CN115186878B - Multi-AGV online task allocation method and system - Google Patents

Multi-AGV online task allocation method and system Download PDF

Info

Publication number
CN115186878B
CN115186878B CN202210716146.0A CN202210716146A CN115186878B CN 115186878 B CN115186878 B CN 115186878B CN 202210716146 A CN202210716146 A CN 202210716146A CN 115186878 B CN115186878 B CN 115186878B
Authority
CN
China
Prior art keywords
task
agv
tasks
bidding
new
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
CN202210716146.0A
Other languages
Chinese (zh)
Other versions
CN115186878A (en
Inventor
王晓伟
吴嘉璇
吴松屿
秦兆博
谢国涛
秦洪懋
边有钢
胡满江
秦晓辉
徐彪
丁荣军
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.)
Wuxi Institute Of Intelligent Control Hunan University
Original Assignee
Wuxi Institute Of Intelligent Control Hunan 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 Wuxi Institute Of Intelligent Control Hunan University filed Critical Wuxi Institute Of Intelligent Control Hunan University
Priority to CN202210716146.0A priority Critical patent/CN115186878B/en
Publication of CN115186878A publication Critical patent/CN115186878A/en
Application granted granted Critical
Publication of CN115186878B publication Critical patent/CN115186878B/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
    • G06Q10/04—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
    • G06Q10/06311—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
    • G06Q10/06—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
    • G06Q10/0835—Relationships between shipper or supplier and carriers
    • G06Q10/08355—Routing methods

Landscapes

  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Engineering & Computer Science (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Marketing (AREA)
  • Development Economics (AREA)
  • Tourism & Hospitality (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Game Theory and Decision Science (AREA)
  • Educational Administration (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Multi Processors (AREA)

Abstract

The invention discloses a multi-AGV online task allocation method and a system, wherein the method comprises the following steps: step 1, sequencing the priority of each task according to the arrival time of each task in a task pool; step 2, selecting the tasks with the priority levels positioned in the front N in the task pool as tasks to be distributed according to a preset auction period, performing sequential auctions, and sending relevant information of the tasks to be distributed to each AGV; step 3, AGV optimizes the task execution sequence; the tasks comprise tasks which are not completed currently and new tasks which are received; step 4, the AGV judges whether to participate in bidding according to the self state, if yes, a bidding value is calculated; step 5, distributing the task to the AGV with the maximum bid value, and returning to the step 3; and 6, repeatedly executing the steps 2-5 until the number of tasks in the task pool is 0.

Description

Multi-AGV online task allocation method and system
Technical Field
The invention relates to the technical field of AGV (Automated Guided Vehicle) task scheduling, in particular to a multi-AGV online task allocation method and system.
Background
In warehouse logistics, automatic guided vehicles AGVs are often used for carrying cargoes, the number of carrying tasks is not fixed, the carrying tasks are dynamically generated, the randomness is achieved, the task concentration is high, and the task timeliness requirement is high. The reasonable task allocation strategy is set, so that the time for the AGV to complete the total task can be effectively shortened, the response speed of the AGV to the task is accelerated, and the overall efficiency of the system is improved.
The existing scheduling system generally adopts a centralized task allocation scheme, the number of tasks is generally fixed, the scene that the tasks arrive dynamically cannot be responded quickly, and the higher the calculation complexity is along with the increase of the number of the tasks. Secondly, optimization objective consideration is single, multi-objective optimization is less studied, uneven task load can be caused, excessive or insufficient number of individual AGVs can be caused, and time consumption for all AGVs to complete all tasks is increased. Furthermore, the scheduling system does not consider the problem that the execution sequence of the tasks affects the efficiency of the AGVs in completing the tasks in the task scheduling process.
Disclosure of Invention
It is an object of the present invention to provide a multi-AGV online task allocation method and system that overcomes or at least alleviates at least one of the above-identified deficiencies in the prior art.
In order to achieve the above object, the present invention provides a multi-AGV online task allocation method, which includes:
step 1, sequencing the priority of each task according to the arrival time of each task in a task pool;
step 2, selecting the tasks with the priorities positioned in the front N in the task pool as tasks to be distributed according to a preset auction period, performing sequential auctions, and sending related information of the tasks to be distributed to each AGV;
step 3, the AGV optimizes the task execution sequence; the tasks comprise tasks which are not completed currently and new tasks which are received;
step 4, the AGV judges whether to participate in bidding according to the self state, and if yes, calculates a bidding value;
step 5, distributing tasks to the AGVs corresponding to the bidding values at maximum, and returning to the step 3;
and 6, repeating the executing steps 2-5 until the number of tasks in the task pool is 0.
Further, the method for ordering the unassigned tasks in the task pool in step 1 includes:
calculating the priority of unassigned tasks in the task pool, and then sequencing the unassigned tasks in sequence from small to large according to the priority:
H i ＝(t cur -t i )*P i
wherein H is i Indicating the priority, t, of the task numbered i cur Representing the current time, t i Representing the time of task generation numbered i, P i Indicating the importance of the task numbered i.
Further, the step 2 further includes: and if the number of the unallocated tasks in the task pool is smaller than N, selecting all the unallocated tasks in the task pool for sequential auction.
Further, the method for optimizing the task execution sequence by the AGV in the step 3 specifically includes:
step 31, acquiring position information, current electric quantity information and an unfinished task set of the AGV;
and step 32, optimizing the task execution sequence of each AGV by adopting a particle swarm algorithm, so that the path cost and the energy consumption cost of the AGVs for completing the tasks according to the new task sequence are the lowest.
Further, the particle swarm algorithm specifically includes:
step 321, generating an initial particle swarm by a greedy algorithm;
step 322, generating a new population of particles by cross-variation;
step 323, calculating the fitness of all the individuals in the new particle swarm of step 322 by using the following formula, judging whether the fitness of the individual with the smallest fitness in the particle swarm is smaller than the fitness of the optimal individual in all the particle swarm, if yes, updating the optimal individual in all the particle swarm into the individual, cycling the iteration times T once to be added with 1, and using the optimal individual to cross with other individuals to generate a new individual in the next cycle when the cycle times T are larger than the maximum allowable iteration times T max Ending the iteration, wherein the optimal individuals in all the current particle swarms are the optimal task execution sequence in the current task set, and obtaining the optimal fitness corresponding to the optimal task execution sequence;
Figure BDA0003708840140000021
where fitness represents the fitness of the AGV numbered k to complete all tasks,
Figure BDA0003708840140000022
representing a task set T incomplete at AGVs numbered k k Task k in j Distance to be travelled, R (k) j ，k j+1 ) Representing task k j And task k j+1 And a correlation distance between them.
Further, the method for calculating the bidding value of the AGV participating in bidding in step 4 specifically includes:
step 41, calculating the path cost of the AGV for receiving the new task according to the distance that the AGV completes all the incomplete tasks and the new task needs to travel minus the distance that the AGV completes all the incomplete tasks needs to travel;
step 42, calculating the energy consumption cost of the AGV for receiving the new task according to the residual electric quantity after completing all the incomplete tasks and the new task minus the residual electric quantity after completing all the incomplete tasks by the AGV;
in step 43,calculating the bidding value Q of the AGV numbered k according to the reciprocal of the weighted sum of the path cost and the energy consumption cost by the following formula k Beta is the weighted value of two costs, 0 < beta < 1:
Figure BDA0003708840140000031
wherein L is k ' represents the estimated total path of the AGV with the number of k for completing all tasks according to the optimal task execution sequence after adding new tasks, L k Representing estimated total paths of AGVs with number k for completing all tasks according to optimal task execution sequence and SOC ave Indicating the amount of power consumed by the AGV per unit length of travel.
Further, in the step 4, it is determined whether the AGV participates in bidding according to one of the following two methods:
the method comprises the following steps: judging whether the number of tasks in the self-unfinished task set of the AGV with the number of k is larger than the average unfinished task number multiplied by the proportionality coefficient alpha, if so, the AGV with the number of k does not participate in bidding; otherwise, participating in bidding;
the second method is as follows: estimated remaining capacity of AGV with judgment number k
Figure BDA0003708840140000032
Whether or not to be lower than threshold SOC 0 If yes, the AGV with the number k does not participate in bidding; otherwise participate in bidding.
The invention also provides a multi-AGV online task distribution system, which comprises:
a task pool unit for sorting the priorities of each task according to the arrival time of each task in the task pool, and for marking the assigned task as assigned;
the scheduling unit is used for selecting the tasks with the priorities positioned in the front N in the task pool as tasks to be distributed according to a preset auction period, conducting sequential auctions, and sending relevant information of the tasks to be distributed to each AGV;
the AGV equipment unit is used for optimizing the task execution sequence, judging whether to participate in bidding according to the self state, calculating a bidding value if the judgment is yes, feeding back the bidding value to the scheduling unit, and notifying the AGV equipment unit of the task, wherein the scheduling unit is used for distributing the task to the AGV corresponding to the bidding value at maximum and notifying the AGV equipment unit of the task, and the selected AGV equipment unit updates the task set of the AGV equipment unit; the tasks comprise tasks which are not completed currently and new tasks which are received.
Further, the method for optimizing the task execution sequence by the AGV equipment unit specifically comprises the following steps:
step 31, acquiring position information, current electric quantity information and an unfinished task set of the AGV;
step 32, optimizing the task execution sequence of each AGV by adopting a particle swarm algorithm to minimize the path cost and the energy consumption cost of the AGV completing the task according to the new task sequence, wherein the particle swarm algorithm specifically comprises:
step 321, generating an initial particle swarm by a greedy algorithm;
step 322, generating a new population of particles by cross-variation;
step 323, calculating the fitness of all the individuals in the new particle swarm of step 322 by using the following formula, judging whether the fitness of the individual with the smallest fitness in the particle swarm is smaller than the fitness of the optimal individual in all the particle swarm, if yes, updating the optimal individual in all the particle swarm into the individual, cycling the iteration times T once to be added with 1, and using the optimal individual to cross with other individuals to generate a new individual in the next cycle when the cycle times T are larger than the maximum allowable iteration times T max Ending the iteration, wherein the optimal individuals in all the current particle swarms are the optimal task execution sequence in the current task set, and obtaining the optimal fitness corresponding to the optimal task execution sequence;
Figure BDA0003708840140000041
in the fig, fitness represents the adaptation of the AGV numbered k to complete all tasksThe degree of the heat dissipation,
Figure BDA0003708840140000042
representing a task set T incomplete at AGVs numbered k k Task k in j Distance to be travelled, R (k) j ，k j + 1 ) Representing task k j And task k j+1 And a correlation distance between them.
Further, the method for calculating the bidding value of the AGV participating in bidding in step 4 specifically includes:
step 41, calculating the path cost of the AGV for receiving the new task according to the distance that the AGV completes all the incomplete tasks and the new task needs to travel minus the distance that the AGV completes all the incomplete tasks needs to travel;
step 42, calculating the energy consumption cost of the AGV for receiving the new task according to the residual electric quantity after completing all the incomplete tasks and the new task minus the residual electric quantity after completing all the incomplete tasks by the AGV;
step 43, calculating the bidding value Q of the AGV with the number k according to the inverse of the weighted sum of the path cost and the energy consumption cost k Beta is the weighted value of two costs, 0 < beta < 1:
Figure BDA0003708840140000043
wherein L is k ' represents the estimated total path of the AGV with the number of k for completing all tasks according to the optimal task execution sequence after adding new tasks, L k Representing estimated total paths of AGVs with number k for completing all tasks according to optimal task execution sequence and SOC ave Indicating the amount of power consumed by the AGV per unit length of travel.
Due to the adoption of the technical scheme, the invention has the following advantages:
1. the method is suitable for various scenes such as warehousing and logistics, the scheduling system can quickly respond to the tasks, the tasks are distributed to the proper AGVs, the total path cost of the AGVs for executing the tasks is reduced, and the overall benefit of the system is improved.
2. Aiming at the scene that the tasks are dynamically generated and the number of the tasks is unpredictable, the task pool is constructed to buffer the tasks, and the tasks are ordered according to the time and importance of the task generation, so that the tasks with high comprehensive priority are distributed and executed first.
3. The multi-AGV task allocation strategy provided by the invention considers the task load balance of each AGV during task allocation, avoids the situations of excessive number of individual AGV tasks and excessive idle time of individual AGVs, fully utilizes all AGV equipment resources and averages the task number of each AGV.
4. The distributed task allocation algorithm is constructed, and is different from the centralized task allocation algorithm, the calculation complexity is low, each AGV only needs to calculate the bidding value of the AGV during task allocation and send the bidding value to the scheduling system, the scheduling system only needs to select the AGV with the optimal bidding value, and the task is allocated to the optimal AGV for processing, so that the efficiency of completing the task is improved, and the overall efficiency of the system is improved.
5. The AGV task execution sequence optimization algorithm provided by the invention can dynamically adjust the AGV task execution sequence, and can reduce the total path cost of the AGV for completing all tasks.
Drawings
Fig. 1 is a schematic flow chart of task allocation according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of task ordering optimization flow provided by an embodiment of the present invention.
FIG. 3 is a schematic diagram of a multi-AGV scheduling system according to an embodiment of the present invention.
Detailed Description
In the drawings, the same or similar reference numerals are used to denote the same or similar elements or elements having the same or similar functions. Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
In the description of the present invention, the terms "center", "longitudinal", "lateral", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", and the like indicate an orientation or a positional relationship based on that shown in the drawings, only for convenience of description and simplification of the description, and do not indicate or imply that the apparatus or element to be referred to must have a specific orientation, be configured and operated in a specific orientation, and therefore should not be construed as limiting the scope of protection of the present invention.
In the case of no conflict, the technical features in the embodiments and the implementation modes of the present invention may be combined with each other, and are not limited to the embodiments or implementation modes where the technical features are located.
The invention will be further described with reference to the drawings and the specific embodiments, it being noted that the technical solution and the design principle of the invention will be described in detail with only one optimized technical solution, but the scope of the invention is not limited thereto.
The following terms are referred to herein, and for ease of understanding, the meaning thereof is described below. It will be understood by those skilled in the art that other names are possible for the following terms, but any other name should be construed to be consistent with the terms set forth herein without departing from their meaning.
As shown in fig. 1, the method for distributing the online tasks of the multiple AGVs provided by the embodiment of the invention includes:
and step 1, sequencing the priority of each task according to the arrival time of each task in the task pool. The task pool is constructed in advance and used for dynamically storing the task information which arrives in real time, recording the starting point, the end point, the generation time and the importance degree of the task and sequencing the priority of the task.
In one embodiment, the prioritization of tasks is generally based on task arrival time and priority, which includes:
and step 1.1, recording dynamically generated task information by a task pool. The related information of the task mainly comprises a task starting point, a task ending point, a task generating time and a task importance degree, S i Indicating task related information numbered i.
S i ＝(x i ，y i ，x i ′，y i ′，t i ，P i ) (1)
In (x) i 、y i ) Representing the two-dimensional coordinates of the starting point of the task numbered i, (x) i ′、y i ') represents the endpoint two-dimensional coordinates, t, of the task numbered i i Representing the time of task generation numbered i, P i Represents the importance level of the task with the number i which is preset, P i The value range of (2) is 1-2, the higher the importance degree is, P i The larger the value is.
And 1.2, sequencing unassigned tasks in the task pool. Acquiring current time t cur The priorities of all tasks are calculated.
H i ＝(t cur -t i )*P i (2)
Wherein H is i Indicating the priority of the task numbered i. And then sorting according to the priority of the tasks, and arranging the tasks with large priority values in front.
The method and the system are suitable for various scenes such as logistics and warehousing, the scheduling system can rapidly respond to tasks, the tasks are distributed to proper AGVs, the total path cost and the total energy consumption cost of the AGVs for executing the tasks are reduced, and the overall benefit of the system is improved. In the logistics and storage scenes, the tasks to be distributed are dynamically generated and have unpredictability, and the number of the tasks is not fixed, so that the system constructs a task pool for dynamically storing the tasks, the tasks are prevented from being excessively large or excessively small, and the tasks with high priority are distributed and executed first according to the task generation time and importance degree. Of course, prioritization may be performed in other ways as desired.
And 2, selecting the tasks with the priorities positioned in the front N in the task pool as tasks to be distributed according to a preset auction period, performing sequential auctions, and sending related information of the tasks to be distributed to each AGV.
Wherein, the auction period T is preset fix May be derived from, but not limited to, historical data settings, such as: in the historical data, M carrying tasks are generated in average every hour, then the auction period can be set to be less than N/M hours, then T fix Can be set as N2M hours. N is a positive integer. The task related information includes a task start point, a task end point, a task generation time, and a task importance degree.
Step 2 also includes a special case: and if the number of the unallocated tasks in the task pool is smaller than N, selecting all the unallocated tasks in the task pool for sequential auction.
And 3, the AGV optimizes the task execution sequence. The tasks comprise tasks which are not completed currently and new tasks which are received.
In one embodiment, the method for optimizing the task execution sequence by the AGV in step 3 specifically includes:
and step 31, acquiring position information, current electric quantity information and an unfinished task set of the AGV.
Figure BDA0003708840140000071
The two-dimensional coordinates of the position of the AGVs with the number k are represented, the value range of k is 1-M, and M represents the total number of the AGVs.
Figure BDA0003708840140000072
Representing the current power of the AGV numbered k. T (T) k Indicating the set of non-completed tasks of AGVs themselves numbered k +.>
Figure BDA0003708840140000073
For task set T k Tasks, k j Numbering the task in the task pool. The order of the tasks in the task set is the execution order of the tasks. J (J) k Representing the number of tasks in the task set that the AGV numbered k does not complete itself.
Figure BDA0003708840140000074
And step 32, optimizing the task execution sequence of each AGV by adopting a particle swarm algorithm, so that the path cost and the energy consumption cost of the AGVs for completing the tasks according to the new task sequence are the lowest. AGV completes tasks according to the sequence of tasks in the task set, assuming any task in the task setThe order of the transactions is as shown in formula (3). Then the AGV completes the estimated total path L of all tasks k Equal to the distance travelled by a single mission
Figure BDA0003708840140000075
Adding the associated distance R (k) j ，k j+1 ). And estimating the total distance, namely subsequently estimating the individual fitness of the individual quality of the particle swarm. Distance that AGV needs to travel to complete a single task +.>
Figure BDA0003708840140000076
And the associated distance R (k) for the AGV to transfer from the last task end to the next task start j ，k j+1 ). Manhattan distance estimation is used. The starting point of the task is (x j ，y j ) The end point is (x j ′，y j ′)。(x j+1 ，y j+1 ) Is the starting point for the next task.
In another embodiment, the Euclidean distance may also be used or a path planning algorithm may be used to find the distance instead of the Manhattan distance in the above embodiments.
In one embodiment, the particle swarm algorithm specifically includes:
step 321, generating an initial particle swarm by a greedy algorithm: assuming that the current AGV task set has N tasks in total, firstly selecting one task i as a starting point, then selecting the task with the shortest Manhattan distance from the end point of the task i in the rest tasks, and then, analogizing the task with the shortest end point from the last task each time. Traversing the above process for N times, selecting different tasks as starting points each time, and recording the path cost and task execution sequence calculated by taking the different tasks as starting points. The number of cycles T is set to 1. The individual fitness function fitnes of the particle swarm is the path cost for completing all tasks. And comparing N task execution sequence schemes, and selecting an individual with the minimum adaptability as an optimal individual of the particle swarm and an optimal individual in all the particle swarms.
Step 322, generating a new population of particles by cross-variation: crossover refers to the selection of different individuals in a population of particles to exchange for different sequences, and mutation refers to the random adjustment of the sequences of the individuals in the population of particles. Firstly, selecting the optimal individual in all particle swarms, and randomly exchanging different sequences between other individuals and the individual to generate a new individual. And part of individuals carry out mutation operation and exchange sequence, wherein the individual exchange sequence means the execution sequence of exchange tasks.
Step 323, calculating the fitness of all the individuals in the new particle swarm of step 322 by using the following formula, judging whether the fitness of the individual with the smallest fitness in the particle swarm is smaller than the fitness of the optimal individual in all the particle swarm, if yes, updating the optimal individual in all the particle swarm into the individual, cycling the iteration times T once to be added with 1, and using the optimal individual to cross with other individuals to generate a new individual in the next cycle when the cycle times T are larger than the maximum allowable iteration times T max Ending the iteration, wherein the optimal individuals in all the current particle swarms are the optimal task execution sequence in the current task set, and obtaining the optimal fitness corresponding to the optimal task execution sequence;
Figure BDA0003708840140000081
Figure BDA0003708840140000082
R(k j ，k j+1 )＝|x j ′-x j+1 |+|y j ′-y j+1 | (6)
where fitness represents the fitness of the AGV numbered k to complete all tasks,
Figure BDA0003708840140000083
representing a task set T incomplete at AGVs numbered k k Task k in j Distance to be travelled, R (k) j ，k j+1 ) Representing task k j And task k j+1 And a correlation distance between them. Maximum allowable iteration number T max Typically greater than 500.
Assume that the original set of incomplete tasks for AGVs numbered k is T k The task set after adding new tasks in the original task set is T k ′。
All AGV original task sets T are calculated by using the particle swarm task optimization ordering algorithm k The optimal task execution sequence and the optimal fitness of the system, namely the estimated total path L for completing all tasks according to the optimal task execution sequence k The task set T after adding the new task is calculated by using the method k ' optimal task execution order and estimated total path L for completing all tasks according to the order k ′。
According to the method for optimizing the task execution sequence of the AGV task set in the embodiment, the execution sequence of the tasks distributed by the AGV is optimized by adopting the improved particle swarm algorithm, and the total path cost and the total energy consumption cost of the AGV for completing all the tasks can be effectively reduced.
And step 4, the AGV judges whether to participate in bidding according to the self state, and if yes, calculates a bidding value.
In one embodiment, in step 4, it may be determined whether the AGV is involved in bidding as follows:
the method comprises the following steps: judging whether the number of tasks in the self-unfinished task set of the AGV with the number of k is larger than the average unfinished task number multiplied by the proportionality coefficient alpha, if so, the AGV with the number of k does not participate in bidding; otherwise participate in bidding.
In this embodiment, the AGV may understand, according to its own status, that the AGV predicts the remaining power after completing the existing incomplete tasks and the new tasks, and further evaluates whether to participate in bidding, if the number of the incomplete tasks of the current AGV is greater than the average number of the incomplete tasks multiplied by the scaling factor α, the value range of α is set between 1 and 1.5, and the smaller the value of α, the smaller the value of α is, which indicates that the number of tasks in the task set that allows a single AGV to complete itself exceeds the average number of the incomplete tasks, and if equation (8) is satisfied, the AGV does not participate in bidding, and the bidding value is set to 0. That is, determining whether the AGVs are involved in bidding requires calculating an average number J of task sets for all AGVs that are not completed ave As a basis:
Figure BDA0003708840140000091
J k ＞J ave *α (8)
in another embodiment, in step 4, it may be determined whether the AGV is involved in bidding as follows:
the second method is as follows: estimated remaining capacity of AGV with judgment number k
Figure BDA0003708840140000092
Whether or not to be lower than threshold SOC 0 If so, the AGV with the number k does not participate in bidding as shown in the following formula (9); otherwise participate in bidding.
Figure BDA0003708840140000093
In the method, in the process of the invention,
Figure BDA0003708840140000094
task set T representing completion of self-incomplete task and new task by AGV numbered k k ' post estimated remaining charge. Estimated remaining capacity +.>
Figure BDA0003708840140000095
Equal to the current electric quantity->
Figure BDA0003708840140000096
Subtracting the completion task set T k Estimated total path L of all tasks in k ' multiplying the power consumed per unit length of AGV travel SOC ave . That is, a threshold value SOC is set 0 If estimated remaining capacity of AGV numbered k +.>
Figure BDA0003708840140000097
Below threshold SOC 0 The AGV does not participate in the bid, and the bid value is set to 0.
And each AGV calculates a bid value according to the current incomplete task information, the new task information and the current electricity quantity, and sends the bid value to the scheduling system. Where the new task refers to the task being auctioned. The electric quantity threshold value is generally set to be 20% -40% of the total electric quantity.
The method for calculating the bidding value of the AGV participating in bidding in step 4 specifically comprises the following steps:
step 41, calculating the path cost of the AGV for receiving the new task according to the distance that the AGV completes all the incomplete tasks and the new task needs to travel minus the distance that the AGV completes all the incomplete tasks needs to travel;
step 42, calculating the energy consumption cost of the AGV for receiving the new task according to the residual electric quantity after completing all the incomplete tasks and the new task minus the residual electric quantity after completing all the incomplete tasks by the AGV;
step 43, calculating the bidding value Q of the AGV with the number k according to the inverse of the weighted sum of the path cost and the energy consumption cost k Beta is a weighted value of two costs, beta is more than 0 and less than 1, and after each AGV calculates the bid value, the bid information is sent to a dispatching center:
Figure BDA0003708840140000101
wherein L is k ' represents the estimated total path of the AGV with the number of k for completing all tasks according to the optimal task execution sequence after adding new tasks, L k Representing estimated total paths of AGVs with number k for completing all tasks according to optimal task execution sequence and SOC ave Indicating the amount of power consumed by the AGV per unit length of travel.
And 5, distributing the task to the AGV corresponding to the maximum bidding value, returning to the step 3, and marking the task as distributed in a task pool. The multi-AGV task allocation strategy provided by the invention considers the task load balance of each AGV during task allocation, avoids the situations of excessive number of individual AGV tasks and excessive idle time of individual AGVs, fully utilizes all AGV equipment resources and averages the task number of each AGV. And in addition, bidding values of all AGVs are comprehensively considered during task allocation, and the tasks can be allocated to the optimal AGVs for processing, so that the execution efficiency of the completed tasks is improved, and the overall efficiency of the system is improved.
And 6, repeating the executing steps 2-5 until the number of tasks in the task pool is 0.
As shown in fig. 3, the embodiment of the invention further provides a multi-AGV online task allocation system, which includes a task pool unit, a scheduling unit, and an AGV device unit, wherein:
on the other hand, the invention also provides a scheduling system for multi-AGV task allocation. The system comprises a task pool unit, a scheduling unit and an AGV equipment unit.
The task pool unit is used for sequencing the priority of each task according to the arrival time of each task in the task pool. Further, the task pool unit is responsible for dynamically storing the task information arriving in real time, recording the start point, the end point, the generation time and the importance degree of the task, and prioritizing the task.
The scheduling unit is used for selecting the tasks with the priority levels positioned in the front N in the task pool as tasks to be distributed according to a preset auction period, and performing sequential auctions. The scheduling unit is also used for issuing task information and sending the related information of the task to be distributed to each AGV equipment unit through the communication unit. The communication unit may be wifi, 4G, 5G or other wireless data transmission communication modes. Of course, the communication unit may be built in the scheduling unit, and each AGV device unit in the communication unit in the scheduling unit may send the task to be allocated.
The AGV equipment unit is used for receiving the task information transmitted by the scheduling unit, judging whether to participate in bidding according to the self state, if yes, calculating a bidding value, and feeding back the bidding value to the scheduling unit through the communication unit.
The scheduling unit distributes tasks to the AGVs with the maximum bid values, and the selected AGVs are informed of the selected AGV devices through the communication unit, so that the selected AGVs update task sets of the AGVs. The task pool unit marks the task as allocated.
In one embodiment, the method for optimizing the task execution sequence of the AGV equipment unit specifically includes:
step 31, acquiring position information, current electric quantity information and an unfinished task set of the AGV;
step 32, optimizing the task execution sequence of each AGV by using a particle swarm algorithm, so that the path cost and the energy consumption cost of the AGV for completing the task according to the new task sequence are the lowest, as shown in FIG. 2, wherein the particle swarm algorithm specifically comprises:
in step 321, an initial population of particles is generated by a greedy algorithm.
At step 322, a new population of particles is generated by cross-variation.
Step 323, calculating the fitness of all the individuals in the new particle swarm of step 322 by using the following formula, judging whether the fitness of the individual with the smallest fitness in the particle swarm is smaller than the fitness of the optimal individual in all the particle swarm, if yes, updating the optimal individual in all the particle swarm into the individual, cycling the iteration times T once to be added with 1, and using the optimal individual to cross with other individuals to generate a new individual in the next cycle when the cycle times T are larger than the maximum allowable iteration times T max Ending the iteration, wherein the optimal individuals in all the current particle swarms are the optimal task execution sequence in the current task set, and obtaining the optimal fitness corresponding to the optimal task execution sequence;
Figure BDA0003708840140000111
where fitness represents the fitness of the AGV numbered k to complete all tasks,
Figure BDA0003708840140000112
representing a task set T incomplete at AGVs numbered k k Task k in j Distance to be travelled, R (k) j ，k j+1 ) Representing task k j And task k j+1 And a correlation distance between them.
In one embodiment, the method for calculating the bidding value of the AGV participating in the bidding by the AGV equipment unit specifically comprises the following steps:
step 41, calculating the path cost of the AGV for receiving the new task according to the distance that the AGV completes all the incomplete tasks and the new task needs to travel minus the distance that the AGV completes all the incomplete tasks needs to travel;
step 42, calculating the energy consumption cost of the AGV for receiving the new task according to the residual electric quantity after completing all the incomplete tasks and the new task minus the residual electric quantity after completing all the incomplete tasks by the AGV;
step 43, calculating the bidding value Q of the AGV with the number k according to the inverse of the weighted sum of the path cost and the energy consumption cost k Beta is the weighted value of two costs, 0 < beta < 1:
Figure BDA0003708840140000121
wherein L is k ' represents the estimated total path of the AGV with the number of k for completing all tasks according to the optimal task execution sequence after adding new tasks, L k Representing estimated total paths of AGVs with number k for completing all tasks according to optimal task execution sequence and SOC ave Indicating the amount of power consumed by the AGV per unit length of travel.
Finally, it should be pointed out that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting. Those of ordinary skill in the art will appreciate that: the technical schemes described in the foregoing embodiments may be modified or some of the technical features may be replaced equivalently; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims (8)

1. The multi-AGV online task allocation method is characterized by comprising the following steps of:
step 1, sequencing the priority of each task according to the arrival time of each task in a task pool;
step 2, selecting the tasks with the priorities positioned in the front N in the task pool as tasks to be distributed according to a preset auction period, performing sequential auctions, and sending related information of the tasks to be distributed to each AGV;
step 3, the AGV optimizes the task execution sequence; the tasks comprise tasks which are not completed currently and new tasks which are received;
step 4, the AGV judges whether to participate in bidding according to the self state, and if yes, calculates a bidding value;
step 5, distributing tasks to the AGVs corresponding to the bidding values at maximum, and returning to the step 3;
step 6, repeating the executing steps 2-5 until the number of tasks in the task pool is 0;
the method for calculating the bidding value of the AGV participating in bidding in the step 4 specifically comprises the following steps:
step 41, calculating the path cost of the AGV for receiving the new task according to the distance that the AGV completes all the incomplete tasks and the new task needs to travel minus the distance that the AGV completes all the incomplete tasks needs to travel;
step 42, calculating the energy consumption cost of the AGV for receiving the new task according to the residual electric quantity after completing all the incomplete tasks and the new task minus the residual electric quantity after completing all the incomplete tasks by the AGV;
step 43, calculating the bidding value Q of the AGV with the number k according to the inverse of the weighted sum of the path cost and the energy consumption cost k Beta is the weighted value of two costs, 0<β<1：
Figure QLYQS_1
Wherein L is k ' represents the estimated total path of the AGV with the number of k for completing all tasks according to the optimal task execution sequence after adding new tasks, L k Representing estimated total paths of AGVs with number k for completing all tasks according to optimal task execution sequence and SOC ave Indicating the amount of power consumed by the AGV per unit length of travel.
2. The multi-AGV online task allocation method according to claim 1, wherein the method for ordering the unallocated tasks in the task pool in step 1 comprises:
calculating the priority of unassigned tasks in the task pool, and then sequencing the unassigned tasks in sequence from small to large according to the priority:
H i ＝(t cur -t i )*P i
wherein H is i Indicating the priority, t, of the task numbered i cur Representing the current time, t i Representing the time of task generation numbered i, P i Indicating the importance of the task numbered i.
3. The multi-AGV online task allocation method according to claim 1, wherein said step 2 further comprises: and if the number of the unallocated tasks in the task pool is smaller than N, selecting all the unallocated tasks in the task pool for sequential auction.
4. The multi-AGV online task allocation method according to claim 1, wherein the method for optimizing the task execution sequence by the AGV in step 3 specifically comprises:
step 31, acquiring position information, current electric quantity information and an unfinished task set of the AGV;
and step 32, optimizing the task execution sequence of each AGV by adopting a particle swarm algorithm, so that the path cost and the energy consumption cost of the AGVs for completing the tasks according to the new task sequence are the lowest.
5. The multi-AGV online task allocation method according to claim 4, wherein the particle swarm algorithm specifically comprises:
step 321, generating an initial particle swarm by a greedy algorithm;
step 322, generating a new population of particles by cross-variation;
step 323, calculate the new step 322 using the following formulaThe fitness of all individuals in the particle swarm is judged, whether the fitness of the individual with the smallest fitness in the particle swarm is smaller than the fitness of the optimal individual in all the particle swarm is judged, if yes, the optimal individual in all the particle swarm is updated to be the individual, the iteration number T is increased by 1 in one cycle, and in the next cycle, the optimal individual is used for intersecting with other individuals to generate a new individual, and when the cycle number T is greater than the maximum allowable iteration number T max Ending the iteration, wherein the optimal individuals in all the current particle swarms are the optimal task execution sequence in the current task set, and obtaining the optimal fitness corresponding to the optimal task execution sequence;
Figure QLYQS_2
where fitness represents the fitness of the AGV numbered k to complete all tasks,
Figure QLYQS_3
representing a task set T incomplete at AGVs numbered k k Task k in j Distance to be travelled, R (k) j ,k j+1 ) Representing task k j And task k j+1 And a correlation distance between them.
6. The multi-AGV online task allocation method according to claim 1, wherein in step 4, it is determined whether the AGV participates in bidding according to one of two methods:
the method comprises the following steps: judging whether the number of tasks in the self-unfinished task set of the AGV with the number of k is larger than the average unfinished task number multiplied by the proportionality coefficient alpha, if so, the AGV with the number of k does not participate in bidding; otherwise, participating in bidding;
the second method is as follows: estimated remaining capacity of AGV with judgment number k
Figure QLYQS_4
Whether or not to be lower than threshold SOC 0 If yes, the AGV with the number k does not participate in bidding; otherwise participate in bidding.
7. A multi-AGV online task distribution system comprising:
a task pool unit for sorting the priorities of each task according to the arrival time of each task in the task pool, and for marking the assigned task as assigned;
the scheduling unit is used for selecting the tasks with the priorities positioned in the front N in the task pool as tasks to be distributed according to a preset auction period, conducting sequential auctions, and sending relevant information of the tasks to be distributed to each AGV;
the AGV equipment unit is used for optimizing the task execution sequence, judging whether to participate in bidding according to the self state, calculating a bidding value if the judgment is yes, feeding back the bidding value to the scheduling unit, and notifying the AGV equipment unit of the task, wherein the scheduling unit is used for distributing the task to the AGV corresponding to the bidding value at maximum and notifying the AGV equipment unit of the task, and the selected AGV equipment unit updates the task set of the AGV equipment unit; the tasks comprise tasks which are not completed currently and new tasks which are received;
the method for calculating the bidding value of the AGV participating in bidding specifically comprises the following steps:
step 41, calculating the path cost of the AGV for receiving the new task according to the distance that the AGV completes all the incomplete tasks and the new task needs to travel minus the distance that the AGV completes all the incomplete tasks needs to travel;
step 42, calculating the energy consumption cost of the AGV for receiving the new task according to the residual electric quantity after completing all the incomplete tasks and the new task minus the residual electric quantity after completing all the incomplete tasks by the AGV;
step 43, calculating the bidding value Q of the AGV with the number k according to the inverse of the weighted sum of the path cost and the energy consumption cost k Beta is the weighted value of two costs, 0<β<1：
Figure QLYQS_5
Wherein L is k ' represents the estimated total path of the AGV with the number of k for completing all tasks according to the optimal task execution sequence after adding new tasks, L k Representing estimated total paths of AGVs with number k for completing all tasks according to optimal task execution sequence and SOC ave Indicating the amount of power consumed by the AGV per unit length of travel.
8. The multi-AGV online task allocation system of claim 7 wherein the method of the AGV equipment unit optimizing the order of task execution specifically comprises:
step 31, acquiring position information, current electric quantity information and an unfinished task set of the AGV;
step 32, optimizing the task execution sequence of each AGV by adopting a particle swarm algorithm to minimize the path cost and the energy consumption cost of the AGV completing the task according to the new task sequence, wherein the particle swarm algorithm specifically comprises:
step 321, generating an initial particle swarm by a greedy algorithm;
step 322, generating a new population of particles by cross-variation;
step 323, calculating the fitness of all the individuals in the new particle swarm of step 322 by using the following formula, judging whether the fitness of the individual with the smallest fitness in the particle swarm is smaller than the fitness of the optimal individual in all the particle swarm, if yes, updating the optimal individual in all the particle swarm into the individual, cycling the iteration times T once to be added with 1, and using the optimal individual to cross with other individuals to generate a new individual in the next cycle when the cycle times T are larger than the maximum allowable iteration times T max Ending the iteration, wherein the optimal individuals in all the current particle swarms are the optimal task execution sequence in the current task set, and obtaining the optimal fitness corresponding to the optimal task execution sequence;
Figure QLYQS_6
in the field, the field represents the AGV completion position numbered kThe degree of fitness of the task is provided,
Figure QLYQS_7
representing a task set T incomplete at AGVs numbered k k Task k in j Distance to be travelled, R (k) j ,k j+1 ) Representing task k j And task k j+1 And a correlation distance between them. />
CN202210716146.0A 2022-06-22 2022-06-22 Multi-AGV online task allocation method and system Active CN115186878B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210716146.0A CN115186878B (en) 2022-06-22 2022-06-22 Multi-AGV online task allocation method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210716146.0A CN115186878B (en) 2022-06-22 2022-06-22 Multi-AGV online task allocation method and system

Publications (2)

Publication Number Publication Date
CN115186878A CN115186878A (en) 2022-10-14
CN115186878B true CN115186878B (en) 2023-05-16

Family

ID=83515927

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210716146.0A Active CN115186878B (en) 2022-06-22 2022-06-22 Multi-AGV online task allocation method and system

Country Status (1)

Country Link
CN (1) CN115186878B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116245257B (en) * 2023-05-06 2023-09-12 季华实验室 Multi-robot scheduling method and device

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107861511A (en) * 2017-11-07 2018-03-30 河北科技大学 A kind of control device and its control method for ground handling trolley
CN110264062A (en) * 2019-08-12 2019-09-20 南京邮电大学 Distributed more AGV dynamic task allocations and its paths planning method and system
CN110456633A (en) * 2019-06-29 2019-11-15 西南电子技术研究所（中国电子科技集团公司第十研究所） Airborne multi-platform distributed task scheduling distribution method
CN111474926A (en) * 2020-03-24 2020-07-31 浙江中烟工业有限责任公司 Waste smoke recovery method based on multiple AGV time window path optimization algorithm
CN112100899A (en) * 2020-09-22 2020-12-18 河南中烟工业有限责任公司 AGV multithreading operation global optimization method for cigarette auxiliary material scheduling
CN113870602A (en) * 2021-09-28 2021-12-31 湖南大学 Method and system for dispatching multiple AGV parking
CN114154819A (en) * 2021-11-20 2022-03-08 大连理工大学 Task similarity-based multi-AGV distributed scheduling method and system
CN114240076A (en) * 2021-11-26 2022-03-25 成都睿芯行科技有限公司 Multi-AGV task allocation method based on improved particle swarm algorithm

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10255582B2 (en) * 2016-08-18 2019-04-09 Amazon Technologies, Inc. AGV traffic management system
KR20200069925A (en) * 2018-12-07 2020-06-17 현대자동차주식회사 Automated guided vehicle control system and method thereof
US11780676B2 (en) * 2019-08-14 2023-10-10 Opex Corporation Systems and methods for managing inventory

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107861511A (en) * 2017-11-07 2018-03-30 河北科技大学 A kind of control device and its control method for ground handling trolley
CN110456633A (en) * 2019-06-29 2019-11-15 西南电子技术研究所（中国电子科技集团公司第十研究所） Airborne multi-platform distributed task scheduling distribution method
CN110264062A (en) * 2019-08-12 2019-09-20 南京邮电大学 Distributed more AGV dynamic task allocations and its paths planning method and system
CN111474926A (en) * 2020-03-24 2020-07-31 浙江中烟工业有限责任公司 Waste smoke recovery method based on multiple AGV time window path optimization algorithm
CN112100899A (en) * 2020-09-22 2020-12-18 河南中烟工业有限责任公司 AGV multithreading operation global optimization method for cigarette auxiliary material scheduling
CN113870602A (en) * 2021-09-28 2021-12-31 湖南大学 Method and system for dispatching multiple AGV parking
CN114154819A (en) * 2021-11-20 2022-03-08 大连理工大学 Task similarity-based multi-AGV distributed scheduling method and system
CN114240076A (en) * 2021-11-26 2022-03-25 成都睿芯行科技有限公司 Multi-AGV task allocation method based on improved particle swarm algorithm

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
一种为地面节点充电的多UAV任务分配与路线规划方法;胡洁;兰玉彬;欧阳帆;;电讯技术(第04期);第18-26页 *

Also Published As

Publication number Publication date
CN115186878A (en) 2022-10-14

Similar Documents

Publication Publication Date Title
Lin et al. mTS: Temporal-and spatial-collaborative charging for wireless rechargeable sensor networks with multiple vehicles
Wei et al. Intelligent parking garage EV charging scheduling considering battery charging characteristic
He et al. Optimal scheduling for charging and discharging of electric vehicles
CN110428165B (en) Electric vehicle charging scheduling method considering reservation and queuing in charging station
US20160380440A1 (en) Electric charging power management
Zhu et al. Joint transportation and charging scheduling in public vehicle systems—A game theoretic approach
Alizadeh et al. Optimized path planning for electric vehicle routing and charging
JP6457266B2 (en) Charging the battery group
Chekired et al. Fog-computing-based energy storage in smart grid: A cut-off priority queuing model for plug-in electrified vehicle charging
CN113505962B (en) Electric automobile navigation and charging reservation method considering complete journey
Chung et al. Intelligent charging management of electric vehicles considering dynamic user behavior and renewable energy: A stochastic game approach
CN115186878B (en) Multi-AGV online task allocation method and system
CN113283623A (en) Electric vehicle electric quantity path planning method compatible with energy storage charging pile
CN108944500B (en) Electric vehicle charging scheduling method based on distributed station joint control
WO2019085411A1 (en) Battery equalization method and system
Chen et al. Deadline scheduling for large scale charging of electric vehicles with renewable energy
Usman et al. A coordinated framework for optimized charging of EV fleet in smart grid
Theodoropoulos et al. A load balancing control algorithm for EV static and dynamic wireless charging
Vayá et al. Optimal bidding of plug-in electric vehicles in a market-based control setup
Tsaousoglou et al. A stochastic approximation method for price-based assignment of electric vehicles to charging stations
CN104755313A (en) Recharging of a pool of batteries
CN111429014A (en) Electric vehicle charging station selection method based on double-queue dynamic pricing model
Dang et al. An advanced framework for electric vehicles interaction with distribution grids based on q-learning
Gerding et al. Fair online allocation of perishable goods and its application to electric vehicle charging
CN113715669B (en) Ordered charging control method, system and equipment for electric automobile and readable storage medium

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