CN108171394B - Multi-robot task allocation method based on layered structure and resource integration - Google Patents

Multi-robot task allocation method based on layered structure and resource integration Download PDF

Info

Publication number
CN108171394B
CN108171394B CN201711168423.4A CN201711168423A CN108171394B CN 108171394 B CN108171394 B CN 108171394B CN 201711168423 A CN201711168423 A CN 201711168423A CN 108171394 B CN108171394 B CN 108171394B
Authority
CN
China
Prior art keywords
resource
manager
task
layer
robot
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.)
Active
Application number
CN201711168423.4A
Other languages
Chinese (zh)
Other versions
CN108171394A (en
Inventor
曹志强
任亮
于莹莹
庞磊
谭民
喻俊志
周超
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Institute of Automation of Chinese Academy of Science
Original Assignee
Institute of Automation of Chinese Academy of Science
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Institute of Automation of Chinese Academy of Science filed Critical Institute of Automation of Chinese Academy of Science
Priority to CN201711168423.4A priority Critical patent/CN108171394B/en
Publication of CN108171394A publication Critical patent/CN108171394A/en
Application granted granted Critical
Publication of CN108171394B publication Critical patent/CN108171394B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06315—Needs-based resource requirements planning or analysis
    • 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)
  • Manipulator (AREA)

Abstract

The invention relates to the field of multi-robot task allocation, in particular to a multi-robot task allocation method based on a layered structure and resource integration. The multi-robot task allocation method disclosed by the invention is used for allocating tasks based on a pyramid-shaped layered structure. When a pyramid is constructed, the robot is used as the bottom layer from bottom to top, the rest layers are managers, the number of the managers is reduced layer by layer, and only one manager is arranged at the top layer; then, based on each robot satisfying the task constraint condition, the resources owned by each manager are calculated layer by layer from bottom to top. When tasks are distributed, all lowest-level managers capable of meeting task requirements are found from top to bottom, robots which are directly or indirectly managed by all the lowest-level managers and meet task constraint conditions are combined, and a robot alliance with the best matching degree is selected as a robot alliance for executing the tasks. The calculation method of the invention has high efficiency and can effectively improve the real-time performance of task allocation.

Description

Multi-robot task allocation method based on layered structure and resource integration
Technical Field
The invention relates to the field of multi-robot task allocation, in particular to a multi-robot task allocation method based on a layered structure and resource integration.
Background
In recent years, multi-robot systems have attracted general attention due to their flexibility, parallelism, robustness, and the like. Research of the multi-robot system comprises aspects of group architecture, task allocation, coordinated cooperation mechanism and the like, wherein the task allocation is important research content.
The task allocation result directly influences the quality of task completion of the multi-robot system, and the task is allocated to a proper robot to be executed, so that the optimization of system resources is facilitated. Early conventional task assignment methods considered that each task could be completed by one robot, which is generally applicable to situations where tasks are relatively simple. With the increase of task complexity, the resource of a single robot cannot meet the requirement of the task, and at this time, the task needs to be distributed to a union formed by a plurality of robots. However, due to the high computational complexity of the existing alliance-based task allocation method, the real-time performance of task allocation in a large-scale multi-robot system still needs to be improved.
In this situation, the effectiveness of the human social hierarchical management mechanism provides a solution. Inspired by the human society hierarchical management mechanism, researchers begin to pay attention to the hierarchical construction of multi-robot systems, and then task allocation is carried out under a hierarchical structure. However, the existing research on the multi-robot task allocation method based on the hierarchical structure has not effectively solved the problem of resource integration under the hierarchical structure, which will affect the efficiency of task allocation. Meanwhile, the real-time performance of the existing multi-robot task allocation method based on the layered structure is further improved.
Disclosure of Invention
In order to solve the problems in the prior art, the invention provides a multi-robot task allocation method based on a layered structure and resource integration, which improves the efficiency of multi-robot task allocation and effectively improves the instantaneity of allocation.
The invention provides a multi-robot task allocation method based on a layered structure and resource integration, which comprises the following steps:
step S10, based on the pyramid layered structure, according to the task constraint conditions and the resource vectors of each robot, screening out the robots meeting the task constraint conditions from the bottom layer of the pyramid;
step S20, calculating resource vectors owned by each manager in each layer of the pyramid-shaped hierarchical structure from bottom to top according to the robots meeting the task constraint conditions and the resource vectors of each robot;
step S30, according to the task demand resource vector and the resource vector owned by each manager, each lowest-level manager meeting the task demand in the pyramid-shaped hierarchical structure is found from top to bottom;
step S40, searching the robots which are directly or indirectly managed by the lowest-level managers meeting the task requirements and meet the task constraint conditions, selecting a robot combination which can meet the task requirements as a candidate robot alliance, and further forming a candidate robot alliance set;
step S50, selecting the robot alliance with the best matching degree from the candidate robot alliance set as the robot alliance which is selected to execute the task;
wherein,
the pyramid-shaped layered structure is constructed by the following steps:
step A10, taking all robots as the bottom layer of the pyramid, grouping the robots according to the preset robot grouping member number, and respectively setting a manager for each group of robots;
step A20, grouping managers according to the number of preset manager grouping members, and setting a superior manager for each group of managers; grouping the superior managers according to the preset number of the manager grouping members, and respectively setting the superior managers for each group; and so on until the top layer of the pyramid has only one manager;
the administrator is a computer.
Preferably, step a10 is specifically:
step A11, taking the robots with the total number of robotNum as the bottom layer of the pyramid, and recording the number of layers as j being 1; grouping the robots, and if (robotNum% P) is 0, then grouping the number of groups1robotNum/P, P robots per group; otherwise, group number groupNum1robotNum/P +1, 1 st to groupNum1-1 groups of P robots, the first group Num1In the group, robotNum- (groupNum)1-1)*P robots;
step A12, setting a manager for each group of robots at the bottom layer as a member of the layer 2, and calling each robot in the group managed by the manager as a subordinate member of the manager; the number of managers is recorded as mangagerNum2＝groupNum1；
Wherein,
p is the preset robot group member number; groupNum1The group number of the grouped robots in the 1 st layer of the pyramid-shaped layered structure; mangagerNum2The number of managers in layer 2.
Preferably, step a20 is specifically:
step A21, the layer number j is 2;
step A22, grouping the managers at the j layer, if (managerNum)jWhen% Q) is 0, the number of groups is groupNumj＝managerNumjQ, Q managers in each group; otherwise, group number groupNumj＝managerNumj(ii)/Q +1, 1 st to groupNumj-Q managers in each group in group 1, the first group NumjThere is mangagerNum in the groupj-(groupNumj-1)*Q managers;
step A23, setting a superior manager for each group manager of the jth layer as a member of the jth +1 layer, and calling each manager in the group managed by the superior manager as a subordinate member of the superior manager; superior manager number managerNumj+1＝groupNumj；
Step a24, if j is j +1, if the number of j-level managers is greater than 1, go to step a 22; otherwise, go to step A25;
step A25, recording the total layer number of the pyramid as N ═ j;
wherein,
q is the preset administrator group member number; groupNumjThe group number of the managers at the j-th layer in the pyramid-shaped hierarchical structure is grouped; mangagerNumjThe number of the managers in the j-th layer.
Preferably, the robot meeting the task constraint condition is a robot in an idle state and having a resource vector greater than or equal to the task constraint resource vector.
Preferably, the comparison between the resource vectors is performed by comparing each resource therein, if the value of each resource in the resource vector f (a) is greater than or equal to the value of the corresponding resource in the resource vector f (b), f (a) is greater than or equal to f (b);
wherein, f (a) and f (b) are two different resource vectors in the resource space, respectively; the resource space is described by F, whose dimension is dim (F), and the resource vector in F is represented by F ═ F1f2...fdim(F)]，fkA value representing the kth resource of the resource vector,
Figure BDA0001476734590000031
1,2,. d, (f); f, (a) and F (b) have dimensions dim (F), F (a) and F (b) are respectively:
F(a)＝[f1(a)f2(a)…fdim(F)(a)]
F(b)＝[f1(b)f2(b)…fdim(F)(b)]。
preferably, step S20 is specifically:
based on the robot meeting the task constraint conditions, performing resource integration from bottom to top in the sequence from the 2 nd layer, the 3 rd layer to the Nth layer, and updating the manager layer by layer
Figure BDA0001476734590000041
Resource vector of
Figure BDA0001476734590000042
Wherein,
for each manager at the layer 2, integrating the resource vectors of the robots meeting the task constraint conditions directly managed by the manager according to the following formula, thereby realizing the update of the resource vectors of the managers:
Figure BDA0001476734590000043
Figure BDA0001476734590000044
as managers in layer 2
Figure BDA0001476734590000045
Resource vector of (2), subscript g2、i2Layer number, group number and serial number of the layer manager;
Figure BDA0001476734590000046
for robots in layer 1
Figure BDA0001476734590000047
Resource vector of (1), subscript g1、i1Respectively is the layer number, group number and serial number of the robot; omegarlThe method comprises the steps of (1) collecting all robots meeting task constraint conditions;
for each manager from the layer 3 to the layer N, integrating the resource vectors of subordinate members directly managed by the manager according to the following formula, and further updating the resource vector of the manager:
Figure BDA0001476734590000048
Figure BDA0001476734590000049
is the nth manager
Figure BDA00014767345900000410
Resource vector of (3), subscripts n, gn、inLayer number, group number and serial number of the layer manager; n is 3, 4.
Preferably, the integration of the resource vector is accomplished by integrating each resource;
for each resource, adopting a corresponding integration rule according to the type of the resource; the resource types include: accumulation type, maximum type, minimum type; the integration rule of the accumulation type resources is to sum corresponding resources; the integration rule of the maximum value type resource is to find the maximum value of the corresponding resource; the integration rule of the minimum value type resource is to find the minimum value of the corresponding resource.
Preferably, step S30 is specifically:
comparing the task demand resource vector with the resource vectors owned by the managers in a top-down order from the Nth layer of the pyramid-shaped hierarchical structure; when the resource vector owned by a manager is larger than or equal to the task demand resource vector, indicating that the manager meets the task demand, continuously comparing the task demand resource vector with the resource vectors of all the subordinate members directly managed by the manager, if the subordinate members meeting the task demand still exist, repeating the comparison process until the level 2, and taking the level 2 members meeting the task demand as the lowest-level manager; and if a manager positioned at the 3 rd layer or higher meets the task requirement but the subordinate members directly managed by the manager do not meet the task requirement and the lowest-layer manager meeting the task requirement does not exist, taking the manager as the lowest-layer manager.
Preferably, step S40 is specifically:
step S41, judging the number of layers of each manager at the lowest layer meeting the task requirement, and if the manager at the 2 nd layer is in the step S42; otherwise, go to step S43;
step S42, finding out all robots meeting task constraint conditions directly managed by the lowest-level manager; and to all satisfied task constraintsThe robots with the conditions are exhaustive of all possible combinations, when a resource vector of one robot combination is larger than or equal to a task demand resource vector, the robot combination is shown to meet the task demand, and the robot combination meeting the task demand is used as a candidate robot alliance ATAdded to the set of candidate robot unions omegarcPerforming the following steps;
step S43, finding out all robots satisfying task constraint conditions indirectly managed by the lowest-level manager, and combining all robots satisfying task constraint conditions as candidate robot alliance ATAdded to the set of candidate robot unions omegarcIn (1).
Preferably, step S50 is specifically:
set omega based on candidate robot alliancercSelecting the constraint condition satisfying the matching degree
σ1≤|F(T),F(AT)|≤σ2
And the robot alliance with the best matching degree:
Figure BDA0001476734590000051
wherein, | F (T), F (A)T) I is task demand resource vector F (T) and candidate robot alliance ATResource vector F (A)T) The matching degree between the two is measured by the weighted Euclidean distance between the two; sigma1And σ2Respectively setting a preset minimum matching degree threshold value and a preset maximum matching degree threshold value;
Figure BDA0001476734590000052
is the best league of robots to be assigned to perform task T.
Preferably, the method for calculating the matching degree between the resource vectors is as follows:
Figure BDA0001476734590000053
F(a)＝[f1(a)f2(a)…fdim(F)(a)]
F(b)＝[f1(b)f2(b)…fdim(F)(b)]
where F is used to describe the resource space, its dimension is dim (F), and the resource vector in F can be expressed as F ═ F1f2...fdim(F)]；fkA value representing the kth resource of the resource vector,
Figure BDA0001476734590000061
1,2,. d, (f); f (a) and F (b) are two different resource vectors in resource space F, respectively.
The invention has the beneficial effects that:
the multi-robot task allocation method disclosed by the invention is used for allocating tasks based on a pyramid-shaped layered structure. When a pyramid is constructed, the robot is used as the bottom layer from bottom to top, the rest layers are managers, the number of the managers is reduced layer by layer, and only one manager is arranged at the top layer; then, based on each robot satisfying the task constraint condition, the resources owned by each manager are calculated layer by layer from bottom to top. When distributing tasks, searching managers capable of meeting task requirements layer by layer from top to bottom until the lowest-layer manager capable of meeting the task requirements is found, combining robots which are directly or indirectly managed by the managers and meet task constraint conditions, and selecting a robot alliance with the best matching degree as a robot alliance for executing the tasks. The invention effectively solves the problem of resource integration under a layered structure, and quickly determines the optimal robot alliance to be dispatched to execute the task by resource integration from bottom to top and resource comparison from top to bottom in the layered structure.
Drawings
FIG. 1 is a schematic flow chart diagram of an embodiment of a multi-robot task assignment method of the present invention;
FIG. 2 is a schematic view of a pyramidal layered structure according to an embodiment of the present invention.
Detailed Description
Preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only for explaining the technical principle of the present invention, and are not intended to limit the scope of the present invention.
Multi-robot task assignment can be viewed as a matching process between the resources owned by the multi-robot system and the resources required by the task. We describe the resource space with F, whose dimension is dim (F), and the resource vector in F can be expressed as F ═ F1f2...fdim(F)]Wherein f iskA value of the k-th resource of the resource vector, k being 1,2kHas a numerical value range of
Figure BDA0001476734590000062
For a multi-robot system, resources owned by an individual robot and resources required by a task are described by using resource vectors in a resource space F, and in order to avoid selecting a robot with inappropriate performance, the resource constraint of the task on the individual robot is considered and the resource constraint is represented by using the task constraint resource vectors.
When the individual robot does not have a certain resource, the value of the corresponding resource item in the resource vector is 0; when a certain resource is not needed in the task requirement, the value of the corresponding resource item in the task requirement resource vector is 0; when the task does not constrain a certain resource of the individual robot, the value of the corresponding resource item in the task constraint resource vector is 0. The robot resource vector, the task demand resource vector and the task constraint resource vector all have the same dimension, and correspond to the same resource in each dimension.
The resource vectors can be integrated, and the integration between the resource vectors is completed by integrating each resource. Each resource has a corresponding integration rule according to the resource type. The resource types include accumulation type resource, maximum value type resource and minimum value type resource. The integration rule of the accumulation type resources is to sum the corresponding resources; the integration rule of the maximum value type resource is to find the maximum value of the corresponding resource; the integration rule of minimum value type resources is to find the most corresponding resourcesA small value. Resource vector F (a) of dimension 3s) And F (b)s) For example, the two resource vectors are shown in equations (1) and (2), respectively:
F(as)＝[f1(as)f2(as)f3(as)] (1)
F(bs)＝[f1(bs)f2(bs)f3(bs)] (2)
the 1 st, 2 nd and 3 rd resources in the resource vector are respectively accumulation type resource, maximum value type resource and minimum value type resource, then F (a)s) And F (b)s) Is F (a)s)+F(bs) As shown in equation (3):
F(as)+F(bs)＝[f1(as)+f1(bs)max(f2(as),f2(bs))min(f3(as),f3(bs))] (3)
wherein, max (f)2(as),f2(bs) To find f2(as) And f2(bs) Maximum value of (d), min (f)3(as),f3(bs) To find f3(as) And f3(bs) Is measured.
The resource vectors may also be compared, and f (a) and f (b) are both resource vectors with dimension dim (f), and the two resource vectors are respectively shown in formulas (4) and (5):
F(a)＝[f1(a)f2(a)…fdim(F)(a)] (4)
F(b)＝[f1(b)f2(b)…fdim(F)(b)] (5)
when the value of each resource in F (a) is greater than or equal to the value of the corresponding resource in F (b), F (a) is greater than or equal to F (b), as shown in formula (6):
Figure BDA0001476734590000081
the matching degree between the resource vectors can be expressed by weighted euclidean distances | f (a), f (b) |, and the calculation method is shown in formula (7):
Figure BDA0001476734590000082
the embodiment of the multi-robot task allocation method based on the layered structure and the resource integration, as shown in fig. 1, comprises the following steps:
step S10, based on the pyramid layered structure, according to the task constraint conditions and the resource vectors of each robot, screening out the robots meeting the task constraint conditions from the bottom layer of the pyramid;
step S20, calculating resource vectors owned by each manager in each layer of the pyramid-shaped hierarchical structure from bottom to top according to the robots meeting the task constraint conditions and the resource vectors of each robot;
step S30, according to the task demand resource vector and the resource vector owned by each manager, each lowest-level manager meeting the task demand in the pyramid-shaped hierarchical structure is found from top to bottom;
step S40, searching the robots which are directly or indirectly managed by the lowest-level managers meeting the task requirements and meet the task constraint conditions, selecting a robot combination which can meet the task requirements as a candidate robot alliance, and further forming a candidate robot alliance set;
step S50, selecting the robot alliance with the best matching degree from the candidate robot alliance set as the robot alliance which is selected to execute the task;
wherein:
the manager is a computer, and in the hierarchical structure, the manager plays a role in storage, calculation and communication. According to the actual deployment requirement, each manager can be integrated on the same computer, different computers can be used respectively, or some managers can be integrated on the same computer.
The pyramid-shaped layered structure is constructed by the following steps:
in step a10, all robots are set as the bottom layer of the pyramid, and the robots are grouped by a preset number of robot grouping members, and one manager is set for each group of robots.
Each group of robots is taken as a subordinate member of a corresponding manager, and each manager stores the serial number of each robot managed by the manager; these managers also act as subordinates to the higher level managers at the same time.
In step a20, grouping managers according to the number of preset manager grouping members, and setting a superior manager for each group of managers; grouping the superior managers according to the preset number of manager grouping members, and respectively setting the superior managers for each group; and so on until there is only one manager at the top level of the pyramid.
The manager at the top layer is the manager at the top layer, so that the construction of the hierarchical structure of the multi-robot system is completed. The hierarchical structure of the multi-robot system constructed here is determined by the multi-robot system itself and does not change from task to task. Therefore, the structure can be constructed in advance before task allocation, and as long as the number of the robots in the multi-robot system is unchanged and the resource vector of each robot is unchanged, the pyramid-shaped hierarchical structure does not need to be reconstructed when the tasks are allocated.
Specifically, step a10 in this embodiment may be:
step A11, taking the robots with the total number of robotNum as the bottom layer of the pyramid, and recording the number of layers as j being 1; grouping the robots, and if (robotNum% P) is 0, then grouping the number of groups1robotNum/P, P robots per group; otherwise, group number groupNum1robotNum/P +1, 1 st to groupNum1-1 groups of P robots, the first group Num1In the group, robotNum- (groupNum)1-1)*P robots;
step A12, setting one robot for each group of bottom layerA manager which is a member of the layer 2 and designates each robot in the group managed by the manager as a subordinate member of the manager; the number of managers is recorded as mangagerNum2＝groupNum1；
Wherein:
p is the preset robot group membership, that is, after the bottom layer robots are grouped, the number of the robots in each robot group is less than P in the last group; groupNum1Grouping the robots in the layer 1 of the pyramid layered structure; mangagerNum2The number of managers in layer 2.
The idea of the above steps a11 to a12 is:
taking each robot in the multi-robot system as a member of the bottommost layer of the layered structure, grouping according to the standard of a group of P robots, and taking the robots as a group independently if the P robots are left; for each group of robots, a manager is set, and the robot corresponding to the manager is called as a subordinate member of the manager.
Specifically, step a20 in this embodiment may be:
step A21, the layer number j is 2;
step A22, grouping the managers at the j layer, if (managerNum)jWhen% Q) is 0, the number of groups is groupNumj＝managerNumjQ, Q managers in each group; otherwise, group number groupNumj＝managerNumj(ii)/Q +1, 1 st to groupNumj-Q managers in each group in group 1, the first group NumjThere is mangagerNum in the groupj-(groupNumj-1)*Q managers;
step A23, setting a superior manager for each group manager of the jth layer as a member of the jth +1 layer, and calling each manager in the group managed by the superior manager as a subordinate member of the superior manager; superior manager number managerNumj+1＝groupNumj；
Step a24, if j is j +1, if the number of j-level managers is greater than 1, go to step a 22; otherwise, go to step A25;
step A25, recording the total layer number of the pyramid as N ═ j;
wherein:
q is the preset administrator group membership, that is, after the administrators in the same layer are grouped, the membership in each administrator group may be less than Q in the last group; groupNumjThe group number of the managers at the j-th layer in the pyramid-shaped hierarchical structure after grouping; mangagerNumjThe number of the managers in the j-th layer.
The idea of the above steps a21 to a25 is:
continuously grouping the managers in the same layer according to the standard of Q groups, and if the number of the remaining managers is less than Q, independently taking the managers as a group; setting a superior manager for each group, wherein each manager is a subordinate member of the corresponding superior manager; the superior managers store the serial numbers of the subordinate members; these managers also act as subordinates to the higher level managers at the same time. And circularly grouping and setting managers until only one manager exists in a certain layer, and the manager is marked as the highest-layer manager.
As shown in fig. 2, a total of 37 robots each have a pyramidal layered structure with N ═ 5, in which P ═ 5 and Q ═ 2 are constructed. Level 5 has 1 top manager M5,1,1(ii) a Layer 4 has 2 managers, M respectively4,1,1、M4,1,2(ii) a Layer 3 has 4 managers, M respectively3,1,1、M3,1,2、M3,2,3、M3,2,4(ii) a Layer 2 has 8 managers, M respectively2,1,1、M2,1,2、M2,2,3、M2,2,4、M2,3,5、M2,3,6、M2,4,7、M2,4,8(ii) a Layer 1 has 37 robots, divided into 8 groups, each of the first 7 groups having 5 robots, and the last group having 2 robots.
As can be seen from the above construction method, the layered structure of the multi-robot system is respectively the Nth layer, the N-1 st layer, … and the 1 st layer from top to bottom. All robots are located at the bottom layer, i.e. layer 1, exceptThe members of other layers outside the layer 1 are managers, and the highest layer manager is located on the Nth layer. The robot in the hierarchical structure can transmit information to and from the administrator and different administrators through a wireless local area network or the like. We use
Figure BDA0001476734590000111
Describe the robot on level 1, where subscripts 1, g1And i1Respectively representing layer number, group number and serial number, and resource vector for resource owned by each robot
Figure BDA0001476734590000112
Is represented by i1robotNum, which is the total number of robots;
using a Boolean-type variable for the state of the robot
Figure BDA0001476734590000113
Indicating when the robot is performing a task or is malfunctioning
Figure BDA0001476734590000114
When the robot can normally run and is idle at present
Figure BDA0001476734590000115
For managers starting from level 2 and going up to the highest level
Figure BDA0001476734590000116
Wherein the subscripts n, gnAnd inRespectively representing layer number, group number and serial number, the values of N are 2,3, …, N, and the manager
Figure BDA0001476734590000117
Resource vector for owned resources
Figure BDA0001476734590000118
And (4) showing.
Specifically, in this embodiment, the robot that satisfies the task constraint condition in step S10 may be a robot that is in an idle state and has a resource vector that is greater than or equal to the task constraint resource vector.
For task T, remember
Figure BDA0001476734590000119
Constraining the resource vectors for the task, as an individual robot
Figure BDA00014767345900001110
In an idle state, i.e.
Figure BDA00014767345900001111
And satisfy
Figure BDA00014767345900001112
Then, the robot is shown to meet the task constraint condition, namely is qualified to participate in the task allocation, and all the robots qualified to participate in the task allocation form a set omegarl。
Specifically, step S20 in this embodiment may be:
based on set omegarlThe robot in (1) integrates resources from bottom to top in the order from the 2 nd layer, the 3 rd layer to the Nth layer, and updates the manager layer by layer
Figure BDA00014767345900001113
Resource vector of
Figure BDA00014767345900001114
The values of N are 2,3, … and N respectively, and the method comprises the following steps:
and (3) integrating the resource vectors of the robots which are directly managed by the managers at the layer 2 and are qualified to participate in the task allocation according to a formula (8), so that the resource vectors of the managers are updated:
Figure BDA0001476734590000121
the managers from the layer 3 to the layer N integrate according to the resource vectors of the subordinates which are directly managed by the managers and the formula (9), thereby realizing the updating of the resource vectors of the managers:
Figure BDA0001476734590000122
specifically, step S30 in this embodiment may be:
comparing the task demand resource vector with the resource vectors owned by the managers in a top-down order from the Nth layer of the pyramid-shaped hierarchical structure; when one manager meets the task requirement, namely the resource vector owned by the manager is greater than or equal to the resource vector needed by the task, the resource vector needed by the task is continuously compared with the resource vectors of all the subordinate members directly managed by the manager, if the subordinate members meeting the task requirement still exist, the comparison process is repeated until the level 2, and the members of the level 2 meeting the task requirement are taken as the manager of the lowest level; and if a manager positioned at the 3 rd layer or higher meets the task requirement but the subordinate members directly managed by the manager do not meet the task requirement and the lowest-layer manager meeting the task requirement does not exist, taking the manager as the lowest-layer manager.
The following illustrates how to find the lowest level manager that meets the task requirements:
assume that a pyramid-shaped hierarchical structure as shown in fig. 2 has been constructed and that the integration of resources by the managers in each layer has been completed.
When a pyramid-shaped hierarchical structure is searched from top to bottom, the pyramid-shaped hierarchical structure is a tree-shaped branch structure, and when a manager meeting the task requirement at the lowest layer is searched, each branch which possibly meets the task requirement needs to be traversed.
Example 1:
(1) if F (M)5,1,1)≥F(T)；
(2) View M4,1,1Discovery of F (M)4,1,1)≥F(T)；
(3) View M3,1,1Discovery of F (M)3,1,1)≥F(T)，And F (M)2,1,1)≥F(T)、F(M2,1,2)<F (T), then M2,1,1Is a lowest-level manager of its corresponding branch that meets the task requirements;
(4) view M3,1,2Discovery of F (M)3,1,2) ≧ F (T), and F (M)2,2,3)≥F(T)、F(M2,2,4) Not less than F (T), then M2,2,3、M2,2,4The manager is a lowest-level manager which meets the task requirement and corresponds to the branch;
(5) view M4,1,2Discovery of F (M)4,1,2)≥F(T)；
(6) View M3,2,3But F (M)3,2,3)<F(T)；
(7) View M3,2,4But F (M)3,2,4)<F(T)；
So far, example 1 has already found out the whole hierarchical structure, and found out 3 lowest-level managers meeting task requirements: m2,1,1、M2,2,3、M2,2,4。
Example 2:
(1) if F (M)5,1,1)≥F(T)；
(2) View M4,1,1Discovery of F (M)4,1,1)≥F(T)；
(3) View M3,1,1But F (M)3,1,1)<F(T)；
(4) View M3,1,2But F (M)3,1,2)<F(T)；
(5) Since so far no one lowest manager has been found in the overall hierarchical structure to meet the task requirements, we will M4,1,1As a lowest manager of its corresponding branch that meets task requirements;
(6) view M4,1,2Discovery of F (M)4,1,2)≥F(T)；
(7) View M3,2,3Discovery of F (M)3,2,3) ≧ F (T), and F (M)2,3,5)≥F(T)、F(M2,3,6) Not less than F (T), then M2,3,5、M2,3,6The managers at the lowest layer which meet the task requirements and correspond to the branches are respectively;
(8) check theSee M3,2,4Discovery of F (M)3,2,4) ≧ F (T), and F (M)2,4,7)≥F(T)、F(M2,4,8)<F, (T), then M2,4,7Is the lowest manager of the corresponding branch which meets the task requirement;
to date, example 2 has examined the entire hierarchical structure, and found 4 lowest level managers meeting task requirements: m4,1,1、M2,3,5、M2,3,6、M2,4,7。
Specifically, step S40 in this embodiment may be:
step S41, judging the number of layers of each manager at the lowest layer meeting the task requirement, and if the manager at the 2 nd layer is in the step S42; otherwise, go to step S43;
step S42, finding out all robots meeting task constraint conditions directly managed by the lowest-level manager; and exhausting all possible combinations for all robots meeting task constraint conditions, wherein when a resource vector of one robot combination is greater than or equal to a task demand resource vector, the robot combination is shown to meet task demands, and the robot combination meeting the task demands is taken as a candidate robot alliance ATAdded to the set of candidate robot unions omegarcPerforming the following steps;
step S43, finding out all robots satisfying task constraint conditions indirectly managed by the lowest-level manager, and combining all robots satisfying task constraint conditions as candidate robot alliance ATAdded to the set of candidate robot unions omegarcIn (1).
The following explains the concepts of "direct management" and "indirect management":
in a hierarchical structure with the total number of N, for the manager at the nth (N is more than or equal to 2 and less than or equal to N), the member at the N-1 th layer which has the membership with the manager is the subordinate member directly managed by the manager; the members of the n-2 th, n-3 rd, … and 1 st levels who have membership to them are all subordinate members indirectly managed by them. For the manager at the layer 2, only the directly managed subordinate members, namely the robots, are available, and the indirectly managed subordinate members are not available.
For example, for manager M at layer 2 in FIG. 22,1,1In other words, the underlying robot R has membership to it1,1,1、R1,1,2、…、R1,1,5It is the subordinate member it directly manages; as another example, for manager M at level 4 in FIG. 24,1,2In particular, the manager M at level 3 with whom it has membership3,2,3And M3,2,4Is a member of the subordinate it directly manages, and has membership in layer 2 to it2,3,5、M2,3,6、M2,4,7And M2,4,8Is indirectly managed by it, a member R in level 1 having membership to it1,5,21、R1,5,22、…、R1,8,37(i.e., the lowest level from R in the figure)1,5,21Starting to the right until R1,8,37All robots) is also indirectly managed by it.
The idea of the above steps S30, S41 to S43 is:
we describe the task demand resource vector by F (T), when a manager satisfies
Figure BDA0001476734590000141
Meaning that the manager has sufficient resources to meet the task requirements; starting from the Nth layer of the layered structure, comparing F (T) with resource vectors owned by managers of each layer in the order from top to bottom, when one manager meets the task requirement, continuing to compare F (T) with the resource vectors of the subordinates of the manager, if the subordinates meeting the task requirement still exist, repeating the comparison process until the layer 2, for each manager of the layer 2 meeting the task requirement, combining the subordinates qualified to participate in the task distribution, exhaustively exhausting all possible robot combinations, and for each robot combination CrCalculating its resource vector F (C) using equation (10)r)：
Figure BDA0001476734590000142
If F (C)r)≥F (T), then combine the corresponding robots CrAs a coalition of candidate robots, use ATTo describe, ATAdd to candidate robot federation set ΩrcIn (1).
If a manager at layer 3 or higher satisfies the task requirement but its subordinate members do not satisfy the task requirement, at ΩrcIn the case of an empty set, searching the members with membership from the manager to the 1 st level layer by layer, and combining all robots meeting task constraint conditions related to the 1 st level as a candidate robot alliance ATAdded to the set of candidate robot unions omegarcIn (1).
Specifically, step S50 in this embodiment may be:
set omega based on candidate robot alliancercThe matching degree constraint condition satisfying the formula (11) is selected:
σ1≤|F(T),F(AT)|≤σ2 (11)
and a best matching degree robot union satisfying the formula (12):
Figure BDA0001476734590000151
wherein, | F (T), F (A)T) I is task demand resource vector F (T) and candidate robot alliance ATResource vector F (A)T) The matching degree between the two is measured by the weighted Euclidean distance between the two; sigma1And σ2Respectively setting a preset minimum matching degree threshold value and a preset maximum matching degree threshold value;
Figure BDA0001476734590000152
is the best league of robots to be assigned to perform task T.
Specifically, in the present embodiment, the matching degree calculation method between resource vectors is as shown in formula (13):
Figure BDA0001476734590000153
wherein:
f (a) and f (b) are two different resource vectors, as shown in equations (14), (15):
F(a)＝[f1(a)f2(a)…fdim(F)(a)] (14)
F(b)＝[f1(b)f2(b)…fdim(F)(b)] (15)
dim (F) is the dimensionality of F (a) and F (b); f. ofkA value representing the kth resource of the resource vector,
Figure BDA0001476734590000154
k＝1,2,...,dim(F)；
the following is a specific example to further illustrate the multi-robot task assignment method:
for example, in a multi-robot system, there are 10 robots in total, where P is 5, the 10 robots are averagely divided into 2 groups, each group is provided with 1 manager, and there are 2 managers in total on the 2 nd layer; for the 2 managers, Q is 2, namely the managers are grouped according to the standard of 2 managers in each group, only 1 group is needed, and 1 superior manager is set; the 1 upper manager is the highest manager.
To this end, a 3-layer pyramid structure is constructed, i.e., N-3. Layer 1 is composed of 10 robots equally divided into 2 groups, wherein 5 robots of group 1 are each represented by R1,1,1、R1,1,2、R1,1,3、R1,1,4、R1,1,5Describing, the 5 robots of the 2 nd group are respectively R1,2,6、R1,2,7、R1,2,8、R1,2,9、R1,2,10The description is carried out; layer 2 is composed of 2 managers, M2,1,1And M2,1,2Describing that 5 robots of the 1 st group and 5 robots of the 2 nd group in the 1 st layer are managed separately; layer 3 is formed by a manager, with M3,1,1Describing, managing M of layer 22,1,1And M2,1,2。
The 3 managers of layer 2 and layer 3 are all integrated on one computer to realize all storage, calculation and communication functions. The wireless local area network is built based on the gorgeous router Pro (WS 851). The dimension of the resource vector of each robot is 4, and the resource vector corresponds to a motion capability resource (unit: m/s), a visual sensor resolution resource (unit: pixel), a laser sensor detection angle resource (unit: degree) and a laser sensor measurement distance resource (unit: meter), and the 4 resources are maximum-value resources. The maximum and minimum values of the values of these 4 resources are given as follows:
Figure BDA0001476734590000161
Figure BDA0001476734590000162
assuming that all of the 10 robots can normally operate and are in an idle state, the specific resource vectors are shown in equations (16) - (25), respectively:
F(R1,1,1)＝[0.5640×48027010] (16)
F(R1,1,2)＝[0.71280×72027010] (17)
F(R1,1,3)＝[0.61280×72027010] (18)
F(R1,1,4)＝[0.61280×7203608] (19)
F(R1,1,5)＝[0.61280×72027020] (20)
F(R1,2,6)＝[0.7640×4803606] (21)
F(R1,2,7)＝[0.51280×7203606] (22)
F(R1,2,8)＝[0.51280×7203608] (23)
F(R1,2,9)＝[0.51920×108027030] (24)
F(R1,2,10)＝[0.61920×108027030] (25)
for task T, the task demand resource vector is shown in equation (26):
F(T)＝[0.6 1280×720 270 10] (26)
the task-constrained resource vector is shown in equation (27):
Figure BDA0001476734590000171
according to the multi-robot task allocation method, individual robots are screened firstly, and it is determined that the robots qualified to participate in the task allocation have R1,1,2、R1,1,3、R1,1,4、R1,1,5And R1,2,10The robots form a robot set omega participating in the task distributionrl(ii) a Performing resource integration from bottom to top in the order from layer 2 to layer 3, and updating manager M layer by layer2,1,1、M2,1,2、M3,1,1The updated resource vector is shown in equations (28) - (30):
F(M2,1,1)＝[0.7 1280×720 360 20] (28)
F(M2,1,2)＝[0.6 1920×1080 270 30] (29)
F(M3,1,1)＝[0.7 1920×1080 360 30] (30)
then comparing the task demand resource vector F (T) with the resource vectors owned by the managers at all layers from top to bottom layer by layer, since F (M)3,1,1) F (T) is continued to follow M3,1,1Member of the genus M2,1,1、M2,1,2Is compared due to F (M)2,1,1)≥F(T)，F(M2,1,2) ≧ F (M), thus respectively from M2,1,1And M2,1,2The managed subordinate robots qualified to participate in the task allocation are exhausted, and for each robot combination CrCalculate its resource vector F (C)r) If F (C)r) If not less than F (T), combining the corresponding robots with the robot CrAs a coalition of candidate robots, use ATTo describe, ATAdd to candidate robot federation set ΩrcPerforming the following steps; on the basis of the above, according to
Figure BDA0001476734590000172
And combined with degree of matchingConstraint sigma1≤|F(T),F(AT)|≤σ2Wherein σ is1＝0.5，σ2Get the best robot federation { R) selected to perform task T ═ 11,1,2,R1,1,4,R1,1,5}。
The invention can quickly determine the optimal robot alliance selected to execute the task by the bottom-up resource integration and the top-down resource comparison in the layered structure, has good real-time performance and provides technical support for the application of the multi-robot system task allocation and the like.
Those of skill in the art will appreciate that the method steps of the examples described in connection with the embodiments disclosed herein may be embodied in electronic hardware, computer software, or combinations of both, and that the components and steps of the examples have been described above generally in terms of their functionality in order to clearly illustrate the interchangeability of electronic hardware and software. Whether such functionality is implemented as electronic hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
So far, the technical solutions of the present invention have been described in connection with the preferred embodiments shown in the drawings, but it is easily understood by those skilled in the art that the scope of the present invention is obviously not limited to these specific embodiments. Equivalent changes or substitutions of related technical features can be made by those skilled in the art without departing from the principle of the invention, and the technical scheme after the changes or substitutions can fall into the protection scope of the invention.

Claims (10)

1. A multi-robot task allocation method based on a layered structure and resource integration is characterized by comprising the following steps:
step S10, based on the pyramid layered structure, according to the task constraint conditions and the resource vectors of each robot, screening out the robots meeting the task constraint conditions from the bottom layer of the pyramid;
step S20, calculating resource vectors owned by each manager in each layer of the pyramid-shaped hierarchical structure from bottom to top according to the robots meeting the task constraint conditions and the resource vectors of each robot;
step S30, according to the task demand resource vector and the resource vector owned by each manager, each lowest-level manager meeting the task demand in the pyramid-shaped hierarchical structure is found from top to bottom;
step S40, searching the robots which are directly or indirectly managed by the lowest-level managers meeting the task requirements and meet the task constraint conditions, selecting a robot combination which can meet the task requirements as a candidate robot alliance, and further forming a candidate robot alliance set;
step S50, selecting the robot alliance with the best matching degree between the task demand resource vector and the resource vector of the candidate robot alliance from the candidate robot alliance set as the selected robot alliance for executing the task;
wherein,
the pyramid-shaped layered structure is constructed by the following steps:
step A10, taking all robots as the bottom layer of the pyramid, grouping the robots according to the preset robot grouping member number, and respectively setting a manager for each group of robots;
step A20, grouping managers according to the number of preset manager grouping members, and setting a superior manager for each group of managers; grouping the superior managers according to the preset number of the manager grouping members, and respectively setting the superior managers for each group; and so on until the top layer of the pyramid has only one manager;
the administrator is a computer;
wherein, step S50 specifically includes:
set omega based on candidate robot alliancercSelecting the constraint condition satisfying the matching degree
σ1≤|F(T),F(AT)|≤σ2
And the robot alliance with the best matching degree:
Figure FDA0002761342270000021
wherein, | F (T), F (A)T) I is task demand resource vector F (T) and candidate robot alliance ATResource vector F (A)T) The matching degree between the two is measured by the weighted Euclidean distance between the two; sigma1And σ2Respectively setting a preset minimum matching degree threshold value and a preset maximum matching degree threshold value;
Figure FDA0002761342270000022
the best robot alliance is selected to execute the task T;
wherein, the resource space is described by F, the dimension is dim (F), and the resource vector in F is represented as F ═ F1 f2...fdim(F)]，fkA value representing the kth resource of the resource vector,
Figure FDA0002761342270000023
k＝1,2,...,dim(F)，
Figure FDA0002761342270000024
and
Figure FDA0002761342270000025
respectively the minimum value and the maximum value of the numerical value of the kth resource of the preset resource vector.
2. The multi-robot task assignment method according to claim 1, wherein step a10 specifically comprises:
step A11, taking the robots with the total number of robotNum as the bottom layer of the pyramid, and recording the number of layers as j being 1; grouping the robots, and if (robotNum% P) is 0, then grouping the number of groups1robotNum/P, P robots per group(ii) a Otherwise, group number groupNum1robotNum/P +1, 1 st to groupNum1-1 groups of P robots, the first group Num1In the group, robotNum- (groupNum)1-1) P robots;
step A12, setting a manager for each group of robots at the bottom layer as a member of the layer 2, and calling each robot in the group managed by the manager as a subordinate member of the manager; the number of managers is recorded as mangagerNum2＝groupNum1；
Wherein,
p is the preset robot group member number; groupNum1The group number of the grouped robots in the 1 st layer of the pyramid-shaped layered structure; mangagerNum2The number of managers in layer 2.
3. The multi-robot task assignment method according to claim 2, wherein step a20 specifically comprises:
step A21, the layer number j is 2;
step A22, grouping the managers at the j layer, if (managerNum)jWhen% Q) is 0, the number of groups is groupNumj＝managerNumjQ, Q managers in each group; otherwise, group number groupNumj＝managerNumj(ii)/Q +1, 1 st to groupNumj-Q managers in each group in group 1, the first group NumjThere is mangagerNum in the groupj-(groupNumj-1) Q managers;
step A23, setting a superior manager for each group manager of the jth layer as a member of the jth +1 layer, and calling each manager in the group managed by the superior manager as a subordinate member of the superior manager; superior manager number managerNumj+1＝groupNumj；
Step a24, if j is j +1, if the number of j-level managers is greater than 1, go to step a 22; otherwise, go to step A25;
step A25, recording the total layer number of the pyramid as N ═ j;
wherein,
q is the managerThe number of group members; groupNumjThe group number of the managers at the j-th layer in the pyramid-shaped hierarchical structure is grouped; mangagerNumjThe number of the managers in the j-th layer.
4. The method for assigning a task to multiple robots according to claim 3, wherein the robot satisfying the task constraint condition is a robot in an idle state and having a resource vector greater than or equal to the task constraint resource vector.
5. The method of claim 4, wherein the comparison between resource vectors is performed by comparing the resources in the resource vectors F (a), and if the values of the resources in the resource vector F (a) are all greater than or equal to the values of the corresponding resources in the resource vector F (b), F (a) is greater than or equal to F (b);
wherein, f (a) and f (b) are two different resource vectors in the resource space, respectively; f, (a) and F (b) have dimensions dim (F), F (a) and F (b) are respectively:
F(a)＝[f1(a) f2(a)...fdim(F)(a)]
F(b)＝[f1(b) f2(b)...fdim(F)(b)]。
6. the multi-robot task assignment method according to claim 5, wherein the step S20 is specifically:
based on the robot meeting the task constraint conditions, performing resource integration from bottom to top in the sequence from the 2 nd layer, the 3 rd layer to the Nth layer, and updating the manager layer by layer
Figure FDA0002761342270000031
Resource vector of
Figure FDA0002761342270000032
Wherein,
for each manager at the layer 2, integrating the resource vectors of the robots meeting the task constraint conditions directly managed by the manager according to the following formula, thereby realizing the update of the resource vectors of the managers:
Figure FDA0002761342270000041
Figure FDA0002761342270000042
as managers in layer 2
Figure FDA0002761342270000043
Resource vector of (2), subscript g2、i2Layer number, group number and serial number of the layer manager;
Figure FDA0002761342270000044
for robots in layer 1
Figure FDA0002761342270000045
Resource vector of (1), subscript g1、i1Respectively is the layer number, group number and serial number of the robot; omegarlThe method comprises the steps of (1) collecting all robots meeting task constraint conditions;
for each manager from the layer 3 to the layer N, integrating the resource vectors of subordinate members directly managed by the manager according to the following formula, and further updating the resource vector of the manager:
Figure FDA0002761342270000046
Figure FDA0002761342270000047
is the nth manager
Figure FDA0002761342270000048
Resource vector of (3), subscripts n, gn、inLayer number, group number and serial number of the layer manager; n is 3, 4.
7. The multi-robot task assignment method of claim 6, wherein the integration of the resource vectors is accomplished by integrating each resource;
for each resource, adopting a corresponding integration rule according to the type of the resource; the resource types include: accumulation type, maximum type, minimum type; the integration rule of the accumulation type resources is to sum corresponding resources; the integration rule of the maximum value type resource is to find the maximum value of the corresponding resource; the integration rule of the minimum value type resource is to find the minimum value of the corresponding resource.
8. The multi-robot task assignment method according to claim 7, wherein the step S30 is specifically:
comparing the task demand resource vector with the resource vectors owned by the managers in a top-down order from the Nth layer of the pyramid-shaped hierarchical structure; when the resource vector owned by a manager is larger than or equal to the task demand resource vector, indicating that the manager meets the task demand, continuously comparing the task demand resource vector with the resource vectors of all the subordinate members directly managed by the manager, if the subordinate members meeting the task demand still exist, repeating the comparison process until the level 2, and taking the level 2 members meeting the task demand as the lowest-level manager; and if a manager positioned at the 3 rd layer or higher meets the task requirement but the subordinate members directly managed by the manager do not meet the task requirement and the lowest-layer manager meeting the task requirement does not exist, taking the manager as the lowest-layer manager.
9. The multi-robot task assignment method according to claim 8, wherein the step S40 is specifically:
step S41, judging the number of layers of each manager at the lowest layer meeting the task requirement, and if the manager at the 2 nd layer is in the step S42; otherwise, go to step S43;
step S42, finding out all robots meeting task constraint conditions directly managed by the lowest-level manager; and exhausting all possible combinations for all robots meeting task constraint conditions, wherein when a resource vector of one robot combination is greater than or equal to a task demand resource vector, the robot combination is shown to meet task demands, and the robot combination meeting the task demands is used as a candidate robot alliance ATAdded to the set of candidate robot unions omegarcPerforming the following steps;
step S43, finding out all robots satisfying task constraint conditions indirectly managed by the lowest-level manager, and combining all robots satisfying task constraint conditions as candidate robot alliance ATAdded to the set of candidate robot unions omegarcIn (1).
10. A multi-robot task assignment method according to any of claims 1-9, characterized in that the degree of matching between resource vectors is calculated by:
Figure FDA0002761342270000051
F(a)＝[f1(a) f2(a)...fdim(F)(a)]
F(b)＝[f1(b) f2(b)...fdim(F)(b)]
wherein, F (a) and F (b) are two different resource vectors in the resource space F, respectively.
CN201711168423.4A 2017-11-21 2017-11-21 Multi-robot task allocation method based on layered structure and resource integration Active CN108171394B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711168423.4A CN108171394B (en) 2017-11-21 2017-11-21 Multi-robot task allocation method based on layered structure and resource integration

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711168423.4A CN108171394B (en) 2017-11-21 2017-11-21 Multi-robot task allocation method based on layered structure and resource integration

Publications (2)

Publication Number Publication Date
CN108171394A CN108171394A (en) 2018-06-15
CN108171394B true CN108171394B (en) 2021-02-02

Family

ID=62527177

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711168423.4A Active CN108171394B (en) 2017-11-21 2017-11-21 Multi-robot task allocation method based on layered structure and resource integration

Country Status (1)

Country Link
CN (1) CN108171394B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110070289B (en) * 2019-04-19 2023-03-24 苏州达家迎信息技术有限公司 Task allocation method, device, equipment and storage medium
CN112859887B (en) * 2019-11-28 2022-06-14 中国科学院沈阳自动化研究所 Multi-underwater robot autonomous task allocation method based on space-based center

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101945492A (en) * 2010-08-09 2011-01-12 哈尔滨工程大学 Clustering-based multi-robot task allocation method
CN102033536A (en) * 2010-12-22 2011-04-27 大连理工大学 Scheduling, organization and cooperation system and method for multi-robot system
CN102831319A (en) * 2012-08-27 2012-12-19 北京科技大学 Task allocation algorithm of fine-grained multi-robot collaboration system
CN103064743A (en) * 2012-12-27 2013-04-24 深圳先进技术研究院 Resource scheduling method for multiple robots and resource scheduling system thereof

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101178847B1 (en) * 2008-12-08 2012-08-31 한국전자통신연구원 Apparatus and method for controlling multi-robot which responding to virtual space

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101945492A (en) * 2010-08-09 2011-01-12 哈尔滨工程大学 Clustering-based multi-robot task allocation method
CN102033536A (en) * 2010-12-22 2011-04-27 大连理工大学 Scheduling, organization and cooperation system and method for multi-robot system
CN102831319A (en) * 2012-08-27 2012-12-19 北京科技大学 Task allocation algorithm of fine-grained multi-robot collaboration system
CN103064743A (en) * 2012-12-27 2013-04-24 深圳先进技术研究院 Resource scheduling method for multiple robots and resource scheduling system thereof

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
Resource Constrained Multirobot Task Allocation with A Leader-Follower Coalition Method;Jian Chen等;《The 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems》;20101022;第5093-5098页 *
多机器人任务分配的研究与进展;张嵛等;《智能系统学报》;20080430;第3卷(第2期);第115-120页 *
多机器人系统任务分配的研究进展;黎萍等;《计算机工程与应用》;20081231;第44卷(第17期);第201-205页 *
多机器人系统任务分配问题的建模与求解;苏丽颖等;《中南大学学报（自然科学版）》;20130731;第44卷;第122-125页 *
自组织多机器人系统任务分配建模与设计;周菁;《中国博士学位论文全文数据库(电子期刊)信息科技辑》;20170115;I140-22 *

Also Published As

Publication number Publication date
CN108171394A (en) 2018-06-15

Similar Documents

Publication Publication Date Title
CN112016812B (en) Multi-unmanned aerial vehicle task scheduling method, system and storage medium
CN104461740B (en) A kind of cross-domain PC cluster resource polymerization and the method for distribution
CN104281496B (en) The distribution method and distributor of virtual resource
CN104050042B (en) The resource allocation methods and device of ETL operations
JP2011118899A (en) Method and system for data distribution in high performance computing cluster
CN108171394B (en) Multi-robot task allocation method based on layered structure and resource integration
Mondal et al. Scheduling of time-varying workloads using reinforcement learning
CN108038576A (en) Based on the logistics distribution routing resource and system for improving dijkstra&#39;s algorithm
CN108647771A (en) The layout method of research-on-research flow data under a kind of mixing cloud environment
CN109409773A (en) A kind of earth observation resource dynamic programming method based on Contract Net Mechanism
CN111313957B (en) Hybrid satellite communication system resource allocation method based on classification multi-objective optimization
CN103200242A (en) Method for constructing cross-layer face data analysis hub based on internet of things
CN114386570A (en) Heterogeneous federated learning training method based on multi-branch neural network model
CN112632615A (en) Scientific workflow data layout method based on mixed cloud environment
Zhang et al. Service composition based on discrete particle swarm optimization in military organization cloud cooperation
CN113570118A (en) Workshop scheduling and analyzing method based on scheduling rule
CN117094435A (en) Novel cloud manufacturing self-adaptive robust service combination and optimization selection method
Shahid et al. A precedence based load balancing strategy for batch of DAGs for computational grid
CN116364256A (en) Emergency medical rescue command dispatching method and dispatching system
CN112148452B (en) Task resource minimum entropy criterion distribution method for single-platform radar comprehensive detection system
CN115689201A (en) Multi-criterion intelligent decision optimization method and system for enterprise resource supply and demand allocation
US20190102406A1 (en) System for multi-criteria decision analysis
Regunathan et al. Neural based QoS aware mobile cloud service and its application to preeminent service selection using back propagation
Raidl et al. Applications of Evolutionary Computing: EvoWorkshop 2003: EvoBIO, EvoCOP, EvoIASP, EvoMUSART, EvoROB, and EvoSTIM, Essex, UK, April 14-16, 2003, Proceedings
TWI812018B (en) Multiple Job Assignment Methods

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