CN111027875A - Intelligent warehousing multi-robot task allocation method based on self-adaptive task pool - Google Patents

Intelligent warehousing multi-robot task allocation method based on self-adaptive task pool Download PDF

Info

Publication number
CN111027875A
CN111027875A CN201911299936.8A CN201911299936A CN111027875A CN 111027875 A CN111027875 A CN 111027875A CN 201911299936 A CN201911299936 A CN 201911299936A CN 111027875 A CN111027875 A CN 111027875A
Authority
CN
China
Prior art keywords
task
robot
threshold
tasks
pool
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
Application number
CN201911299936.8A
Other languages
Chinese (zh)
Other versions
CN111027875B (en
Inventor
杨洪勇
刘飞
唐莉
刘一凡
侯典立
韩辅君
张淑宁
刘慧霞
杨怡泽
于美妍
孙玉娇
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.)
Ludong University
Original Assignee
Ludong University
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 Ludong University filed Critical Ludong University
Priority to CN201911299936.8A priority Critical patent/CN111027875B/en
Publication of CN111027875A publication Critical patent/CN111027875A/en
Application granted granted Critical
Publication of CN111027875B publication Critical patent/CN111027875B/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
    • G06Q10/06311—Scheduling, planning or task assignment for a person or group
    • 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
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • 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
    • G06Q10/06316—Sequencing of tasks or work
    • 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
    • G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G06Q10/083—Shipping

Abstract

The invention discloses an intelligent warehousing multi-robot task allocation method based on a self-adaptive task pool, which comprises the following steps: s1, the initialization task pool P and the task sequence of each robot are empty. S2, setting a threshold value of the task amount stored in the task pool P. And S3, adding a new task into the P, and executing S4 when the number of tasks in the task pool exceeds a threshold value. And S4, performing overall optimization processing on the tasks in the task pool, and distributing the tasks to all robots. S5, each robot inserts the assigned task sequence into the end of the respective current task sequence. And S6, emptying the task pool P, and jumping to S2. Therefore, the invention adopts the self-adaptive task pool to process dynamically added tasks, combines the CMA-ES algorithm, realizes the dynamic overall optimization allocation scheme of the warehousing system in dynamic change, and runs the algorithm and the robot execution task in parallel, thereby saving time and improving the running efficiency of the warehousing system.

Description

Intelligent warehousing multi-robot task allocation method based on self-adaptive task pool
Technical Field
The invention relates to the field of task allocation, in particular to an intelligent warehousing multi-robot task allocation method based on a self-adaptive task pool.
Background
In recent years, electronic commerce is rapidly developed, the order quantity of each e-commerce platform is increased, the warehouse logistics system becomes more and more complex, and the task quantity and scale of a distribution center are more and more huge, so that great challenges are brought to the traditional logistics industry. According to the related data, workers spend 60% to 70% of the time in the traditional warehouse to take goods, so that the traditional mode of manually sorting goods needs to consume a large amount of manpower and material resources, and the efficiency is extremely low.
With the development of robotics, more and more automated machine devices are applied to the field of warehouse logistics. Many enterprises begin to adopt a novel "goods to people" intelligent warehousing system, like amazon's Kiva system, under this system, the workstation is transported from goods shelves storage area to the goods shelves that the robot will need, and the staff only need wait at the platform, and after goods shelves arrived the workstation, take off the goods that will need from the goods shelves or deposit the goods into goods shelves. The fact proves that the intelligent warehousing system greatly saves labor cost and improves the operation efficiency of the warehouse.
In a warehouse, there are often a large number of tasks such as replenishment and picking and a large number of robots to perform these tasks, and meanwhile, the cost for different robots to perform a certain task is different, so that which robots perform which tasks determines the operating efficiency of the warehouse, that is, the cooperative control of multiple mobile robots is the key to achieve intelligent warehousing. This is a typical multi-robot task allocation problem, and as the warehouse operates, the task and the warehouse environment will change continuously, and how to find a better task allocation scheme in such a highly dynamic environment is a problem that needs to be solved urgently by those skilled in the art.
One of the existing processing methods is to consider the multi-robot task allocation problem as a 0-1 integer linear programming problem and solve the problem by adopting a simplex method, a branch-and-bound method, a Hungarian algorithm and the like, but the algorithm can only process the situation that one robot corresponds to one task, has large calculated amount and is only suitable for the situation that the number of the robots and the number of the tasks are small.
The second existing processing method is to solve the problem of task allocation of multiple robots by adopting a genetic algorithm, but the algorithm only considers the problem of task allocation under the condition of a fixed task number in a static environment, and how to process the condition of dynamic change of tasks in actual warehousing application is not considered.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: an intelligent warehousing multi-robot task allocation method based on a self-adaptive task pool is provided.
In order to solve the technical problems, the technical scheme adopted by the invention is as follows:
an intelligent warehousing multi-robot task allocation method based on an adaptive task pool comprises the following steps:
s1, initializing the task pool P and leaving the task sequence of each robot as empty;
s2, setting a threshold value of the task amount stored in the task pool P;
s3, adding the new tasks into the P, and jumping to S4 after the number of the tasks in the task pool exceeds a threshold value;
s4, performing overall optimization processing on the tasks in the task pool, and distributing the tasks to all robots;
s5, each robot inserts the assigned task sequence into the last of the respective current task sequence;
and S6, emptying the task pool P, and jumping to S2.
On the basis of the technical scheme, the invention can be further improved as follows.
Preferably, the step S2 sets a threshold for the task pool P to store the task amount, which includes setting an initial threshold and a dynamically adjusted threshold, where the initial threshold is obtained by calculation according to the formula (5):
initialThreshold＝(γ*stations+robots)/2 (5)
wherein initialThreshold represents an initial threshold value, gamma is the average task number of each workstation, states is the number of workstations, and robots is the number of robots.
Preferably, in the step S2, a threshold of the task amount stored in the task pool P is set, the threshold is automatically adjusted once every time interval I, and the time interval I is an adjustable constant; the threshold is automatically adjusted according to the following steps:
s21, judging whether the current time exceeds the time interval I from the last threshold adjustment time, if so, executing S22, otherwise, executing S21;
s22, judging whether the total task taskfullated completed by the robot during the period from the last threshold adjusting time to the current time is zero, if yes, executing S23, otherwise executing S24;
s23, adjusting the threshold newThreshold to be half of the current threshold, and recording that the adjustment direction of the threshold is reduced; jumping to step S27;
s24, judging whether the total task tasskCompleted completed by the robot during the period from the last threshold adjusting time to the current time is more than or equal to the total task lasttasskCompleted completed by the robot during the period from the last threshold adjusting time to the last threshold adjusting time, if yes, executing S25, otherwise executing S26;
s25, continuously adjusting the threshold newThreshold according to the previous threshold adjusting direction, namely if the previous adjustment is to reduce the threshold, the threshold is continuously reduced in the current adjustment, and if the previous adjustment is to increase the threshold, the threshold is continuously increased in the current adjustment, and the threshold adjusting direction is recorded; jumping to step S27;
s26, adjusting a threshold newThreshold according to a direction opposite to the adjustment direction of the previous threshold, namely if the adjustment of the previous time is to reduce the threshold, increasing the threshold in the adjustment of the current time, and if the adjustment of the previous time is to increase the threshold, reducing the threshold in the adjustment of the current time; recording the threshold adjustment direction;
and S27, finishing the threshold value adjustment.
Preferably, the step S4 performs overall optimization processing on the tasks in the task pool, specifically, regards the tasks in the task pool as an optimization problem in a static environment, and finds an optimal solution by using a CMA-ES algorithm.
Preferably, the optimal solution is found by using the CMA-ES algorithm, and the task allocation scheme is represented by the following codes:
X＝[x1,x2,x3···xm]where m is the number of tasks, xiIs a real number, xiIndicating that the ith task is composed of Int (x)i) Personal robot to execute, Int (x)i) Is expressed for real number xiGetting the whole; if Int (x)i)＝Int(xj) I ≠ j, indicating task xiAnd xjIs assigned to the same robot, x at this timeiAnd xjThe task represented by the smaller number is executed first if xi＝xjThe execution order of these two tasks is random.
Preferably, the optimal solution is found by using a CMA-ES algorithm, and is used as a fitness function according to the following formula:
f＝αC'time+(1-α)C'dist,0≤α≤1 (1)
wherein f is a fitness function, α is a constant and can be adjusted according to actual requirements C'timeRepresents the total time (determined by the maximum time a single robot spends completing the task) from the current time of day that the robot completed the tasks in all task sequences, C'distIndicating that all robots complete tasks in all task sequences from the current timeThe total distance traveled;
preferably, in the step of finding the optimal solution by using the CMA-ES algorithm,
Figure BDA0002321604050000041
wherein C' (r)i) To make the robot r from the current timeiThe cost of executing the tasks in the current task sequence and then executing the tasks according to the allocation scheme of the candidate solution.
Preferably, in the step of finding the optimal solution by using the CMA-ES algorithm,
Figure BDA0002321604050000042
preferably, in the step of finding the optimal solution by using the CMA-ES algorithm,
C'(ri)＝C'1(ri)+C'2(ri) (4)
wherein, C'1(ri) Is a robot riCost, C ', spent executing uncompleted tasks in the current sequence of tasks'2(ri) Is a robot riCost, C ', spent performing tasks according to the allocation scheme of the candidate solution'1(ri) And C'2(ri) The travel distance of the robot is used for representation.
Compared with the prior art, the invention has the following technical effects:
on one hand, the self-adaptive task pool is adopted to process dynamically added tasks, and a CMA-ES algorithm is combined, so that a scheme for dynamically and integrally optimizing and distributing the warehousing system in dynamic change is realized, and the operation of the algorithm and the execution of the tasks by the robot are carried out in parallel, so that the time is saved, and the operation efficiency of the warehousing system is improved.
Drawings
FIG. 1 is a flow chart of an intelligent warehousing multi-robot task allocation method based on an adaptive task pool according to the present invention;
FIG. 2 is a program code segment for dynamically adjusting a threshold according to an embodiment of the present invention.
Detailed Description
The principles and features of this invention are described below in conjunction with the following drawings, which are set forth by way of illustration only and are not intended to limit the scope of the invention.
First, two tasks, task a and task B,
task A Task B
Robot nail
10 minutes 15 minutes
Robot B 20 minutes 30 minutes
If the task A or the task B is independently allocated, the task A or the task B is the best processing efficiency of the robot A;
however, if task a and task B exist simultaneously, there are different schemes:
the first scheme is as follows: the two tasks are both given to the robot A, the robot A always takes 25 minutes, and the robot B is idle;
scheme II: the task A is given to the robot A, the task B is given to the robot B, and the total time of the two tasks is 30 minutes;
the third scheme is as follows: and the task A is given to the robot B, the task B is given to the robot A, and the total time of the two tasks is 20 minutes.
From the above, in the third solution, although the task a is not processed the fastest, the task a consumes the least time for all tasks as a whole, and is the best solution of the three solutions.
In a warehousing system, new orders are added at any time, new tasks are continuously generated along with the addition of new orders, and the tasks are completed as soon as possible, but based on the above example, if a new task is generated, a lowest-cost robot is selected from all robots to operate immediately, the single task is the optimal scheme, but the system continuously generating new orders is not the optimal scheme, in order to obtain dynamic overall optimization processing in a dynamic warehousing system continuously generating new orders, a task pool P is created, when a new task is generated, the new task is added to the P immediately, when the number of tasks in the P meets a threshold value, the CMA-ES algorithm is used for carrying out overall planning on the tasks in the task pool, and the robot inserts a divided new task sequence behind the previous incomplete task sequence, the system then empties the task pool, and the robot executes the tasks according to the assigned task sequence, from which the executed tasks are deleted. The task is added to P again as a new task is generated again.
Fig. 1 is a schematic structural diagram of an intelligent warehousing multi-robot task allocation method based on an adaptive task pool according to the present invention. The intelligent warehousing multi-robot task allocation method based on the self-adaptive task pool comprises the following steps:
s1, the initialization task pool P and the task sequence of each robot are empty.
S2, setting a threshold value of the task amount stored in the task pool P.
And S3, adding the new tasks into the P, and jumping to S4 after the number of the tasks in the task pool exceeds a threshold value.
And S4, performing overall optimization processing on the tasks in the task pool, and distributing the tasks to all robots.
S5, each robot inserts the assigned task sequence into the end of the respective current task sequence.
And S6, emptying the task pool P, and jumping to S2.
Further, step S4 performs an overall optimization process on the tasks in the task pool, which is an optimization process using the CMA-ES algorithm.
For a particular problem, a suitable code must be designed to represent the solution to the problem. For the problem under study in this patent, a set of parameters (a candidate solution) is used to represent a set of task allocation schemes.
We have designed the following coding:
for a task allocation problem with m tasks and n robots, one candidate solution is generated as X ═ X1,x2,x3···xm]. X contains m real numbers, X being for each real numberiSatisfies the following conditions: x is more than or equal to 1i≤n,i＝1,2,3,···,m。
Wherein x isiIndicating that the ith task is composed of Int (x)i) Personal robot to execute, Int (x)i) Is expressed for real number xiAnd (6) taking the whole. If Int (x)i)＝Int(xj) I ≠ j, indicating task xiAnd xjAre all assigned to the same robot, at which time xiAnd xjThe task represented by the smaller number of (a) is executed first. If xi＝xjThe order of execution of the two tasks is randomly decided.
For example, if there are 8 tasks (denoted by numbers 1,2, 3.., 8), and the task assignment problem for 3 robots (denoted by numbers 1,2, 3), an individual is generated [1.7,3.8,2.2,1.3,2.8,1.5,3.3,3.7]. then the task sequence assigned by robot number 1 is: 4 → 6 → 1; the task sequence allocated by the robot No. 2 is as follows: 3 → 5; the task sequence allocated by the robot No. 3 is as follows: 7 → 8 → 2.
The fitness function is used to evaluate the candidate solution, and in the CMA-ES algorithm, individuals with smaller fitness are better. For the whole system, we propose two optimization objectives: one is the time C 'at which the robot completes all tasks'time(ii) a Second, total travel distance C 'of all robots'dist. Each planning can be seen as a sub-problem of the whole, for eachSub-problem, to optimize the whole, we still consider these two goals, so we calculate the fitness function f by:
f＝αC'time+(1-α)C'dist,0≤α≤1 (1)
Figure BDA0002321604050000071
Figure BDA0002321604050000072
wherein α is a constant, which can be adjusted according to actual requirements C' (r)i) To make the robot r from the current timeiThe cost of executing the tasks in the current task sequence and then executing the tasks according to the allocation scheme of the candidate solution. C'timeRepresenting the total time for the robot to complete all tasks from the current time (determined by the maximum time it takes for a single robot to complete a task). C'distIndicating the total distance traveled by all robots to complete all tasks from the current time.
At the current time, there may be tasks planned before incomplete in the task sequence of the robot, and the robot must complete these tasks before executing the tasks planned at the current time. Thus, for C' (r)i) We calculate in two parts:
C'(ri)＝C'1(ri)+C'2(ri) (4)
wherein, C'1(ri) Is a robot riCost, C ', spent executing uncompleted tasks in the current sequence of tasks'2(ri) Is a robot riThe cost of performing the task according to the allocation scheme of the candidate solution. C'1(ri) And C'2(ri) All use the robot riThe travel distance of (c).
Through the fitness function, the optimal solution from the current moment is searched during each planning, and the global optimal solution can be continuously approached through the method.
Further, when the number of tasks in the task pool reaches a threshold value, the tasks in the task pool are allocated to the robot, and the size of the threshold value plays a decisive role in the allocation efficiency. If the planned tasks include all tasks in the system, the optimal solution found by the planning is the optimal solution of the whole system, but the warehousing tasks are dynamic, so that no method is available for considering all the tasks at one time, and the larger the threshold value is, the more schemes are considered and planned at the same time, the larger the threshold value is, the closer the optimal solution of the whole system can be. However, since the orders in the warehouse are dynamically added over time, tasks are also dynamically generated, and as the threshold becomes larger, the time required for the number of tasks in the task pool to reach the threshold also becomes longer, which may occur: the robot has completed all tasks allocated to it, but the number of tasks in the task pool has not reached the threshold yet, the next planning cannot be started, and at this time, the robot can only wait without being left, which causes waste of resources and cannot meet the requirement of the real-time property of the warehousing system. Furthermore, since each station has a limit on the order size, there is an upper limit on the total number of tasks in the system, and if the threshold size of the task pool exceeds this upper limit, the number of tasks in the task pool can never reach the threshold, and the system will stay there. Therefore, it is very important to set a threshold value of a suitable size.
Obviously, different threshold values should be set for different warehouses according to actual conditions, even if the same warehouse is used, the number of robots may be adjusted, the speed of order generation may change at different times, and therefore, it is not appropriate to set the threshold value to a fixed value. Therefore, an adaptive control strategy is needed to dynamically adjust the task pool threshold.
First, an initial threshold is set for the system, which determines the speed at which the algorithm finds the best threshold. The size of the initial threshold value is related to the number of robots and the upper limit of the number of tasks in the system, and the upper limit of the number of tasks in the system is related to the number of workstations and the capacity of each workstation, so the initial threshold value can be calculated according to the following heuristic formula:
initialThreshold＝(γ*stations+robots)/2 (5)
wherein initialThreshold represents an initial threshold value, gamma is the average task number of each workstation, states is the number of workstations, and robots is the number of robots.
After the initial threshold is set, a mechanism is needed to dynamically adjust the threshold. The specific program code is shown in fig. 2, where I is a time interval, I is a constant, and may be adjusted according to actual needs, and the threshold is adjusted once every I seconds (line 1 in the program code).
lastAction is used to record the last adjustment action. tasskcompleted represents the total number of tasks completed by the robot from the time of the last adjustment to the current time, and lasttasskcompleted represents the total number of tasks completed by the robot from the time of the last adjustment to the time of the last adjustment.
If taskfulled is 0, indicating that the threshold is set too large and the number of tasks has not reached the threshold, then the threshold is directly halved and lastAction is set to-1 ( lines 2,3, 4 in the program code).
If tasskcompleted is greater than or equal to lasttasskcompleted, it indicates that the last tuning action has a positive impact on the system and we still perform the last tuning action ( lines 5 and 6 in the program code).
If tasskcomplete is less than lasttasskcomplete, indicating that the last adjustment has a negative impact on the system, we should perform the opposite adjustment from the last and negate lastAction ( lines 7, 8, 9 in the program code).
Thereby, automatic and dynamic adjustment of the task pool threshold is realized.
The intelligent warehousing multi-robot task allocation method based on the self-adaptive task pool adopts the self-adaptive task pool to process dynamically added tasks, combines the CMA-ES algorithm, realizes a dynamic overall optimization allocation scheme for the dynamically changing warehousing system, and runs the algorithm and the robot execution task in parallel, thereby saving time and improving the running efficiency of the warehousing system.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

Claims (9)

1. An intelligent warehousing multi-robot task allocation method based on an adaptive task pool is characterized by comprising the following steps:
s1, initializing the task pool P and leaving the task sequence of each robot as empty;
s2, setting a threshold value of the task amount stored in the task pool P;
s3, adding the new tasks into the P, and executing S4 when the number of the tasks in the task pool exceeds a threshold value;
s4, performing overall optimization processing on the tasks in the task pool, and distributing the tasks to all robots;
s5, each robot inserts the assigned task sequence into the last of the respective current task sequence;
and S6, emptying the task pool P, and jumping to S2.
2. The intelligent warehousing multi-robot task allocation method based on the adaptive task pool as claimed in claim 1, wherein the step S2 sets a threshold value for the task amount stored in the task pool P, including setting an initial threshold value and a dynamically adjusted threshold value, the initial threshold value is obtained by calculation according to formula (5):
initialThreshold＝(γ*stations+robots)/2 (5)
where initialThreshold represents the initial threshold, γ is the average number of tasks per workstation,
statons is the number of workstations and robots is the number of robots.
3. The intelligent-warehousing multi-robot task allocation method based on the adaptive task pool as claimed in claim 1 or 2, wherein in the step S2, the threshold is automatically adjusted once every time interval I, and the time interval I is an adjustable constant; the threshold is automatically adjusted according to the following steps:
s21, judging whether the current time exceeds the time interval I from the last threshold adjustment time, if so, executing S22, otherwise, executing S21;
s22, judging whether the total task taskfullated completed by the robot during the period from the last threshold adjusting time to the current time is zero, if yes, executing S23, otherwise executing S24;
s23, adjusting the threshold newThreshold to be half of the current threshold, and recording that the adjustment direction of the threshold is reduced; jumping to step S27;
s24, judging whether the total task tasskCompleted completed by the robot during the period from the last threshold adjusting time to the current time is more than or equal to the total task lasttasskCompleted completed by the robot during the period from the last threshold adjusting time to the last threshold adjusting time, if yes, executing S25, otherwise executing S26;
s25, continuously adjusting the threshold newThreshold according to the previous threshold adjusting direction, namely if the previous adjustment is to reduce the threshold, the threshold is continuously reduced in the current adjustment, and if the previous adjustment is to increase the threshold, the threshold is continuously increased in the current adjustment, and the threshold adjusting direction is recorded; jumping to step S27;
s26, adjusting a threshold newThreshold according to a direction opposite to the adjustment direction of the previous threshold, namely if the adjustment of the previous time is to reduce the threshold, increasing the threshold in the adjustment of the current time, and if the adjustment of the previous time is to increase the threshold, reducing the threshold in the adjustment of the current time; recording the threshold adjustment direction;
and S27, finishing the threshold value adjustment.
4. The intelligent warehousing multi-robot task allocation method based on the adaptive task pool as claimed in claim 1 or 2, wherein the step S4 performs overall optimization processing on the tasks in the task pool, specifically, the tasks in the task pool are regarded as an optimization problem in a static environment, and a CMA-ES algorithm is used to find an optimal solution.
5. The intelligent-warehousing multi-robot task allocation method based on the adaptive task pool as claimed in claim 4, wherein the optimal solution is found by using CMA-ES algorithm, and the task allocation scheme is represented by the following codes:
X＝[x1,x2,x3···xm]
where m is the number of tasks, xiIs a real number, xiIndicating that the ith task is composed of Int (x)i) Personal robot to execute, Int (x)i) Is expressed for real number xiGetting the whole; if Int (x)i)＝Int(xj) I ≠ j, indicating task xiAnd xjIs assigned to the same robot, x at this timeiAnd xjThe task represented by the smaller number is executed first if xi＝xjThe execution order of these two tasks is random.
6. The intelligent-warehousing multi-robot task allocation method based on the adaptive task pool as claimed in claim 4, wherein the optimal solution is found by using a CMA-ES algorithm, and is used as a fitness function according to the following formula:
f＝αC'time+(1-α)C'dist,0≤α≤1(1)
wherein f is a fitness function, α is a constant, and can be adjusted according to actual requirements, C'timeRepresenting the total time, C ', from the current time until the robot completes all tasks in the task sequence'distIndicating the total distance traveled by the ensemble of robots to complete all of the tasks in the task sequence from the current time.
7. The intelligent-warehousing multi-robot task allocation method based on the adaptive task pool as claimed in claim 6, wherein the optimal solution is found by using CMA-ES algorithm,
Figure FDA0002321604040000031
wherein C' (r)i) To make the robot r from the current timeiThe tasks in the current task sequence are executed firstAnd then the cost of executing the task according to the allocation scheme of the candidate solution.
8. The intelligent-warehousing multi-robot task allocation method based on the adaptive task pool as claimed in claim 7, wherein the optimal solution is found by using CMA-ES algorithm,
Figure FDA0002321604040000032
9. the intelligent-warehousing multi-robot task allocation method based on the adaptive task pool as claimed in claim 8, wherein the optimal solution is found by using CMA-ES algorithm,
C'(ri)＝C'1(ri)+C'2(ri)(4)
wherein, C'1(ri) Is a robot riCost, C ', spent executing uncompleted tasks in the current sequence of tasks'2(ri) Is a robot riCost, C ', spent performing tasks according to the allocation scheme of the candidate solution'1(ri) And C'2(ri) The travel distance of the robot is used for representation.
CN201911299936.8A 2019-12-17 2019-12-17 Intelligent warehouse multi-robot task allocation method based on self-adaptive task pool Active CN111027875B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911299936.8A CN111027875B (en) 2019-12-17 2019-12-17 Intelligent warehouse multi-robot task allocation method based on self-adaptive task pool

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911299936.8A CN111027875B (en) 2019-12-17 2019-12-17 Intelligent warehouse multi-robot task allocation method based on self-adaptive task pool

Publications (2)

Publication Number Publication Date
CN111027875A true CN111027875A (en) 2020-04-17
CN111027875B CN111027875B (en) 2023-09-26

Family

ID=70210129

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911299936.8A Active CN111027875B (en) 2019-12-17 2019-12-17 Intelligent warehouse multi-robot task allocation method based on self-adaptive task pool

Country Status (1)

Country Link
CN (1) CN111027875B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112749927A (en) * 2021-02-03 2021-05-04 香港中文大学（深圳） Dispatching method of storage robot and related equipment
CN112785132A (en) * 2021-01-14 2021-05-11 北京理工大学 Task allocation method for multi-robot mobile shelf for intelligent warehouse
CN112990831A (en) * 2021-04-27 2021-06-18 北京京东振世信息技术有限公司 Vehicle scheduling method and device, electronic equipment and computer readable medium

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2014120031A (en) * 2012-12-18 2014-06-30 International Business Maschines Corporation Task allocation server, task allocation method and task allocation program
CN106228302A (en) * 2016-07-21 2016-12-14 上海仙知机器人科技有限公司 A kind of method and apparatus for carrying out task scheduling in target area
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
CN107784434A (en) * 2017-09-29 2018-03-09 平安科技（深圳）有限公司 A kind of distribution method, storage medium and the server of task of attending a banquet
CN108197778A (en) * 2017-12-15 2018-06-22 远光软件股份有限公司 A kind of method for allocating tasks and device
CN108197865A (en) * 2017-12-29 2018-06-22 珠海国芯云科技有限公司 Logistics distribution task fast dispatch control method and device
CN108416488A (en) * 2017-12-21 2018-08-17 中南大学 A kind of more intelligent robot method for allocating tasks towards dynamic task
CN109614213A (en) * 2018-12-06 2019-04-12 优信拍（北京）信息科技有限公司 A kind of information of vehicles checking method and device
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

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2014120031A (en) * 2012-12-18 2014-06-30 International Business Maschines Corporation Task allocation server, task allocation method and task allocation program
CN106228302A (en) * 2016-07-21 2016-12-14 上海仙知机器人科技有限公司 A kind of method and apparatus for carrying out task scheduling in target area
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
CN107784434A (en) * 2017-09-29 2018-03-09 平安科技（深圳）有限公司 A kind of distribution method, storage medium and the server of task of attending a banquet
CN108197778A (en) * 2017-12-15 2018-06-22 远光软件股份有限公司 A kind of method for allocating tasks and device
CN108416488A (en) * 2017-12-21 2018-08-17 中南大学 A kind of more intelligent robot method for allocating tasks towards dynamic task
CN108197865A (en) * 2017-12-29 2018-06-22 珠海国芯云科技有限公司 Logistics distribution task fast dispatch control method and device
CN109614213A (en) * 2018-12-06 2019-04-12 优信拍（北京）信息科技有限公司 A kind of information of vehicles checking method and device
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

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112785132A (en) * 2021-01-14 2021-05-11 北京理工大学 Task allocation method for multi-robot mobile shelf for intelligent warehouse
CN112749927A (en) * 2021-02-03 2021-05-04 香港中文大学（深圳） Dispatching method of storage robot and related equipment
CN112749927B (en) * 2021-02-03 2023-11-28 香港中文大学（深圳） Scheduling method of storage robot and related equipment
CN112990831A (en) * 2021-04-27 2021-06-18 北京京东振世信息技术有限公司 Vehicle scheduling method and device, electronic equipment and computer readable medium

Also Published As

Publication number Publication date
CN111027875B (en) 2023-09-26

Similar Documents

Publication Publication Date Title
CN111027875A (en) Intelligent warehousing multi-robot task allocation method based on self-adaptive task pool
Asokan et al. Application of adaptive genetic algorithm and particle swarm optimisation in scheduling of jobs and AS/RS in FMS
CN111091238B (en) Automatic container terminal AGV intelligent scheduling method
Wang et al. Application of reinforcement learning for agent-based production scheduling
CN110598941A (en) Bionic strategy-based dual-target scheduling method for particle swarm optimization manufacturing system
CN112561194B (en) Integrated scheduling method and system for production and logistics of mixed flow shop
CN106773686B (en) Path model method for building up is dispatched with piler under the double vehicle operational modes of rail
Petropoulos et al. A particle swarm optimization algorithm for balancing assembly lines
CN113044462B (en) Robot scheduling method, device, system, storage medium and program product
CN114610015A (en) Dynamic scheduling method and device for flexible job shop with carrying robot
CN110717684A (en) Task allocation method based on task allocation coordination strategy and particle swarm optimization
CN112488606A (en) Intelligent optimization and automatic scheduling system for production logistics
Lim et al. A dispatching method for automated guided vehicles by using a bidding concept
Liu et al. Holonic manufacturing system for distributed control of automated guided vehicles
CN113050644B (en) AGV scheduling method based on iterative greedy evolution
CN113326970A (en) Mixed-flow assembly line sequencing optimization method
CN116822884A (en) Multi-crown block cooperative scheduling method in flexible manufacturing system
CN110245809B (en) Intelligent optimization method and system for multi-robot multitask cooperative work
CN116523221A (en) Optimal scheduling method and system for intelligent warehouse picking task
CN116468372A (en) Storage allocation method, system and storage medium
Siswanto et al. Multi-heuristics based genetic algorithm for solving maritime inventory routing problem
CN114707707A (en) Method and system for scheduling AGV task based on improved genetic algorithm
CN115249121A (en) Discrete manufacturing workshop robust scheduling optimization method based on deep reinforcement learning
CN115034143A (en) Multi-resource cooperative intelligent workshop equipment configuration optimization method
CN113821972A (en) Multi-robot collaborative assembly line balancing method considering energy efficiency

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