CN108416488A - A kind of more intelligent robot method for allocating tasks towards dynamic task - Google Patents
A kind of more intelligent robot method for allocating tasks towards dynamic task Download PDFInfo
- Publication number
- CN108416488A CN108416488A CN201711394738.0A CN201711394738A CN108416488A CN 108416488 A CN108416488 A CN 108416488A CN 201711394738 A CN201711394738 A CN 201711394738A CN 108416488 A CN108416488 A CN 108416488A
- Authority
- CN
- China
- Prior art keywords
- task
- intelligent robot
- chromosome
- time
- intelligent
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/123—DNA computing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Molecular Biology (AREA)
- Strategic Management (AREA)
- General Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Evolutionary Computation (AREA)
- Economics (AREA)
- Genetics & Genomics (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Game Theory and Decision Science (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- General Business, Economics & Management (AREA)
- Physiology (AREA)
- Numerical Control (AREA)
Abstract
The present invention provides a kind of more intelligent robot method for allocating tasks towards dynamic task, mainly solves the problems, such as that task status measurer has the multi-task planning of time-varying characteristics.Including:Dynamic task characteristic parameter is obtained first establishes the characteristic equation of task dotted state amount in conjunction with intelligent robot ability parameter;According to characteristic equation, intelligent robot revenue function is designed;Secondly according to revenue function, genetic algorithm fitness function is designed;Further design genetic algorithm difference selection opertor and local mutation operator, and propose algorithm correcting strategy;Finally genetic algorithm is utilized to generate intelligent robot task allocation plan, completes multi-task planning.Method for allocating tasks proposed by the present invention realizes the quick distribution of dynamic multitask, solves the algorithm chromosome Deadlock, search is avoided to be absorbed in local optimum to obtain system maximum return as target;By multistage allocation strategy, the intelligent robot in system can be given full play to and go to participate in completion task, improve system overall efficiency.
Description
Technical field
The present invention relates to intelligent robot task allocation algorithms technical fields, appoint towards dynamic more particularly, to one kind
More intelligent robot method for allocating tasks of business.
Background technology
In recent years, as people go deep into artificial intelligence and complication system understanding, Intelligent multi-robot system regions
Research abundant achievement is all achieved in terms of theoretical and real system, intelligent robot can replace the mankind severe at some
In the environment of help us to complete certain work, such as aerospace, deep-sea detecting, exploration disaster relief etc., the frequent companion of these environment
With the substance that toxic, anaerobic, high temperature and pressure, intense radiation etc. are harmful, the limit that the mankind can bear is had exceeded.
In current multi-robot system, towards task quantity it is big, task dynamic change, traditional genetic algorithm solves this
The problems such as will produce deadlock when class dynamic task allocation problem, be easily trapped into local optimum and repeated dispensing can not be carried out,
While wasting intelligent robot resource, the needs of real-time can not be met, seriously affect the completeness of task.
Invention content
The present invention provide it is a kind of overcoming the above problem or solve the above problems at least partly towards the more of dynamic task
Intelligent robot method for allocating tasks the characteristics of according to dynamic task allocation problem, proposes a kind of to be based on Revised genetic algorithum
Intelligent multi-robot dynamic task allocation method, changed according to dynamic task quantity of state, design dynamic income model, design is new
Genetic operator, the multistage use genetic algorithm, not only make full use of intelligent robot resource, and improve genetic algorithm reality
Shi Xing, environmental suitability, to efficiently solve the assignment problem of multiple dynamic tasks in environment.
In order to achieve the above object, the present invention provides a kind of Intelligent multi-robot task distribution side towards dynamic task
Method, this method include being:
S1:According to the ability parameter of environment information acquisition intelligent robot, environmental information, task initial characteristics parameter and
Relevant constraint;
S2:Changed according to dynamic task quantity of state, designs dynamic income model;
S3:The financial value that each intelligent robot of distribution of computation tasks obtains;
S4:When task occurs, it is based on improved adaptive GA-IAGA, generates task allocation plan, intelligent robot is according to described
Task allocation plan executes corresponding goal task.
In the step S1, further comprise:
S11:Environment is modeled by establishing coordinate system, multiple intelligence for having and executing task ability are distributed in environment
The task and several static-obstacle things that energy robot, attribute change over time;
S12:According to the information of environment, analysis task distributes factor needed to be considered and sorts out relevant constraint.
Further, in the step S11, the environment is intelligent robot working environment, it is assumed that the intelligence in environment
Energy robot, task point and barrier establish rectangular coordinate system, distribution N (N ∈ Z all in same plane+) a task point, M (M
∈Z+) a intelligent robot and B (B ∈ Z+) a static-obstacle thing.
The approximate coordinate for obtaining each task point is (xj,yj), j=1,2 ..., N, the approximate coordinate (x of intelligent roboti,
yi), i=1,2 ..., the approximate coordinate (x of M and barrierb,yb), b=1,2 ..., B.
Air line distance formula between jth item task and intelligent robot i is:
It is assumed that the coordinate of task point j and task point j ' are respectively (xj,yj), (xj′,yj′), then the distance between 2 points formula
For:
Further, in the step S12, the quantity of state of task point changes over time, in the case where intervening without the external world, task point j
The quantity of state of (j=1,2 ..., N) is expressed as with equation:
In formula:For the amount of state variation in the task j unit interval, αjThe state growth rate of task point j.
Intelligent robot i has executive capability βi, the intelligent robot collection being operated on task j is combined into Λj, the shape of task j
State amount is expressed as with equation:
This is a dynamic task allocation problem.The quantity of state of task changes over time, and in the case where intervening without the external world, it increases
Speed is αj, it is carrying out the shadow of the task (the executive capability summation of intelligent robot) by how many intelligent robot
It rings.
Assuming that the m robot executive capability of t moment performance objective task j and beingWhenWhen, show
The m intelligent robot can not complete the task j, sj(t) trend risen is presented;WhenWhen, show m intelligent machine
Device people can complete the task, sj(t) downward trend is presented.
Further, in the step S12, it includes task point feature parameter and intelligence that system task, which distributes factor needed to be considered,
It can roboting features parameter.
Task point j (j=1,2 ..., N) characteristic parameter includes:The quantity of state S of taskj(t), t moment task status is indicated
Amount, works as t=0, is the original state amount S of taskj(0);Rate of rise αjAnd residing coordinate position (x in the environmentj,yj)。
Intelligent robot i (i=1,2 ..., M) ability parameter includes:Movement velocity vi, indicate intelligent robot i in unit
The path length walked in time;Executive capability βi(βi> 0) and residing coordinate position (x in the environmentj,yj)。
Further, in the step S12, there are all multiple constraints, intelligent robot should be kept away during going to goal task
Exempt to collide, with function H1The case where description intelligent robot collides with barrier:
In formula:Zi(t) indicate intelligent robot i in the location of t moment state, ObIndicate the band of position of barrier,
H1(Zi(t),Zb)=1 indicates that intelligent robot i at least collides with a barrier.
Similarly use function H2The case where colliding between description intelligent robot:
In formula:dsafeSafe distance between intelligent robot.
Therefore, constraints of the intelligent robot in going to object procedure is:
Given ambient intelligence robot executive capability is required with that need to be higher than maximum growth rate in task, with formula table
It is shown as:
When carry out task distribution, the sum of intelligent robot executive capability cannot below correspond to goal task dry without the external world
Growth rate under pre-, otherwise there are global deadlocks for system, are formulated as:
Intelligent robot executive capability summation on expression task j.
The step S2, in order to distinguish different task in the importance of moment t, there are one corresponding incomes for each task tool
Function:
Wherein IjIt is the maximum return that completion task j can be obtained, whendφj(t)/dt > 0, whereinFor
Function phij(t) t increases and incremental time threshold at any time, at this time φj(t) it is that the income that completion task j is obtained becomes at any time
" the rise in price function " changed;Whendφj(t)/dt < 0, whereinFor function phij(t) t increases and successively decreases at any time
Time threshold, φ at this timej(t) it is " discounting function " that income that completion task j is obtained changes over time;When
φj(t)=a, wherein a are the constant value of setting.
Consider that the robot for reaching goal task on the way needs to spend the time, settingMaximum is obtained to complete task j
The interests time, whereinAssuming thatThe time that intelligent robot to reach goal task j at first on the way consumes,
Rise in price function may be configured as Discounting function setup is
Given completion task j can get the formula that maximum return changes with task status amount and be:
In formula, w1For weight coefficient, Sj(0) the original state amount for being task j.
In the step S3, it is assumed that complete the time that task j is consumedThe income of acquisition isIn order to close
Reason distributes to the intelligent robot for executing this task, distributes income according to capability vector, specially:
I-th of intelligent robot executes the income that task j is obtained in t moment and is formulated as:
In formula, xij(t) ∈ { 0,1 } indicates whether i-th of intelligent robot executes task j in t moment, no if being then 1
It is then 0, whereinIndicate the intelligent robot executive capability in t moment execution task j
Summation.
The distribution of income that completion task j is obtained is formulated as to intelligent robot i:
After completing whole system task, the total revenue that i-th of intelligent robot obtains is formulated as:
In formula, IiIt indicates to complete income obtained by entire all task intelligent robot i, xij∈ { 0,1 } indicates i-th
Whether a intelligent robot performs task j, is otherwise 0 if being then 1.
Therefore the total revenue that whole system task obtains is completed to be formulated as:
The target of whole system is to make the final task income of acquisitionIt maximizes, i.e.,
In formula,The time that expression task j is completed.
In the step S4, further comprise:
S41:Genetic parameter is initialized, and is encoded, fitness function is set;
S42:Calculate the fitness value of each population;
S43:Selection opertor, the present invention propose difference examination choose, carry out first it is preselected, secondly formally chosen,
Select S genome of different templates at group k;
S44:Carry out single-point intersection, due to the present invention towards be dynamic task, need to detect every chromosome in real time be
It is no to be absorbed in global deadlock's situation, if taking correcting strategy in the presence of if;
S45:Variation, designs new mutation operator, and on chromosome on the basis of non-deadlock task, retained earnings value is maximum
Gene code and position are constant, remaining gene randomly selects two and swaps position.
S46:Group k+1 is generated, judges whether genetic algorithm evolutionary generation reaches the maximum evolutionary generation of setting, if reaching
It arrives, establishes optimum distributing scheme, intelligent robot starts performance objective task;Otherwise, step S32 is returned to
S47:Once there is task completion, the number of tasks completed can be judged in the presence of the idle intelligent robot of no target operation
Whether reach the number of tasks given in system, if so, otherwise output optimal solution returns to step S31, is calculated again with heredity
Method distributes idle intelligent robot new task, until after system task has all been performed, terminates.
Further, in the step S41, the genetic parameter of initialization includes:The maximum evolutionary generation G of setting, intelligence machine
The characteristic parameter of people's characteristic parameter and task.
Task point j (j=1,2 ..., N) property parameters include:The original state amount S of taskj(0), growth rate αjAnd
Residing approximate coordinate position (x in the environmentj,yj)。
Intelligent robot i (i=1,2 ..., M) characteristic parameter includes:Movement velocity vi, executive capability βiAnd in the environment
Residing approximate coordinate position (xi,yi)
Using machine code, the volume of intelligent robot can be indicated to avoid decoding operate, the loci sequence value of chromosome
Number, word string value indicates to remove the target designation of execution task, forms a kind of mapping relations, it is assumed that there are 6 intelligent robots, 4
Task point, as shown in table 1.
1 chromosome coding mode of table:
It is [231323] to distribute its chromosome in this subtask, it indicates the 1st, the 5th intelligent robot execution task
2, the 2nd, the 4th, the 6th intelligent robot execution task 3, the 3rd intelligent robot performance objective task 1, this dyeing
Task 4 is executed without intelligent robot on body.
Task distribution is carried out every time, is chosen for using fitness function:
Further, in the step S43, selection opertor of the present invention examines admission using difference, candidate before intersecting every time
Number of individuals is more than the chromosome number S of final choice, first carries out examination selection part individual, then in preselected individual just
Formula chooses the intersection number of individuals S met the requirements, and superseded individual enters short-list next time;
Carry out first it is preselected, every time evolve candidate individual number C is examined, performance assessment criteria be fitness value size,
It chooses fitness value H (H > S) individuals in the top and is used as short-list, according to the descending sequence of fitness value;
Then formal to choose, the individual of same encoding gene is removed using No. 1 individual as template, gradually with fitness value
High individual selects S group of individuals of different templates at group P (k), if the population size for the requirement that is not being met as template
Number S then removes individual and fills vacancies in the proper order the quantity that group lacks according to fitness value size order, k=0 when initial.
Here differentiate whether template is similar using similarity, formula uses:
Cii′=kii′/ n (i, i ' ∈ (i, i '≤H) ∧ i ≠ i ') (formula 17)
In formula:In formula:kii′It is chromosome i and the public genic value number of chromosome i ' corresponding positions, n is chromosome coding
String length;The threshold value μ for giving a selection opertor, works as Cii′>=μ indicates that chromosome i is similar to chromosome i ', otherwise dissimilar.
Further, in the step S44, there may be the chromosome of a large amount of deadlocks after intersection, traditional method is to throw
Abandon the chromosome, this method be to weak restricted problem it is feasible, but the present invention towards the problem of there are all multiple constraints, should not adopt
Use conventional method.
When detecting that gene code is absorbed in global deadlock on chromosome, i.e., intelligent robot after gene code on chromosome
It is improper to distribute, and the intelligent robot executive capability total value of each task distribution, which is less than under the task is intervened without the external world, increases speed
Rate can cause all tasks that can not be completed, the present invention use correcting strategy, specifically include for:
Once detecting chromosome, there are global deadlocks, the corresponding coding of intelligent robot on chromosome is counted;
Calculate each task point intelligent robot executive capability and with task rate of rise ratio Pj(0 < j < N) simultaneously
It is ranked up
It willTask on the small intelligent robot of executive capability distribute toTask, ifTask not yet
Dead lock, continue byTask on the small intelligent robot of executive capability distribute toTask ..., untilAppoint
Business Dead lock.
It carries into execution a plan through the above technical solutions, the present invention generates complete set, and obtains all tasks and be completed institute
The time needed, and complete the time of whole system required by task.
Based on the above-mentioned technical proposal, the present invention proposes a kind of genetic algorithm towards dynamic task, mainly solves task
Quantity of state has the multi-task planning problem of time-varying characteristics;This algorithm designs genetic algorithm fitness function, and further design is lost
Propagation algorithm difference selection opertor and local mutation operator, and propose algorithm correcting strategy, which can solve chromosome deadlock
Problem avoids search from being absorbed in local optimum, and by multistage allocation strategy, can give full play to the intelligent robot in system
It goes to participate in completion task, improves system overall efficiency..
Description of the drawings
Fig. 1 is the solution of the present invention flow chart.
Fig. 2 is the environmental model figure of the present invention program.
Specific implementation mode
Below according to attached drawing, the present invention will be described in detail, is a kind of preferred reality in numerous embodiments of the present invention
Apply example.
In a preferred embodiment, a kind of more intelligent robot method for allocating tasks towards dynamic task, the present invention
It is proposed a kind of genetic algorithm towards dynamic task, solving multiple dynamics in environment by cooperative cooperating between intelligent robot appoints
The problem of business.
Intelligent robot cooperative cooperating is presented as:Environment is modeled, the information in environment is obtained, when task occurs,
It is based on Revised genetic algorithum according to the information of acquisition and carries out task distribution, once there is task to be completed, intelligence machine in system
People goes other intelligence that cooperate by communicating the intelligent robot progress task distribution for using Revised genetic algorithum to be idle again
Robot executes unfinished task, until all tasks are completed.
Revised genetic algorithum uses following steps:The parameter set of acquisition problem, encodes chromosome, passes through difference
Operator is chosen in examination, and single-point is taken to intersect, and the chromosome for encountering deadlock is solved by recovery policy, appoints when task space has
Business is completed, and detects whether being completed for task reaches requirement, if so, terminating, exports the time that each task is completed, no
Then, Revised genetic algorithum is used to carry out task distribution to idle intelligent robot again.
Referring to Fig.1, the population number C of genetic algorithm candidate is 50 to preferential genetic algorithm, and pre-selection population number H is 35, just
It is 30 that the population number that formula is chosen, which is S, and maximum iteration is that G is 150, w1=1.5;Intelligent robot number M=8, then n=8,
The executive capability of intelligent robot is respectively:β={ β1,β2,…,β8}={ 0.3,0.4,0.3,0.2,0.3,0.1,0.2,
0.2 }, intelligent robot movement velocity is respectively:vi={ v1,v2,…,v8}={ 2,3,1,2,4,2,1,3 };Task number N=
3 intervene lower growth rate α={ α without the external world1,α2,α3}={ 0.8,0.9,0.7 }, original state amount is Sj(0)={ S1(0),S2
(0),S3(0) }={ 1.5,2,2.5 }, threshold value μ=0.9;The limiting time of setting
The maximum value of task is respectively:
Similarly, I can be acquired2=7.05, I3=5.85;Rise
Valence function may be configured asDiscounting function setup isA=0.5.
Selection opertor carries out pre-selection, then formally selects S satisfied chromosome as group P (k), k when initial first
=0.
Assuming that the encoding gene of chromosome A is [1 213211 2], the encoding gene of chromosome B is [1 312
332 2], then two chromosomes correspond to position and overlap gene number be 3, then CAB=0.375 μ=0.9 <, therefore chromosome
A and B is dissimilar, can intersect.
Intersected using single-point, when detecting that gene code is absorbed in global deadlock on chromosome, the intelligence of each task distribution
Energy robot executive capability and value are less than without the lower growth rate of external world's intervention, can cause all tasks that can not complete, pass through at this time
Recovery policy solves Deadlock.
Assuming that two chromosome A and 1 B gene coding are as shown in table 2:
Parent chromosome A and 1 B gene coding do not form Deadlock, and No. 3 gene code positions of two chromosome occur
Intersect, generates shown in two blank child tables 3:
Table 3:
The blank child of A:
The blank child of B:
Then for the blank child of B, intelligent robot executes the executive capability of task 1 and is:It executes
The executive capability of task 2 and it is:The executive capability of execution task 3 and it is:
Then this chromosome meets deadlock constraint, i.e.,It needs to calculate P at this timej
(0 < j < N), respectively 0.75,1,0.71.Therefore number is needed by 6 distribution task 2 of intelligent robot, Dead lock, most
The blank child of whole B is encoded to:
Assuming that chromogene is encoded to [1 213211 2], intelligent robot executive capability in task 1 and it isIt can similarly acquire
Without generating deadlock in the task 1 of the chromosome, retains corresponding 1 gene of task and position is constant, remaining gene
Random selection two replaces, it is assumed that No. 2 and No. 4 position genes replace, and generating new gene code is:
New gene code is [1 312211 2].
When genetic iteration number be more than setting maximum iteration G, then meet condition, generate this suboptimum task distribution side
Case, intelligent robot go to execute respective goal task according to task allocation plan.
Once there is task to be completed, system can be again that idle intelligent robot is appointed with Revised genetic algorithum
Business distribution until all tasks are completed, and obtain all tasks and is completed time of institute, and completion whole system task
The required time.
The present invention is exemplarily described above in conjunction with attached drawing, it is clear that the present invention implements not by aforesaid way
Limitation as long as using the various improvement of inventive concept and technical scheme of the present invention progress, or is directly applied without improving
In other occasions, within protection scope of the present invention.
Claims (9)
1. a kind of more intelligent robot method for allocating tasks towards dynamic task, which is characterized in that include the following steps:
S1:Obtain ability parameter, environmental information and the goal task initial characteristics parameter and related constraint item of intelligent robot
Part;
S2:Changed according to dynamic task quantity of state, designs dynamic income model;
S3:Based on improved adaptive GA-IAGA, dynamic task allocation scheme is generated, intelligent robot is held according to the task allocation plan
Row task.
2. in the method according to claim 11, which is characterized in that in the step S1, further comprise:
S11:Coordinate system is established to model environment, be distributed in environment multiple intelligent robots with Mission Capability,
The task point and several static-obstacle things that attribute and quantity of state change over time;
S12:According to environmental information, analysis task distributes factor needed to be considered and sorts out relevant constraint.
3. according to the method described in claim 2, it is characterized in that, in the step S11, the environment is intelligent robot
Working environment, distribution N (N ∈ Z in environment+) a task point, M (M ∈ Z+) a intelligent robot and B (B ∈ Z+) a static-obstacle
Object determines the approximate coordinate of each task point, intelligent robot and barrier.
4. according to the method described in claim 2, it is characterized in that, in the step S12, changed according to task dotted state amount special
Point, establishes task dotted state amount model, and analysis task distributes factor needed to be considered, including task point feature parameter, intelligent machine
Device people ability parameter and static-obstacle thing property parameters.
Task j (j=1,2 ..., N) point feature parameter includes:Original state amount Sj(0), rate of rise αjAnd institute in the environment
Approximate coordinate position (the x at placej,yj)。
Intelligent robot i (i=1,2 ..., M) ability parameter includes:Movement velocity vi, to execute task amount in the unit interval be βiWith
And residing approximate coordinate position (x in the environmenti,yi);
Assuming that the quantity of state of task is variation, and in the case where intervening without the external world, the quantity of state equation of task j (j=1,2 ..., N)
It is expressed as:
In formula:For the amount of state variation in the task j unit interval, αjFor the state growth rate of task point j.
The intelligent robot collection being operated on task j is combined into Λj, in the case where intervening without the external world, the state of task j (j=1,2 ..., N)
Amount is expressed as with equation:
From the point of view of single intelligent robot, each intelligent robot needs avoid colliding with other intelligent robots,
Also it to avoid colliding with barrier, with function H1The case where description intelligent robot collides with barrier:
In formula:Zi(t) indicate intelligent robot i in the location of t moment state, ObIndicate the band of position of barrier, H1(Zi
(t),Zb)=1 indicates that intelligent robot i at least collides with a barrier.
Similarly use function H2The case where colliding between description intelligent robot:
In formula:dsafeSafe distance between intelligent robot.
Therefore, constraints of intelligent robot during going to goal task is:
5. in the method according to claim 11, which is characterized in that in the step S2, further comprise for:
In order to distinguish different goal tasks in the importance of moment t, there are one corresponding revenue function I for each task tooljφj
(t), wherein IjIt is the maximum return that completion task j can be obtained, whendφj(t)/dt > 0, whereinFor function
φj(t) t increases and incremental deadline at any time, at this time φj(t) it is that the income that completion task j is obtained changes over time
" rise in price function ";Whendφj(t)/dt < 0, whereinFor function phij(t) at any time t increase and the cut-off successively decreased
Time, at this time φj(t) it is " discounting function " that income that completion task j is obtained changes over time;Whenφj(t)
=a, wherein a are the constant value of setting.The target of whole system task is so that the ultimate yield obtainedIt maximizes,
I.e.WhereinTo complete the time that task j needs.
Consider that the intelligent robot for reaching goal task on the way needs to spend the time, settingMaximum is obtained to complete task j
The time of interests, whereinAssuming thatIntelligent robot to reach goal task j at first on the way consume when
Between, rise in price function may be configured asDiscounting function setup is
6. in the method according to claim 11, which is characterized in that in the step S3, further comprise for:
S31:Genetic parameter is initialized, and is encoded, fitness function is set;
S32:Calculate the fitness value of each population;
S33:Selection opertor, the present invention propose that difference examination is chosen, and carry out first preselected, then are formally chosen, selection is not
The S genome with template is at group P (k);
S34:Carry out single-point intersection, due to the present invention towards be dynamic task, need to detect whether every chromosome falls into real time
Enter global deadlock's situation, if taking correcting strategy in the presence of if;
S35:Variation designs mutation operator, no longer some gene position of random variation, but non-deadlock task is basic on chromosome
On, retained earnings is worth maximum gene code and position is constant, and two progress place-exchanges are randomly selected in remaining gene,
The deadlock that variation is brought can effectively be prevented, while the blindness of variation can be reduced again;
S36:Group k+1 is generated, judges whether genetic algorithm evolutionary generation reaches the maximum evolutionary generation of setting, if reaching, really
Vertical optimum distributing scheme, intelligent robot start performance objective task;Otherwise, step S32 is returned to;
S37:Once there is task completion, whether the number of tasks completed can be judged in the presence of the idle intelligent robot of no goal task
Reach the number of tasks given in system, if so, otherwise output optimal solution returns to step S31, uses genetic algorithm pair again
Idle intelligent robot distributes new task, until after system task has all been performed, terminates.
7. in the method according to claim 11, which is characterized in that in the step S32, when each task has been performed,
The income of acquisition needs reasonably to distribute to the intelligent robot for executing this task, using distribution according to work.Assuming that i-th of intelligence
Robot executes task j in t moment, then i-th of intelligent robot income obtained by t moment is formulated as:
In formula, xij(t) ∈ { 0,1 } indicate i-th of intelligent robot whether execute task j in t moment, if then be 1, otherwise for
0, wherein Indicate that the intelligent robot executive capability that task j is executed in t moment is total
With,Indicate the income that completion task j is obtained.
The total revenue that intelligent robot i is obtained after completion task j is Iij, it is formulated as:
After completing whole system task, i-th of intelligent robot obtains total revenue and is formulated as:
In formula:IiIncome, x obtained by intelligent robot i after all tasks of expression completionij∈ { 0,1 } indicates intelligent robot i
Task j whether is performed, is otherwise 0 if being then 1.Fitness value can be calculated by fitness function, therefore fitness letter
Number can be chosen for:
In formula:I is the total revenue that all robots obtain after completing all tasks.
8. in the method according to claim 11, which is characterized in that in the step S33, design new selection opertor, adopt
Difference examination is taken to choose, candidate number of individuals is more than final choice chromosome number S before intersecting every time, first carries out examination selection one
Divide individual, the intersection number of individuals S met the requirements is then formally chosen in preselected individual, superseded individual enters next time
Short-list is as follows:
S81:It is preselected, it evolves every time and candidate individual number C is examined, performance assessment criteria is fitness value size, chooses and adapts to
Angle value H in the top (H > S) individuals are used as short-list, according to the descending sequence of fitness value;
S82:Formal selection, first using No. 1 individual as the individual of template removal and its similar genes value coding, gradually with fitness
It is worth high individual as template, removes the individual of its similar encoding gene, selects S genome of different templates at group P
(k), it if the population size number S for the requirement that is not being met, removes individual and fills vacancies in the proper order institute of group according to fitness value size order
Scarce quantity, k=0 when initial, wherein similarity formula are used:Cii′=kii′/ n (i, i ' ∈ (i, i '≤H) ∧ i ≠ i '), formula
In:kii′It is chromosome i and the public genic value number of chromosome i ' corresponding positions, n is chromosome coding string length;Setting one
The threshold value μ of selection opertor, works as Cii′>=μ indicates that chromosome i is similar to chromosome i ', otherwise dissimilar.
9. in the method according to claim 11, which is characterized in that in the step S34, when detecting base on chromosome
When being absorbed in global deadlock because of coding, indicate that the distribution of intelligent robot task is improper after gene code on chromosome, each task point
The sum of intelligent robot executive capability matched is respectively less than task growth rate αj, it can cause all tasks that can not be completed, because
This present invention use correcting strategy, specifically include for:
S91:Once detecting chromosome, there are global deadlocks, the corresponding coding of intelligent robot on chromosome is counted;
S92:Calculate each task point intelligent robot executive capability and with task growth rate ratio Pj(0 < j < N), and
It is ranked up
S93:It willThe intelligent robot of executive capability minimum is distributed in corresponding taskCorresponding task, ifIt is right
Answering for task there are no Dead lock, continue byThe intelligent robot of current executive capability minimum is distributed in corresponding taskTask ..., untilTask quit deadlock.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711394738.0A CN108416488B (en) | 2017-12-21 | 2017-12-21 | Dynamic task-oriented multi-intelligent-robot task allocation method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711394738.0A CN108416488B (en) | 2017-12-21 | 2017-12-21 | Dynamic task-oriented multi-intelligent-robot task allocation method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108416488A true CN108416488A (en) | 2018-08-17 |
CN108416488B CN108416488B (en) | 2022-05-03 |
Family
ID=63125617
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711394738.0A Active CN108416488B (en) | 2017-12-21 | 2017-12-21 | Dynamic task-oriented multi-intelligent-robot task allocation method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108416488B (en) |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109460294A (en) * | 2018-10-13 | 2019-03-12 | 中国人民解放军战略支援部队信息工程大学 | A kind of reconfigurable task layout method based on area-efficient and dynamic genetic algorithm |
CN109886574A (en) * | 2019-02-20 | 2019-06-14 | 哈尔滨工程大学 | A kind of multi-robot Task Allocation based on improvement threshold method |
CN109986563A (en) * | 2019-05-01 | 2019-07-09 | 湖南大学 | A kind of multiple mobile robot's work compound method and system |
CN110046795A (en) * | 2019-03-01 | 2019-07-23 | 斯坦德机器人(深圳)有限公司 | The method for allocating tasks and device of robot |
CN110084512A (en) * | 2019-04-26 | 2019-08-02 | 河海大学常州校区 | A kind of multi-robot Task Allocation towards intelligent warehousing system |
CN110501905A (en) * | 2019-08-27 | 2019-11-26 | 中国人民解放军国防科技大学 | Multi-agent system self-adaptive method and system based on packing model |
CN111027875A (en) * | 2019-12-17 | 2020-04-17 | 鲁东大学 | Intelligent warehousing multi-robot task allocation method based on self-adaptive task pool |
CN111325454A (en) * | 2020-02-11 | 2020-06-23 | 广州地铁设计研究院股份有限公司 | Designer role control method |
CN112181608A (en) * | 2019-07-03 | 2021-01-05 | 中南大学 | Distributed distribution algorithm for multipoint dynamic aggregation tasks based on local information |
CN113039525A (en) * | 2018-12-04 | 2021-06-25 | 阿菲尼帝有限公司 | Techniques for behavioral pairing in a multi-stage task distribution system |
CN113050549A (en) * | 2019-12-26 | 2021-06-29 | 炬星科技(深圳)有限公司 | Method, device and storage medium for freely networking robot to execute task |
CN114137955A (en) * | 2021-10-26 | 2022-03-04 | 中国人民解放军军事科学院国防科技创新研究院 | Multi-robot rapid collaborative map building method based on improved market method |
CN114330978A (en) * | 2021-11-11 | 2022-04-12 | 深圳大学 | Air-ground robot task dynamic allocation method, storage medium and terminal equipment |
CN114924513A (en) * | 2022-06-07 | 2022-08-19 | 中迪机器人(盐城)有限公司 | Multi-robot cooperative control system and method |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070087756A1 (en) * | 2005-10-04 | 2007-04-19 | Hoffberg Steven M | Multifactorial optimization system and method |
CN105069530A (en) * | 2015-08-13 | 2015-11-18 | 肇庆学院 | Multi-robot task allocation method based on multi-objective optimization |
US9446512B1 (en) * | 2014-03-26 | 2016-09-20 | Stc.Unm | iAnt swarm robotic platform and evolutionary algorithms |
CN106020230A (en) * | 2016-05-20 | 2016-10-12 | 武汉科技大学 | Task distribution method for multiple unmanned planes within constraint of energy consumption |
CN106654987A (en) * | 2016-11-18 | 2017-05-10 | 华北电力大学(保定) | Power line multi-robot collaborative inspection method |
CN106875090A (en) * | 2017-01-09 | 2017-06-20 | 中南大学 | A kind of multirobot distributed task scheduling towards dynamic task distributes forming method |
CN107168054A (en) * | 2017-05-10 | 2017-09-15 | 沈阳工业大学 | Multi-robotic task is distributed and paths planning method |
-
2017
- 2017-12-21 CN CN201711394738.0A patent/CN108416488B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070087756A1 (en) * | 2005-10-04 | 2007-04-19 | Hoffberg Steven M | Multifactorial optimization system and method |
US9446512B1 (en) * | 2014-03-26 | 2016-09-20 | Stc.Unm | iAnt swarm robotic platform and evolutionary algorithms |
CN105069530A (en) * | 2015-08-13 | 2015-11-18 | 肇庆学院 | Multi-robot task allocation method based on multi-objective optimization |
CN106020230A (en) * | 2016-05-20 | 2016-10-12 | 武汉科技大学 | Task distribution method for multiple unmanned planes within constraint of energy consumption |
CN106654987A (en) * | 2016-11-18 | 2017-05-10 | 华北电力大学(保定) | Power line multi-robot collaborative inspection method |
CN106875090A (en) * | 2017-01-09 | 2017-06-20 | 中南大学 | A kind of multirobot distributed task scheduling towards dynamic task distributes forming method |
CN107168054A (en) * | 2017-05-10 | 2017-09-15 | 沈阳工业大学 | Multi-robotic task is distributed and paths planning method |
Non-Patent Citations (1)
Title |
---|
黎萍 等: ""多机器人系统任务分配的研究进展"", 《计算机工程与应用》 * |
Cited By (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109460294A (en) * | 2018-10-13 | 2019-03-12 | 中国人民解放军战略支援部队信息工程大学 | A kind of reconfigurable task layout method based on area-efficient and dynamic genetic algorithm |
CN113039525A (en) * | 2018-12-04 | 2021-06-25 | 阿菲尼帝有限公司 | Techniques for behavioral pairing in a multi-stage task distribution system |
CN109886574A (en) * | 2019-02-20 | 2019-06-14 | 哈尔滨工程大学 | A kind of multi-robot Task Allocation based on improvement threshold method |
CN109886574B (en) * | 2019-02-20 | 2023-02-14 | 哈尔滨工程大学 | Multi-robot task allocation method based on improved threshold method |
CN110046795A (en) * | 2019-03-01 | 2019-07-23 | 斯坦德机器人(深圳)有限公司 | The method for allocating tasks and device of robot |
CN110046795B (en) * | 2019-03-01 | 2021-11-05 | 斯坦德机器人(深圳)有限公司 | Task allocation method and device for robot |
CN110084512A (en) * | 2019-04-26 | 2019-08-02 | 河海大学常州校区 | A kind of multi-robot Task Allocation towards intelligent warehousing system |
CN110084512B (en) * | 2019-04-26 | 2022-04-22 | 河海大学常州校区 | Multi-robot task allocation method for intelligent warehousing system |
CN109986563A (en) * | 2019-05-01 | 2019-07-09 | 湖南大学 | A kind of multiple mobile robot's work compound method and system |
CN109986563B (en) * | 2019-05-01 | 2022-03-15 | 湖南大学 | Multi-mobile-robot cooperative operation method and system |
CN112181608A (en) * | 2019-07-03 | 2021-01-05 | 中南大学 | Distributed distribution algorithm for multipoint dynamic aggregation tasks based on local information |
CN112181608B (en) * | 2019-07-03 | 2023-10-31 | 中南大学 | Distributed distribution method for multipoint dynamic aggregation tasks based on local information |
CN110501905B (en) * | 2019-08-27 | 2022-02-08 | 中国人民解放军国防科技大学 | Multi-agent system self-adaptive method and system based on packing model |
CN110501905A (en) * | 2019-08-27 | 2019-11-26 | 中国人民解放军国防科技大学 | Multi-agent system self-adaptive method and system based on packing model |
CN111027875B (en) * | 2019-12-17 | 2023-09-26 | 鲁东大学 | Intelligent warehouse multi-robot task allocation method based on self-adaptive task pool |
CN111027875A (en) * | 2019-12-17 | 2020-04-17 | 鲁东大学 | Intelligent warehousing multi-robot task allocation method based on self-adaptive task pool |
CN113050549A (en) * | 2019-12-26 | 2021-06-29 | 炬星科技(深圳)有限公司 | Method, device and storage medium for freely networking robot to execute task |
CN113050549B (en) * | 2019-12-26 | 2022-06-10 | 炬星科技(深圳)有限公司 | Method, equipment and storage medium for executing task by robot capable of freely networking |
CN111325454A (en) * | 2020-02-11 | 2020-06-23 | 广州地铁设计研究院股份有限公司 | Designer role control method |
CN114137955A (en) * | 2021-10-26 | 2022-03-04 | 中国人民解放军军事科学院国防科技创新研究院 | Multi-robot rapid collaborative map building method based on improved market method |
CN114330978A (en) * | 2021-11-11 | 2022-04-12 | 深圳大学 | Air-ground robot task dynamic allocation method, storage medium and terminal equipment |
CN114924513A (en) * | 2022-06-07 | 2022-08-19 | 中迪机器人(盐城)有限公司 | Multi-robot cooperative control system and method |
Also Published As
Publication number | Publication date |
---|---|
CN108416488B (en) | 2022-05-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108416488A (en) | A kind of more intelligent robot method for allocating tasks towards dynamic task | |
CN106875090B (en) | Dynamic task-oriented multi-robot distributed task allocation forming method | |
Coello Coello et al. | MOSES: A multiobjective optimization tool for engineering design | |
Gowda et al. | Symbolic clustering using a new similarity measure | |
Coello | An updated survey of GA-based multiobjective optimization techniques | |
CN104572297B (en) | A kind of Hadoop job scheduling methods based on genetic algorithm | |
Saif et al. | Multi-objective artificial bee colony algorithm for simultaneous sequencing and balancing of mixed model assembly line | |
CN104516785B (en) | A kind of cloud computing resources dispatch system and method | |
CN109409773A (en) | A kind of earth observation resource dynamic programming method based on Contract Net Mechanism | |
Melvix | Greedy politics optimization: Metaheuristic inspired by political strategies adopted during state assembly elections | |
Rodzin et al. | New computational models for big data and optimization | |
Rodzin et al. | Theory of bioinspired search for optimal solutions and its application for the processing of problem-oriented knowledge | |
Teo et al. | Automatic generation of controllers for embodied legged organisms: A Pareto evolutionary multi-objective approach | |
CN109102203A (en) | A kind of Target Assignment optimization method based on more string chromosomal inheritance algorithms | |
Samanta et al. | Bioeconomic modelling of a three-species fishery with switching effect | |
Protalinskii et al. | Optimal strategy synthesis for a group of mobile robots with variable structure | |
CN115689116A (en) | Resource-construction period-cost comprehensive optimization method based on improved multi-target particle swarm | |
CN112783210B (en) | Multi-target control parameter optimization method of unmanned aerial vehicle cluster control system | |
Liu | An analysis of particle swarm optimization of multi-objective knapsack problem | |
Thomaidis | A soft computing approach to enhanced indexation | |
Cheng et al. | Integer Encoding Genetic Algorithm for Optimizing Redundancy Allocation of Series-parallel Systems. | |
Hao et al. | Improved algorithms to plan missions for agile earth observation satellites | |
Wu et al. | Intelligent automated control of life support systems using proportional representations | |
Marcolino et al. | Agents vote for the environment: Designing energy-efficient architecture | |
Rostovsky et al. | The model of decision-making of high-tech company on the formation of the optimal portfolio. The results of numerical modeling |
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 |