.95% | 582.35 |
| GA | 3318 | 1524 | 45.93% | 565.15 |
| SA | 3318 | 1324 | 39.90% | 684.60 |
| ACO-LS | 3318 | 1588 | 46.85% | 535.40 |
| ACO-SA | 3318 | 1658 | 50.02% | 606.66 |
Calculated result of the different moonscope dispatching algorithms of table 4 in the 4th group of test scene
| Algorithm | CT | NT | Cov (%) | Time(s) |
| TS | 1862 | 1625 | 87.27% | 541.12 |
| GA | 1862 | 1678 | 90.12% | 655.45 |
| SA | 1862 | 1598 | 85.82% | 743.15 |
| ACO-LS | 1862 | 1699 | 91.25% | 524.36 |
| ACO-SA | 1862 | 1742 | 93.56% | 560.65 |
Calculated result of the different moonscope dispatching algorithms of table 5 in the 5th group of test scene
| Algorithm | CT | NT | Cov (%) | Time(s) |
| TS | 1456 | 1355 | 93.06% | 482.32 |
| GA | 1456 | 1382 | 94.92% | 505.24 |
| SA | 1456 | 1275 | 87.57% | 661.32 |
| ACO-LS | 1456 | 1398 | 96.02% | 458.55 |
| ACO-SA | 1456 | 1418 | 97.39% | 500.48 |
Calculated result of the different moonscope dispatching algorithms of table 6 in the 6th group of test scene
Technical term ACO-SA is inventive algorithm in upper table, and TS is TABU search, and GA is genetic algorithm, and SA is that simulation is moved back
Fire, ACO-LS be using ant colony optimization algorithm in conjunction with local searching strategy, generate the algorithm of the solution of more star programming dispatching problems.
What the present embodiment proposed divides and rule frame with versatility, suitable for the programming dispatching of many types, such as nothing
Man-machine scheduling and vehicle path planning etc..Subsequent work is the further scheduling problem for studying quick satellite and super quick satellite,
And the integrated scheduling problem that imaging sum number passes.
Contain the explanation of the preferred embodiment of the present invention above, this be for the technical characteristic that the present invention will be described in detail, and
Be not intended to for summary of the invention being limited in concrete form described in embodiment, according to the present invention content purport carry out other
Modifications and variations are also protected by this patent.The purport of the content of present invention is to be defined by the claims, rather than by embodiment
Specific descriptions are defined.
Claims (10)
1. a kind of more stars observation layering dispatching method based on divide-and-conquer strategy, which comprises the following steps:
S1, task is distributed to each track circle time using allocation algorithm, forms the task-set of each track circle time；
S2, solved using approximate optimum algorithm each track circle time take up an official post business collection schedule sequences；
S3, the distribution for updating allocation algorithm again according to not the generating schedule sequences of the task that each track circle time schedule sequences are fed back
Scheme, and then form the new task-set of new each track circle time；
S4, step S1, S2, S3 are repeated until reaching the termination condition of more star observation layering scheduling.
2. more stars observation layering dispatching method based on divide-and-conquer strategy according to claim 1, which is characterized in that step S2
In, the approximate optimum algorithm is simulated annealing.
3. more stars observation layering dispatching method based on divide-and-conquer strategy according to claim 2, which is characterized in that the simulation
Annealing algorithm obtain current orbit circle time initial schedule sequence the step of include:
S211, the scheduler task to be planned in monorail circle time task-set is arranged successively from big to small according to weight, is formed single
The initiating task collection U that track circle time is not scheduled；
S212, the maximum task t of weight is selected from Uk, using being inserted into and deleting neighbour structure transformation, tkIt is added to local solution
New local solution S ' is generated in S, if the fitness of S ' is higher than S, S is updated, S=S ' is enabled, and otherwise S is not carried out more
Newly, while by tkU is moved to from UjIn, wherein UjIndicate the task-set not being scheduled in track circle time；
S213, repeat step S212 until U in task all traversal complete, obtain current orbit circle time schedule sequences S with
The task-set U that current orbit circle time is not scheduledj。
4. more stars observation layering dispatching method based on divide-and-conquer strategy according to claim 3, which is characterized in that the simulation
The field searching structure of annealing algorithm includes the neighborhood search structure based on greedy criterion and the neighborhood search based on probabilistic search
Structure, the simulated annealing include neighborhood search structure dynamics selection strategy: according to two kinds of neighborhood search structures before
Optimization performance in allocation plan determines the selection of the field searching structure in subsequent allocations scheme.
5. more stars observation layering dispatching method based on divide-and-conquer strategy according to claim 4, which is characterized in that based on greediness
The simulated annealing of the neighborhood search structure of criterion specifically includes:
The task-set U that S221, the schedule sequences S for obtaining current orbit circle time and current orbit circle time are not scheduledj；
S222, in UjThe maximum task t of one weight not yet dispatched of middle taking-upk, enable Uj=Uj-tk；
S223, by tkIt is inserted into the schedule sequences S of current orbit circle time, judges whether to meet side-sway time-constrain, if meeting
New schedule sequences S " is then directly formed, is deleted if being unsatisfactory for and tkNew schedule sequences S " is formed after the task of conflict,
The task of middle deletion is stored in SuIn, SuIt is the set of tasks that all track circle times are not scheduled；
S224, judge whether new schedule sequences S " meets energy constraint and memory capacity constraint:
If satisfied, then repeating step S221, S222, S223, S224 until UjIn all task all traverse completion, obtain current
Final schedule sequences S " on the track and task that current orbit is not scheduled is stored in set SuIn；
If not satisfied, then successively deleting the smallest task of weight repeats step after meeting energy constraint and memory capacity constraint
S221, S222, S223, S224 are until UjIn all task all traverse completion, obtain the final schedule sequences on current orbit
S " and the task that current orbit is not scheduled deposit set SuIn, wherein SuIt is the task-set that all track circle times are not scheduled
It closes.
6. more stars observation layering dispatching method based on divide-and-conquer strategy according to claim 4, which is characterized in that be based on probability
The simulated annealing of the neighborhood search structure of search specifically includes:
The task-set U that S231, the schedule sequences S for obtaining current orbit circle time and current orbit circle time are not scheduledj；
S232, in UjIt is middle to take out any task tkProbability beWherein indexkFor task tkPriority refers to
Mark, enables Uj=Uj-tk；
S233, by tkIt is inserted into the schedule sequences S of current orbit circle time, judges whether to meet side-sway time-constrain, if meeting
New schedule sequences S " is then directly formed, is deleted if being unsatisfactory for and tkNew schedule sequences S " is formed after the task of conflict,
The task of middle deletion is stored in SuIn, SuIt is the set of tasks that all track circle times are not scheduled；
S234, judge whether new schedule sequences meet energy constraint and memory capacity constraint:
If satisfied, then repeating step S231, S232, S233, S234 until UjIn all task all traverse completion, obtain current
Final schedule sequences S " on the track and task that current orbit is not scheduled is stored in set SuIn；
If not satisfied, then deleting task tjProbability beWherein indexjFor task tjPriority refers to
Mark repeats step S231, S232, S233, S234 until U after meeting energy constraint and memory capacity constraintjIn it is all
Task all traverses completion, the task deposit collection for obtaining the final schedule sequences S " on current orbit and current orbit not being scheduled
Close SuIn, wherein SuIt is the set of tasks that all track circle times are not scheduled.
7. according to claim 1 to more stars observation layering dispatching method described in 6 any one based on divide-and-conquer strategy, feature exists
In the allocation algorithm is ant colony optimization algorithm.
8. more stars observation layering dispatching method based on divide-and-conquer strategy according to claim 7, which is characterized in that in step S3
It specifically includes:
S31, the set of tasks S that all track circle times are not scheduled is obtained according to each track circle time schedule sequencesu；
S32, from SuOne highest task of weight of middle selection inserts it into the conflict the smallest track circle time of weight, if this
The conflict weight of business is less than the weight of the task itself, then is inserted into successfully and updates allocation plan, otherwise do not update allocation plan；
S33, step S32 is repeated until SuIn all not scheduled tasks all traverse completion.
9. according to claim 1 to more stars observation layering dispatching method described in 6 any one based on divide-and-conquer strategy, feature exists
In in step S4, the termination condition of more star observation layering scheduling is the calculating time to reach time preset value or the number of iterations
Reach number preset value.
10. a kind of more stars observation layering scheduling system based on divide-and-conquer strategy, including memory and processor, the memory are deposited
Contain computer program, which is characterized in that the processor is realized in claim 1 to 9 when executing the computer program appoints
The step of one the method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810967096.7A CN109087023B (en) | 2018-08-23 | 2018-08-23 | Multi-satellite observation layered scheduling method and system based on divide-and-conquer strategy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810967096.7A CN109087023B (en) | 2018-08-23 | 2018-08-23 | Multi-satellite observation layered scheduling method and system based on divide-and-conquer strategy |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109087023A true CN109087023A (en) | 2018-12-25 |
| CN109087023B CN109087023B (en) | 2021-03-02 |
Family
ID=64794304
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810967096.7A Active CN109087023B (en) | 2018-08-23 | 2018-08-23 | Multi-satellite observation layered scheduling method and system based on divide-and-conquer strategy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109087023B (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109767128A (en) * | 2019-01-15 | 2019-05-17 | 中国人民解放军国防科技大学 | imaging satellite autonomous task planning method based on machine learning |
| CN110728447A (en) * | 2019-10-10 | 2020-01-24 | 合肥工业大学 | Partitioned satellite task planning method for achieving regional target coverage at earliest |
| CN110766284A (en) * | 2019-09-24 | 2020-02-07 | 合肥工业大学 | Multi-star task synthesis method and system |
| CN111176807A (en) * | 2019-12-09 | 2020-05-19 | 中国空间技术研究院 | Multi-satellite cooperative task planning method |
| CN112016812A (en) * | 2020-08-06 | 2020-12-01 | 中南大学 | Multi-unmanned aerial vehicle task scheduling method, system and storage medium |
| CN112070370A (en) * | 2020-08-21 | 2020-12-11 | 中南大学 | Relay satellite task planning method, system and storage medium |
| CN112613644A (en) * | 2020-12-09 | 2021-04-06 | 合肥工业大学 | Multi-satellite multi-station integrated scheduling method and system |
| CN112737660A (en) * | 2020-12-09 | 2021-04-30 | 合肥工业大学 | Multi-satellite multi-station data downloading scheduling method and system |
| CN113158384A (en) * | 2021-03-03 | 2021-07-23 | 东北石油大学 | Oil and gas pipeline route planning method and system |
| CN113408063A (en) * | 2021-07-20 | 2021-09-17 | 中南大学 | Multi-satellite multi-reconnaissance target-oriented track maneuver optimization 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 |
| CN113919122A (en) * | 2021-08-05 | 2022-01-11 | 合肥工业大学 | Multi-star task scheduling method and system based on simulated annealing algorithm |
| CN114604443A (en) * | 2022-02-17 | 2022-06-10 | 中国电子科技集团公司第十研究所 | Multi-satellite measurement and control scheduling method and system under real complex constraint condition |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104361234A (en) * | 2014-11-15 | 2015-02-18 | 北京理工大学 | Method for optimizing multi-star multitask observation dispatching under complicated constraint condition |
| CN106951312A (en) * | 2016-08-31 | 2017-07-14 | 合肥工业大学 | The dispatching method and system of many star tasks |
| CN108335012A (en) * | 2017-12-26 | 2018-07-27 | 佛山科学技术学院 | A kind of intelligence remote sensing satellite stratification distributed freedom cotasking planning system |
-
2018
- 2018-08-23 CN CN201810967096.7A patent/CN109087023B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104361234A (en) * | 2014-11-15 | 2015-02-18 | 北京理工大学 | Method for optimizing multi-star multitask observation dispatching under complicated constraint condition |
| CN106951312A (en) * | 2016-08-31 | 2017-07-14 | 合肥工业大学 | The dispatching method and system of many star tasks |
| CN108335012A (en) * | 2017-12-26 | 2018-07-27 | 佛山科学技术学院 | A kind of intelligence remote sensing satellite stratification distributed freedom cotasking planning system |
Non-Patent Citations (1)
| Title |
|---|
| WUGUOHUA等: ""An adaptive Simulated Annealing-based satellite observation scheduling method combined with a dynamic task clustering strategy"", 《COMPUTERS&INDUSTRIAL ENGINEERING》 * |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109767128B (en) * | 2019-01-15 | 2021-06-11 | 中国人民解放军国防科技大学 | Imaging satellite autonomous task planning method based on machine learning |
| CN109767128A (en) * | 2019-01-15 | 2019-05-17 | 中国人民解放军国防科技大学 | imaging satellite autonomous task planning method based on machine learning |
| CN110766284A (en) * | 2019-09-24 | 2020-02-07 | 合肥工业大学 | Multi-star task synthesis method and system |
| CN110728447B (en) * | 2019-10-10 | 2021-03-09 | 合肥工业大学 | Partitioned satellite task planning method for achieving regional target coverage at earliest |
| CN110728447A (en) * | 2019-10-10 | 2020-01-24 | 合肥工业大学 | Partitioned satellite task planning method for achieving regional target coverage at earliest |
| CN111176807A (en) * | 2019-12-09 | 2020-05-19 | 中国空间技术研究院 | Multi-satellite cooperative task planning method |
| CN111176807B (en) * | 2019-12-09 | 2023-12-12 | 中国空间技术研究院 | Multi-star collaborative task planning method |
| CN112016812A (en) * | 2020-08-06 | 2020-12-01 | 中南大学 | Multi-unmanned aerial vehicle task scheduling method, system and storage medium |
| CN112016812B (en) * | 2020-08-06 | 2022-07-12 | 中南大学 | Multi-unmanned aerial vehicle task scheduling method, system and storage medium |
| CN112070370A (en) * | 2020-08-21 | 2020-12-11 | 中南大学 | Relay satellite task planning method, system and storage medium |
| CN112070370B (en) * | 2020-08-21 | 2024-03-08 | 中南大学 | Relay satellite task planning method, system and storage medium |
| CN112613644A (en) * | 2020-12-09 | 2021-04-06 | 合肥工业大学 | Multi-satellite multi-station integrated scheduling method and system |
| CN112737660A (en) * | 2020-12-09 | 2021-04-30 | 合肥工业大学 | Multi-satellite multi-station data downloading scheduling method and system |
| CN112613644B (en) * | 2020-12-09 | 2022-11-04 | 合肥工业大学 | Multi-satellite multi-station integrated scheduling method and system |
| CN113158384A (en) * | 2021-03-03 | 2021-07-23 | 东北石油大学 | Oil and gas pipeline route planning method and system |
| CN113408063B (en) * | 2021-07-20 | 2022-06-21 | 中南大学 | Multi-satellite multi-reconnaissance target-oriented track maneuvering optimization method |
| CN113408063A (en) * | 2021-07-20 | 2021-09-17 | 中南大学 | Multi-satellite multi-reconnaissance target-oriented track maneuver optimization 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 |
| CN113919122A (en) * | 2021-08-05 | 2022-01-11 | 合肥工业大学 | Multi-star task scheduling method and system based on simulated annealing algorithm |
| CN114604443A (en) * | 2022-02-17 | 2022-06-10 | 中国电子科技集团公司第十研究所 | Multi-satellite measurement and control scheduling method and system under real complex constraint condition |
| CN114604443B (en) * | 2022-02-17 | 2023-06-06 | 中国电子科技集团公司第十研究所 | Multi-star measurement and control scheduling method and system under real and complex constraint conditions |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109087023B (en) | 2021-03-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109087023A (en) | A kind of more stars observation layering dispatching method and system based on divide-and-conquer strategy | |
| CN112766813B (en) | Space-day collaborative observation complex task scheduling method | |
| CN112016812B (en) | Multi-unmanned aerial vehicle task scheduling method, system and storage medium | |
| CN111858009B (en) | Task scheduling method of mobile edge computing system based on migration and reinforcement learning | |
| CN108055067B (en) | Multi-satellite online cooperative scheduling method | |
| US20180341894A1 (en) | Innovative satellite scheduling method based on genetic algorithms and simulated annealing and related mission planner | |
| CN109829636B (en) | Satellite emergency task scheduling planning method based on dynamic priority | |
| CN109960544B (en) | Task parallel scheduling method based on data driving type agile satellite | |
| Wu et al. | Ensemble of metaheuristic and exact algorithm based on the divide-and-conquer framework for multisatellite observation scheduling | |
| CN109933842A (en) | A kind of mobile target list star mission planning method based on constraint satisfaction genetic algorithm | |
| CN105225003A (en) | A kind of cuckoo searching algorithm solves the method for UAV multitask investigation decision problem | |
| CN109039428B (en) | Relay satellite single-address antenna scheduling random search method based on conflict resolution | |
| CN112346839A (en) | Associated task scheduling method based on evolutionary algorithm | |
| Dai et al. | Dynamic scheduling for emergency tasks in space data relay network | |
| CN108021433B (en) | Target observation method for multi-satellite cluster | |
| CN113568426B (en) | Satellite cluster collaborative planning method based on multiple satellites and multiple loads | |
| CN113051815A (en) | Agile imaging satellite task planning method based on independent pointer network | |
| Zhou et al. | Multi-satellite scheduling problem with marginal decreasing imaging duration: An improved adaptive ant colony algorithm | |
| Wei et al. | Knowledge-transfer based genetic programming algorithm for multi-objective dynamic agile earth observation satellite scheduling problem | |
| Tahooneh et al. | Using artificial bee colony to solve stochastic resource constrained project scheduling problem | |
| CN114580209A (en) | Multi-satellite cooperative task allocation method and system for non-time-sensitive moving target | |
| Seman et al. | Improving energy aware nanosatellite task scheduling by a branch-cut-and-price algorithm | |
| Tran et al. | Explorations of quantum-classical approaches to scheduling a mars lander activity problem | |
| Turabieh et al. | Incorporating tabu search into memetic approach for enrolment-based course timetabling problems | |
| Marie-Sainte | A new hybrid particle swarm optimization algorithm for real-world university examination timetabling problem |
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 |