CN116911483B - Multi-AGV dynamic path planning method based on ant colony optimization algorithm - Google Patents

Multi-AGV dynamic path planning method based on ant colony optimization algorithm Download PDF

Info

Publication number
CN116911483B
CN116911483B CN202311189442.0A CN202311189442A CN116911483B CN 116911483 B CN116911483 B CN 116911483B CN 202311189442 A CN202311189442 A CN 202311189442A CN 116911483 B CN116911483 B CN 116911483B
Authority
CN
China
Prior art keywords
agv
busy
path
idle
task
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
CN202311189442.0A
Other languages
Chinese (zh)
Other versions
CN116911483A (en
Inventor
赵宏
袁锴薪
刘静
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.)
Guangzhou Institute of Technology of Xidian University
Original Assignee
Guangzhou Institute of Technology of Xidian 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 Guangzhou Institute of Technology of Xidian University filed Critical Guangzhou Institute of Technology of Xidian University
Priority to CN202311189442.0A priority Critical patent/CN116911483B/en
Publication of CN116911483A publication Critical patent/CN116911483A/en
Application granted granted Critical
Publication of CN116911483B publication Critical patent/CN116911483B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • 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
    • G06Q10/06313—Resource planning in a project environment
    • 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
    • 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
    • G06Q10/087—Inventory or stock management, e.g. order filling, procurement or balancing against orders

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Physics & Mathematics (AREA)
  • Development Economics (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Operations Research (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Game Theory and Decision Science (AREA)
  • Biophysics (AREA)
  • Accounting & Taxation (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Biodiversity & Conservation Biology (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
  • Educational Administration (AREA)
  • Finance (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Computational Linguistics (AREA)

Abstract

The invention discloses a multi-AGV dynamic path planning method based on an ant colony optimization algorithm, which comprises the steps that a task arrangement system issues a new scheduling request, a multi-AGV system acquires idle AGV, an idle AGV with minimum cost reaching the starting point of a scheduling task is selected to respond to the scheduling request, a minimum cost path reaching the starting point from the current position of the responding trolley and a minimum cost path reaching the end point from the starting point are planned, the path is issued to the trolley to be responded, and the state of the trolley is updated to be busy. The method comprises the steps of obtaining a minimum cost path through an ant colony optimization shortest path algorithm, a deadlock resolution algorithm and a conflict resolution algorithm; various conflict types are summarized, and different conflict resolution algorithms are designed according to different conditions, so that the automation, the intellectualization and the high efficiency of the multi-AGV scheduling system are realized.

Description

Multi-AGV dynamic path planning method based on ant colony optimization algorithm
Technical Field
The invention belongs to the field of multi-AGV dynamic path planning methods, and particularly relates to a multi-AGV dynamic path planning method based on an ant colony optimization algorithm.
Background
In recent years, with the development of the growth of information technology and industrial technology, the acceleration of the transformation of various industries including the manufacturing industry has been promoted. Intelligent, automated packaging manufacturing has become a necessary condition for achieving intelligent production. In the process of building intelligent warehouse, the transportation function of the intelligent device and the development of the matched intelligent scheduling algorithm become research investment hot spots of global related scientific research teams and large manufacturing companies. An automatic guided vehicle AGV as an intelligent transportation device may be called an important tie connecting intelligent production, intelligent factories, and intelligent logistics.
In conventional factory systems, the transportation scheduling of goods and parts mainly relies on fixed scheduling lines or manual handling, which can be barely handled when the transportation scheduling tasks of goods and parts are small. Along with the high-speed development of manufacturing industry and the upgrading and transformation of processing and manufacturing technology, the scheduling problem difficulty of massive goods scheduling demands in fixed storage or factories is gradually highlighted, and the traditional factory logistics modes of manual operation and conveyor belt transportation are difficult to meet the demands. Although these difficulties have been alleviated in recent years with the development of AGV hardware, existing schemes for serial handling of AGVs, in the face of scheduling problems in large complex warehouse, still exhibit debilitation when dealing with these problems, mainly as follows:
1. current task allocation methods lack an assessment of potential conflict situations, resulting in an uneven number of task executions among multiple AGVs. Some AGV carts respond to the issued task at high frequency while others remain idle for long periods of time.
2. The number of the existing AGVs participating in scheduling is usually small, the cooperation capacity and fault tolerance capacity of the AGVs are poor, the problems of conflict or deadlock generated in the multi-AGV scheduling process cannot be processed, and the degree of automation is poor.
3. The existing AGV scheduling method lacks intelligent prediction reasoning capability when a dynamic scheduling system is composed of multiple AGVs, and the generated AGV scheduling scheme lacks high efficiency and flexibility.
In the prior art, the chinese patent publication No. CN105938572a discloses a multi-automatic guided vehicle path planning method for preventing interference of a logistics storage system, and designs an ant colony particle swarm fusion induction method to perform path planning by combining an interference threshold, mainly planning paths through an ant colony algorithm and the like, so that the problems of conflict or deadlock and the like generated in a dispatching process are not disclosed and an intelligent prediction function is not realized.
Therefore, under the drive of intelligent manufacturing, a multi-AGV dynamic path planning method based on an ant colony optimization algorithm is needed to be provided, and real-time path planning and dynamic obstacle avoidance technology of a plurality of AGV trolleys are realized through an intelligent and efficient task allocation and scheduling scheme, so that technical support is provided for intelligent warehouse storage.
Disclosure of Invention
Aiming at the problems existing in the prior art, the invention provides a multi-AGV dynamic path planning method based on an ant colony optimization algorithm, which designs the task dispatch and potential conflict resolution technology of an AGV according to parallel double threads, and simultaneously resolves certain deadlock situations, thereby realizing the automation, the intellectualization and the high efficiency of a multi-AGV scheduling system.
The technical scheme of the invention is realized as follows:
a multi-AGV dynamic path planning method based on an ant colony optimization algorithm is applied to a task system and a path planning system, wherein the task system comprises a task arrangement system and a task scheduling system; the path planning system comprises a logistics storage environment, a multi-AGV trolley system and a conflict resolution system; the task system and the path planning system perform information interaction; the method comprises the following steps:
s1, a task arrangement system issues a new cargo scheduling request, and issues the cargo scheduling request to a task scheduling system;
judging whether the task arrangement system issues a new cargo scheduling request at the current moment, and executing step S2 if the new cargo scheduling request is generated; otherwise, the task scheduling system waits for 5 seconds, and judges whether the task scheduling system issues a new cargo scheduling request at the current moment again;
s2, acquiring all idle AGV trolleys in the warehouse by using a logistics warehouse environment and a multi-AGV trolley system;
judging whether an idle AGV trolley exists in the logistics warehouse at the current moment, and if the idle AGV trolley exists, executing a step S3; otherwise, the task scheduling system waits for 5 seconds, and judges whether an idle AGV trolley exists in the logistics storage at the current moment or not again;
S3, acquiring the positions of all the idle AGV trolleys, calculating the time cost from the position of each idle AGV trolley to the task starting point through a logistics storage environment, a multi-AGV trolley system and a conflict resolution system, and selecting an idle AGV trolley with the minimum time cost to respond to a cargo scheduling request;
s4, acquiring an optimal path of the idle AGV responding to the cargo scheduling request from the self-position to a task starting point according to the idle AGV responding to the cargo scheduling request in the step S3, and calculating an optimal path of the idle AGV responding to the cargo scheduling request from the task starting point to the task ending point according to a calculation method of time cost of scheduling all the idle AGVs from the self-position to the task starting point in the step S3;
s5, sending the optimal path from the self-position to the task starting point and the optimal path from the task starting point to the task key in the step S4 to the idle AGV trolley responding to the goods scheduling request, executing the goods scheduling task, and changing the running state of the idle AGV trolley responding to the goods scheduling request into busy.
Preferably, in step S3, according to the information of all the idle AGV trolleys, the positions of all the idle AGV trolleys are obtained, the time cost from the position of the idle AGV trolley to the task start point is calculated, and an idle AGV trolley with the minimum time cost is selected to respond to the cargo scheduling request, which specifically includes:
S31, acquiring the position information C of all idle AGV dollies AFj in the warehouse road network at the current moment AFj J=1, 2, … … f, f being the total number of all free AGV carts at the current time;
the storage road network is G= (V, E) and is graph data formed by a node set V and an edge set E; the attribute of the node is an entry node AP of a storage rack in storage or other nodes LM in a road network.
Specifically, the graph data of the storage road network is combined in the algorithm of estimating and resolving the conflict and resolving the deadlock, so that the method can be applied to different storage environments.
S32, planning a position C of each idle AGV AFj for the idle AGV according to an ant colony optimization shortest path algorithm AFj Optimal path ArriveSPath for scheduling to task Start Point S j ；
According to the average running speed V of the idle AGV AFj j Calculating the position C of the trolley from the position C AFj Time cost ArriveSTime scheduled to task Start Point S j Wherein, arriveSTime j =ArriveSPath j /V j ；
S33, the conflict resolution system reads the idle AGV AFj from the position C of the system according to a conflict resolution algorithm and a deadlock resolution algorithm AFj The collision type with other busy AGV ABi in the process of reaching the task starting point S, wherein i=1, 2, … … b, b is the total number of all busy AGV carts at the current moment, and the extra time cost ExtraTime is calculated according to the collision type j Then the idle AGV AFj is from its own position C AFj The time cost for reaching the task start point S is ArriveSTime j +ExtraTime j Selecting a production time cost ArriveSTime from all idle AGV carts AFj j +ExtraTime j The smallest free AGV truck AFj responds to the load dispatch request with an optimal path to the task start S of OptArriveSPath j* ；
Optimal path OptArriveSPath to task Start S j* Including idle AGV AFj responding to a load dispatch request from its own location C AFj* Optimal path ArriveSPath for scheduling to task Start Point S j* And the path corresponding to the deadlock resolution avoidance path and the collision avoidance path under the condition that the sum of time cost generated by the respective paths is minimum.
Preferably, in step S33, extra time cost ExtraTime is calculated according to the collision type j The method specifically comprises the following steps:
s331, acquiring current state information of all trolleys and executing task conditions: acquiring current position information C of all busy AGV dollies ABi from logistics storage environment and multi-AGV dolly system ABi Remaining path information remaininpath of executing task i All the position information C where the idle AGV AFj is currently located AFj ；
S332, setting avoidance order rules for all AGV trolleys: when collision occurs in the AGV small workshops, the AGV small trolley with low priority avoids the AGV small trolley with high priority;
S333, reading the AGV trolley with the deadlock according to the position information of the AGV trolley, calculating the extra time cost generated by each AGV trolley in the deadlock for resolving the deadlock, scheduling the AGV trolley with the minimum extra time cost to an idle node for avoiding, resolving the deadlock, and acquiring a deadlock resolving evasion path;
s334, according to the position information C of the busy AGV ABi ABi And remaining path information remaininpath i Position information C of idle AGV AFj The AGV trolley generating the conflict is read, and the conflict type comprises the conflict generated between the busy AGV trolley and the conflict generated between the busy AGV trolley and the idle AGV trolley;
executing a corresponding conflict avoidance scheme according to the conflict type to be generated, and acquiring a conflict avoidance path;
s335, adding all deadlock resolution avoidance paths and conflict avoidance paths to an optimal path ArriveSPath j In (a) and (b);
s336, according to a time cost calculation formula in the step S32, according to the average running speed V of the AGV AFj j Calculating the sum of time costs respectively generated by the deadlock resolution avoidance path and the conflict avoidance path as extra time cost extraTime j ；
With minimum extra time cost ExtraTime j The corresponding deadlock resolution avoidance path and the conflict avoidance path are optimal deadlock resolution avoidance path and conflict avoidance path.
Preferably, the ant colony optimization shortest path algorithm includes:
s321, initializing pheromones with the same size for each road network node according to the node connection relation of the storage road network, and initializing a population containing m ants, wherein the departure position is C AFj ；
S322, selecting an ant, and calculating the probability of the ant for neighboring nodes with different trend according to the intensity and the visibility of the pheromone of the neighbors of the current node, namely, the transition probability;
the smaller the distance between two adjacent nodes, the higher the visibility; the larger the pheromone value of the neighbor of the node is, the higher the probability that the subsequent ants select the node is;
s323, iteratively finding by adopting a roulette method according to the transition probabilityTo a position C where the AGV trolley is located AFj Calculating the pheromone value left by each node of the ant on the path to the path of the task starting point S, and dead the selected ant;
s324, iteratively selecting other ants, calculating the transition probabilities of the ants from the current node to different neighbor nodes, searching paths, and calculating the remained pheromone value until m ants all reach the task starting point S and die;
S325, updating the pheromone values of all nodes of the road network according to the pheromone values left by all ants; updating the slave position C obtained by once exploring m ants AFj The shortest path to the task start S;
s326, meeting the ant colony optimization termination condition, and outputting the position C of the idle AGV AFj from the position C AFj Current optimal path ArriveSPath for task Start S j The method comprises the steps of carrying out a first treatment on the surface of the If the ant colony optimization termination condition is not met, all the populations die, the populations of m ants are regenerated, and the positions C are on the road network for updating the pheromone values AFj And (5) iteratively searching the optimal path again.
Specifically, a path is searched according to pheromone and visibility based on an ant colony optimization shortest path algorithm, the pheromone value of the road network node is updated, and the shortest path is obtained in an iterative mode, namely an optimal path, so that real-time path planning of the AGV trolley is realized.
Preferably, the visibility is the reciprocal of the distance between the node where the ant is currently located and its neighbor.
Specifically, the probability of ants going to different neighbor nodes is calculated according to the combination of pheromones and visibility, and the shortest path algorithm of the ant colony is optimized.
Preferably, the deadlock is a closed avoidance loop generated by avoiding a plurality of AGV trolleys mutually, and all the AGV trolleys are in a pause state.
Specifically, the deadlock detection and resolution algorithm based on the closed loop resolution algorithm can cope with the deadlock situation which is difficult to avoid by some scheduling algorithms, and the barrier-free automatic scheduling of multiple AGVs is realized.
Preferably, the priority of the avoidance order rule is set as follows: the priority of all the idle AGV trolleys is lower than that of the busy AGV trolleys; the earlier all busy AGV carts begin to perform their scheduled tasks, the higher the priority of the busy AGV cart.
Specifically, through the setting of priority, the efficiency of the AGV trolley to execute tasks is improved, extra time cost generated by conflict is reduced, meanwhile, the optimal avoiding mode in the case of conflict is realized, and the high efficiency and the flexibility of an AGV trolley scheduling scheme are improved.
Preferably, the collision generated between the busy AGV trolley and the busy AGV trolley comprises single-point opposite collision, single-point T-shaped collision, single-side collision and multi-point collision; the collision generated between the busy AGV trolley and the idle AGV trolley comprises busy and idle collision;
the conflict avoidance scheme comprises a single-point opposite conflict avoidance scheme, a single-point T-shaped conflict avoidance scheme, a single-side conflict avoidance scheme, a multi-point conflict avoidance scheme and a busy and idle conflict avoidance scheme.
Specifically, various conflict types existing in the multi-AGV trolley system are summarized, so that the possible conflict or impending conflict generated during task execution can be avoided in advance, and the intelligent prediction reasoning capability of the dynamic scheduling system consisting of the multi-AGVs is realized.
Preferably, the single point of the opposite collision is the remaining path remain path of the busy AGV car ABm m And the remaining path remaininPath of busy AGV dolly ABn n A common path section starting from a first position exists, the node number of the common path section is equal to 1, and the running directions of ABm and ABn are opposite, so that single-point opposite collision occurs between two AGV carts;
the single point T-type conflict is the remaining path remain Path of the busy AGV car ABm m And the remaining path remaininPath of busy AGV dolly ABn n A common path section starting from a first position exists, the node number of the common path section is equal to 1, and the running directions of ABm and ABn are mutually perpendicular, so that a single-point T-shaped conflict occurs between two AGV trolleys;
the multipoint conflict is the remaining path remaininPath of the busy AGV car ABm m And the remaining path remaininPath of busy AGV dolly ABn n A common path section starting from a first position exists, the node number of the common path section is larger than 1, and the running directions of ABm and ABn are opposite, so that multipoint conflict occurs between two AGV carts;
The unilateral collision is that the node to be passed by the busy AGV trolley ABm is the node just passed by the busy AGV trolley ABn, and the node to be passed by the busy AGV trolley ABn is the node just passed by the busy AGV trolley ABm, so that the unilateral collision is generated between the two AGV trolleys;
the busy-free collision is the remaining path remaininPath of the idle AGV AFj occurring in the busy AGV ABn n On this, the busy AGV ABn will generate a busy collision with the idle AGV AFj.
Specifically, specific situations of conflict types in the warehouse road network are listed, and the scheduling efficiency of the AGV trolley is improved corresponding to a conflict resolution algorithm.
Preferably, the single-point opposite collision avoidance scheme is that single-point opposite collision occurs between two busy AGV carts, connectivity of the side where the cart with lower priority is located is broken, and a shortest path reaching a task start point or a task end point is planned for the AGV carts again;
the single-point T-shaped collision avoidance scheme is that single-point T-shaped collision occurs between two busy AGV trolleys, the AGV trolley with lower priority pauses first, and whether the AGV trolley with higher priority can pass through a T-shaped intersection or not; if the vehicle can pass through, the vehicle with lower priority pauses, so as to avoid the conflict, and if the vehicle can not pass through, the vehicle with higher priority pauses, so as to avoid the conflict;
The multipoint collision avoidance scheme is characterized in that multipoint collision occurs between two busy AGV trolleys, the meeting position Lmeet of the two AGV trolleys is obtained according to the running speed of the two AGV trolleys, a shortest path which bypasses the meeting position Lmeet is planned for the AGV trolley with lower priority, and the shortest path is sent to the AGV trolley with lower priority;
the unilateral collision avoidance scheme is that unilateral collision occurs between two busy AGV trolleys, connectivity of the side where the trolley with lower priority is located is broken, and a shortest path reaching a task starting point or a task end point is planned for the AGV trolley again;
the busy-idle collision avoidance scheme is that busy-idle collision is generated between a busy AGV trolley and an idle AGV trolley, and the idle AGV trolley is adjusted away from a collision generating position to the nearest idle point.
Specifically, different conflict resolution algorithms are designed according to different conditions, and a resolution scheme with the minimum time cost is selected from all conflict resolution schemes, so that the normal operation of the multi-AGV scheduling system can be quickly recovered.
Compared with the prior art, the invention has the following beneficial effects:
when dispatching a scheduling request of an AGV response system, the invention comprehensively considers the shortest paths of the trolley to the task starting point, the conflict possibly generated and various corresponding candidate conflict resolution schemes, and selects the AGV trolley response task requirement according to the time cost corresponding to the shortest paths and the time cost corresponding to the optimal conflict resolution scheme.
And decomposing the dispatching problem of the dispatching tasks in the warehouse in the multi-AGV dynamic system into two parallel independent task dispatching modules and a conflict resolution and deadlock resolution module. The task scheduling module aims at the minimum time cost, so that most of conflicts or impending conflicts possibly generated during task execution can be avoided. The conflict resolution and deadlock resolution module selects a resolution scheme with the minimum time cost from all conflict resolution schemes aiming at the possible conflicts, and the normal operation of the multi-AGV scheduling system is quickly restored.
Various conflict types existing in the multi-AGV trolley system are summarized, and different conflict resolution algorithms are designed according to different conditions. Meanwhile, the invention provides a deadlock detection and resolution algorithm based on a closed loop resolution algorithm, which can cope with the condition that some scheduling algorithms cannot avoid deadlock. The barrier-free automatic scheduling of the multiple AGVs is realized.
In task scheduling, considering that a feasible path can not be found in a given time by a traditional path planning algorithm such as Dijiestra and the like, the path searching basic algorithm based on the ant colony optimization algorithm gives high efficiency and flexibility to multi-AGV scheduling in path generation.
Compared with the traditional AGV trolley scheduling algorithm, the task dispatching and potential conflict resolution technology of the AGV is designed according to parallel double threads, so that the two modules independently and independently interact with an AGV trolley system. In the task allocation strategy, a thread in which a scheduling algorithm is located distributes tasks issued by a task arrangement system to an AGV trolley with minimum response cost in real time; in the path planning strategy, the threads where the conflict resolution and the deadlock resolution are located can intelligently capture potential conflicts and resolve. Meanwhile, for some deadlock situations, the deadlock digestion algorithm can automatically capture the current trolley state, and the deadlock digestion is completed in a heuristic manner, so that the automation, the intellectualization and the high efficiency of the multi-AGV scheduling system are realized.
Drawings
FIG. 1 is a flow chart of a multi-AGV dynamic path planning method based on an ant colony optimization algorithm provided by the invention;
FIG. 2 is a schematic diagram of a warehouse road network of the multi-AGV dynamic path planning method based on the ant colony optimization algorithm;
FIG. 3 is a deadlock diagram of the multi-AGV dynamic path planning method based on the ant colony optimization algorithm provided by the invention;
FIG. 4 is a waiting and avoiding conflict comparison chart from the time T to the time T+ExtraTime of the multi-AGV dynamic path planning method based on the ant colony optimization algorithm;
FIG. 5 is a schematic diagram of single point opposite collision of the method for planning a dynamic path of a multi-AGV based on the ant colony optimization algorithm according to embodiment 1 of the present invention;
FIG. 6 is a schematic view of resolving single-point opposite collision of a multi-AGV dynamic path planning method based on an ant colony optimization algorithm according to embodiment 1 of the present invention;
FIG. 7 is a schematic diagram of single point T-type collision of the multi-AGV dynamic path planning method based on the ant colony optimization algorithm according to embodiment 2 of the present invention;
FIG. 8 is a schematic view of single-point T-type conflict resolution of the method for planning a dynamic path of a multi-AGV based on an ant colony optimization algorithm according to embodiment 2 of the present invention;
FIG. 9 is a schematic diagram of a multi-point collision of a multi-AGV dynamic path planning method based on an ant colony optimization algorithm according to embodiment 3 of the present invention;
FIG. 10 is a schematic diagram of resolving multi-point conflicts in the multi-AGV dynamic path planning method based on the ant colony optimization algorithm according to embodiment 3 of the present invention;
FIG. 11 is a schematic diagram of single-side collision of a multi-AGV dynamic path planning method based on an ant colony optimization algorithm according to embodiment 4 of the present invention;
FIG. 12 is a schematic diagram showing a single-sided conflict resolution of the method for multi-AGV dynamic path planning based on the ant colony optimization algorithm according to embodiment 4 of the present invention;
Fig. 13 is a busy-idle collision schematic diagram of a multi-AGV dynamic path planning method based on an ant colony optimization algorithm provided in embodiment 5 of the present invention;
fig. 14 is a busy/idle collision resolution schematic diagram of a multi-AGV dynamic path planning method based on an ant colony optimization algorithm according to embodiment 5 of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1
As shown in fig. 1 to fig. 4, a multi-AGV dynamic path planning method based on an ant colony optimization algorithm is applied to a task system and a path planning system, wherein the task system comprises a task arrangement system and a task scheduling system; the path planning system comprises a logistics storage environment, a multi-AGV trolley system and a conflict resolution system; the task system and the path planning system perform information interaction; the method comprises the following steps:
S1, a task arrangement system issues a new cargo scheduling request, and issues the cargo scheduling request to a task scheduling system;
judging whether the task arrangement system issues a new cargo scheduling request at the current moment, and executing step S2 if the new cargo scheduling request is generated; otherwise, the task scheduling system waits for a fixed time, wherein the fixed time can be 5 seconds, and whether the task scheduling system issues a new cargo scheduling request or not is judged again at the current moment;
currently, the task orchestration system issues a new cargo scheduling request, and step S2 is executed.
S2, acquiring all idle AGV trolleys in the warehouse by using a logistics warehouse environment and a multi-AGV trolley system;
judging whether an idle AGV trolley exists in the logistics warehouse at the current moment, and if the idle AGV trolley exists, executing a step S3; otherwise, the task scheduling system waits for a fixed time, wherein the fixed time can be 5 seconds, and whether an idle AGV exists in the logistics storage at the current moment is judged again;
AGV is an abbreviation of Automated Guided Vehicle, meaning "automated guided vehicles", commonly used for transporting goods in storage and other environments, which transports the goods in storage by a transport instruction issued by a task scheduling system;
At the current moment, an idle AGV trolley exists in the logistics warehouse, and the step S3 is executed.
S3, acquiring the positions of all the idle AGV trolleys, calculating the time cost from the position of each idle AGV trolley to the task starting point through the logistics storage environment, the multi-AGV trolley system and the conflict resolution system, and selecting an idle AGV trolley with the minimum time cost to respond to a cargo scheduling request, wherein the method specifically comprises the following steps of:
s31, acquiring the position information C of all idle AGV dollies AFj in the warehouse road network at the current moment AFj J=1, 2, … … f, f being the total number of all free AGV carts at the current time;
at the current moment, there is an idle AGV f, and the position information corresponding to the AGV in the warehouse road network is C AFj ；
The position information corresponding to the warehouse road network is C AFj The warehouse road network G= (V, E) is composed ofGraph data formed by a node set V and an edge set E; the attribute of the node is an entry node AP of a storage rack in storage or other nodes LM in a road network;
the storage map of 5 AGV carts is shown in FIG. 2, wherein the map positions facing up, down, left and right in a shape of "<" are AP nodes, the other positions are LM nodes, and the squares marked with numbers represent different AGV carts;
The graph data of the storage road network is combined in the algorithm of estimating and resolving the conflict and resolving the deadlock, and the method can be applied to different storage environments.
S32, planning a position C of each idle AGV AFj for the idle AGV according to an ant colony optimization shortest path algorithm AFj Optimal path ArriveSPath for scheduling to task Start Point S j ；
The ant colony optimization shortest path algorithm comprises the following steps:
s321, initializing pheromones with the same size for each road network node according to the node connection relation of the storage road network, and initializing a population containing m ants, wherein the departure position is C AFj ；
S322, selecting an ant, and calculating the probability of the ant for neighboring nodes with different trend according to the intensity and the visibility of the pheromone of the neighbors of the current node, namely, the transition probability;
the smaller the distance between two adjacent nodes, the higher the visibility; the larger the pheromone value of the neighbor of the node is, the higher the probability that the subsequent ants select the node is;
the visibility is the reciprocal of the distance between the node where the ant is currently located and the neighbor;
according to the combination of pheromone and visibility, calculating the probability of ants going to different neighbor nodes, and optimizing the shortest path algorithm of the ant colony;
s323, iteratively finding a position C of the AGV trolley by adopting a roulette method according to the transfer probability AFj Calculating the pheromone value left by each node of the ant on the path to the path of the task starting point S, and dead the selected ant;
s324, iteratively selecting other ants, calculating the transition probabilities of the ants from the current node to different neighbor nodes, searching paths, and calculating the remained pheromone value until m ants all reach the task starting point S and die;
s325, updating the pheromone values of all nodes of the road network according to the pheromone values left by all ants; updating the slave position C obtained by once exploring m ants AFj The shortest path to the task start S;
s326, meeting the ant colony optimization termination condition, and outputting the position C of the idle AGV AFj from the position C AFj Current optimal path ArriveSPath for task Start S j The method comprises the steps of carrying out a first treatment on the surface of the If the ant colony optimization termination condition is not met, all the populations die, the populations of m ants are regenerated, and the positions C are on the road network for updating the pheromone values AFj And (5) iteratively searching the optimal path again.
The current m ants all die after reaching the task starting point S, and the slave position C obtained after one-time exploration of the m ants is updated AFj The shortest path to the task starting point S meets the ant colony optimization termination condition, and the idle AGV AFj is output from the position C of the idle AGV AFj AFj Current optimal path ArriveSPath for task Start S j ；
And searching a path according to the pheromone and the visibility based on the shortest path algorithm of ant colony optimization, updating the pheromone value of the road network node, and obtaining the shortest path, namely the optimal path, by an iterative mode, thereby realizing the real-time path planning of the AGV trolley.
According to the average running speed V of the idle AGV AFj j Calculating the position C of the trolley from the position C AFj Time cost ArriveSTime scheduled to task Start Point S j Wherein, arriveSTime j =ArriveSPath j /V j ；
S33, the conflict resolution system reads the idle AGV AFj from the position C of the system according to a conflict resolution algorithm and a deadlock resolution algorithm AFj The collision type with other busy AGV ABi in the process of reaching the task starting point S, wherein i=1, 2, … … b, b is the total number of all busy AGV at the current moment, and the additional time is calculated according to the collision typeThe extraTime j Then the idle AGV AFj is from its own position C AFj The time cost for reaching the task start point S is ArriveSTime j +ExtraTime j The generation time cost (i.e., arriveSTime) is selected from all idle AGV carts AFj j +ExtraTime j Minimum sum) minimum free AGV AFj responds to the load dispatch request by having an optimal path to the task start S of OptArriveSPath j* ；
Optimal path OptArriveSPath to task Start S j* Including idle AGV AFj responding to a load dispatch request from its own location C AFj* Optimal path ArriveSPath for scheduling to task Start Point S j* And the path corresponding to the deadlock resolution avoidance path and the collision avoidance path under the condition that the sum of time cost generated by the respective paths is minimum.
The conflict resolution algorithm executes a corresponding conflict avoidance scheme according to the conflict type to be generated, and a conflict avoidance path is obtained;
the deadlock resolution algorithm carries out deadlock resolution according to the deadlock situation and S333 to obtain a deadlock resolution evasion path;
wherein, extra time cost ExtraTime is calculated according to the conflict type j The method specifically comprises the following steps:
s331, acquiring current state information of all trolleys and executing task conditions: acquiring current position information C of all busy AGV dollies ABi from logistics storage environment and multi-AGV dolly system ABi Remaining path information remaininpath of executing task i All the position information C where the idle AGV AFj is currently located AFj ；
S332, setting avoidance order rules for all AGV trolleys: when collision occurs in the AGV small workshops, the AGV small trolley with low priority avoids the AGV small trolley with high priority;
The priority of the avoidance order rule is set as follows: the priority of all the idle AGV trolleys is lower than that of the busy AGV trolleys; the earlier all busy AGV dollies begin to perform their scheduled tasks, the higher the priority of the busy AGV dollies;
through the setting of priority, the efficiency of the AGV trolley executing tasks is improved, extra time cost generated by conflict is reduced, meanwhile, the optimal avoiding mode in the case of conflict is realized, and the high efficiency and the flexibility of an AGV trolley scheduling scheme are improved.
S333, reading the AGV trolley with the deadlock according to the position information of the AGV trolley, calculating the extra time cost generated by each AGV trolley in the deadlock for resolving the deadlock, scheduling the AGV trolley with the minimum extra time cost to an idle node for avoiding, resolving the deadlock, and acquiring a deadlock resolving evasion path;
the deadlock is a closed avoidance loop generated by avoiding a plurality of AGV trolleys, and all the AGV trolleys are in a pause state;
as in fig. 3, there is a diagram of the deadlock caused by three AGV carts, where cart 4 goes to T1014, cart 5 goes to T0613, and cart 3 goes to T0516.4 the car pauses and avoids 5 cars, and 5 cars pause and avoid 3 cars, and 3 cars pause and avoid 4 cars so as to form a closed avoidance loop, thereby causing deadlock. In order to resolve the deadlock, it can be seen that the trolley 3 cannot avoid the deadlock, the waiting party at the position where the trolley 4 needs to move upwards by two units (T0318) can avoid the deadlock, or the waiting party at the position where the trolley 5 moves leftwards by one unit (T0817) after moving downwards by the next unit can avoid the deadlock;
The deadlock detection and digestion algorithm based on the closed loop digestion algorithm can cope with the deadlock situation which is difficult to avoid by some scheduling algorithms, and the barrier-free automatic scheduling of multiple AGVs is realized.
S334, according to the position information C of the busy AGV ABi ABi And remaining path information remaininpath i Position information C of idle AGV AFj The AGV trolley generating the conflict is read, and the conflict type comprises the conflict generated between the busy AGV trolley and the conflict generated between the busy AGV trolley and the idle AGV trolley;
the conflict generated between the busy AGV trolley and the busy AGV trolley comprises single-point opposite conflict, single-point T-shaped conflict, single-side conflict and multi-point conflict; the collision generated between the busy AGV trolley and the idle AGV trolley comprises busy and idle collision;
the conflict avoidance scheme comprises a single-point opposite conflict avoidance scheme, a single-point T-shaped conflict avoidance scheme, a single-side conflict avoidance scheme, a multi-point conflict avoidance scheme and a busy-idle conflict avoidance scheme;
various conflict types existing in the multi-AGV trolley system are summarized, so that the possible conflict or impending conflict generated during task execution can be avoided in advance, and the intelligent prediction reasoning capability of a dynamic scheduling system consisting of multiple AGVs is realized;
As shown in FIG. 5, AF is used herein for ease of understanding 1 Indicating an idle AGV car, AF 2 Indicating busy AGV, since the avoidance order rule according to the priority is preset, both the two AGVs are busy or idle AGV, and can be classified into the above 5 collision types when in collision, and can be referred to for general use in the following examples 2-5;
the single-point opposite collision is AGV car AF 2 Remaining path remaininPath 2 (dotted line with arrow pointing to left) and AGV carriage AF 1 Remaining path remaininPath 1 (dotted line with arrow pointing to right) there is a common path segment from the first position and the number of nodes of the common path segment is equal to 1, and AF 1 And AF 2 If the running directions of the two AGV trolleys are opposite, single-point opposite collision occurs between the two AGV trolleys;
as shown in FIG. 6, the single point opposite collision avoidance scheme is that single point opposite collision occurs between two AGV carts to break the cart AF with lower priority 2 Connectivity of the edge is changed into AF of the AGV trolley 2 Planning a shortest path to a task start point or end point;
executing a corresponding conflict avoidance scheme according to the conflict type to be generated, and acquiring a conflict avoidance path;
s335, adding all deadlock resolution avoidance paths and conflict avoidance paths to an optimal path ArriveSPath j In (a) and (b);
s336, according to a time cost calculation formula in the step S32, according to the average running speed V of the AGV AFj j Calculating time generation generated by a deadlock resolution avoidance path and a conflict avoidance path respectivelyThe sum of the prices is extra time cost ExtraTime j ；
With minimum extra time cost ExtraTime j The corresponding deadlock resolution avoidance path and conflict avoidance path are the optimal (time cost is minimum) deadlock resolution avoidance path and conflict avoidance path in all feasible deadlock resolution schemes and conflict resolution schemes;
the extra time cost ExtraTime j The method can also comprise the time cost generated by the optimal avoidance scheme adopted for avoiding conflict, such as schemes of delayed departure, path re-search, idle node waiting outside the path and the like;
as shown in fig. 4, if at time t=t, the AGV car AF is idle 1 Beginning to execute the scheduling task according to the dotted line, the AGV car AB busy according to the solid line 1 A scheduling conflict is generated on the common path e (LM 1, LM 2) (the edge connecting LM1 and LM2 in the figure); to avoid collision, let the trolley AF 1 Pause waiting AB 1 The walk-through path length is ConflictLength 11 After e (LM 1, LM 2) at time t=t+extratime, car AF 1 And (5) starting. I.e. car AF 1 The scheduling cost of (1) includes the time cost ArriveSTime to the origin 1 And latency extraTime to avoid potential conflict generation 1 。
S4, acquiring an optimal path of the idle AGV responding to the cargo scheduling request from the self-position to a task starting point according to the idle AGV responding to the cargo scheduling request in the step S3, and calculating an optimal path of the idle AGV responding to the cargo scheduling request from the task starting point to the task ending point according to a calculation method of time cost of scheduling all the idle AGVs from the self-position to the task starting point in the step S3;
i.e. AGV trolley AFj. Optimal path optArriveepath from task start S to task end E j* ；
S5, sending the optimal path from the self-position to the task starting point and the optimal path from the task starting point to the task key in the step S4 to the idle AGV trolley responding to the cargo scheduling request, executing the cargo scheduling task, and changing the running state of the idle AGV trolley responding to the cargo scheduling request into busy;
i.e. path OptArriveSPath j* And path OptArriveepath j* And the operation state of the trolley is changed into busy state by sending the operation state to the AGV trolley and executing the cargo dispatching task.
When dispatching a scheduling request of an AGV response system, the invention comprehensively considers the shortest paths of the trolley to the task starting point, the conflict possibly generated and various corresponding candidate conflict resolution schemes, and selects the AGV trolley response task requirement according to the time cost corresponding to the shortest paths and the time cost corresponding to the optimal conflict resolution scheme.
And decomposing the dispatching problem of the dispatching tasks in the warehouse in the multi-AGV dynamic system into two parallel independent task dispatching modules and a conflict resolution and deadlock resolution module. The task scheduling module aims at the minimum time cost, so that most of conflicts or impending conflicts possibly generated during task execution can be avoided. And after the conflict is generated, the conflict resolution and deadlock resolution module selects a resolution scheme with the minimum time cost from all conflict resolution schemes, and the normal operation of the multi-AGV scheduling system is quickly restored.
Various conflict types existing in the multi-AGV trolley system are summarized, and different conflict resolution algorithms are designed according to different conditions. Meanwhile, the invention provides a deadlock detection and resolution algorithm based on a closed loop resolution algorithm, which can cope with the condition that some scheduling algorithms cannot avoid deadlock. The barrier-free automatic scheduling of the multiple AGVs is realized.
In task scheduling, considering that a feasible path can not be found in a given time by a traditional path planning algorithm such as Dijiestra and the like, the path searching basic algorithm based on the ant colony optimization algorithm gives high efficiency and flexibility to multi-AGV scheduling in path generation.
Compared with the traditional AGV trolley scheduling algorithm, the multi-AGV dynamic path planning technology based on the ant colony optimization algorithm designs the task dispatching and potential conflict resolution technology of the AGV according to parallel double threads, so that the two modules independently interact with the AGV trolley system. In the task allocation strategy, a thread in which a scheduling algorithm is located distributes tasks issued by a task arrangement system to an AGV trolley with minimum response cost in real time; in the path planning strategy, the threads where the conflict resolution and the deadlock resolution are located can intelligently capture potential conflicts and resolve. Meanwhile, for some deadlock situations, the deadlock digestion algorithm can automatically capture the current trolley state, and the deadlock digestion is completed in a heuristic manner, so that the automation, the intellectualization and the high efficiency of the multi-AGV scheduling system are realized.
Example 2
As shown in fig. 1 to fig. 4, a multi-AGV dynamic path planning method based on an ant colony optimization algorithm is applied to a task system and a path planning system, wherein the task system comprises a task arrangement system and a task scheduling system; the path planning system comprises a logistics storage environment, a multi-AGV trolley system and a conflict resolution system; the task system and the path planning system perform information interaction; the method comprises the following steps:
S1-S5, see specifically the steps S1-S5 in example 1, which are not described in detail herein.
In step S334, an embodiment of the conflict type being a single point T-type conflict is as follows:
as shown in FIG. 7, the single point T-collision is AGV car AF 1 Remaining path remaininPath 1 And AGV trolley AF 2 Remaining path remaininPath 2 There is a common path segment from the first position and the number of nodes of the common path segment is equal to 1, and AF 1 And AF 2 The running directions of the AGV are mutually perpendicular, and then single-point T-shaped conflict occurs between the two AGV trolleys;
as shown in FIG. 8, the single point T-type collision avoidance scheme is that a single point T-type collision occurs between two busy AGV trolleys, and the AGV trolleys AF with lower priority 2 AGV car AF with higher priority and pausing first 1 Whether the T-shaped intersection can be passed or not; AGV trolley AF 1 Through, the AGV trolley AF with lower priority level 2 Pause, avoid collisions.
Example 3
As shown in fig. 1 to fig. 4, a multi-AGV dynamic path planning method based on an ant colony optimization algorithm is applied to a task system and a path planning system, wherein the task system comprises a task arrangement system and a task scheduling system; the path planning system comprises a logistics storage environment, a multi-AGV trolley system and a conflict resolution system; the task system and the path planning system perform information interaction; the method comprises the following steps:
S1-S5, see specifically the steps S1-S5 in example 1, which are not described in detail herein.
In step S334, an embodiment of the collision type being a multipoint collision is as follows:
as shown in FIG. 9, the multipoint collision is AGV car AF 1 Remaining path remaininPath 1 And AGV trolley AF 2 Remaining path remaininPath 2 There is a common path segment from the first position and the number of nodes of the common path segment is greater than 1, and AF 1 And AF 2 If the traveling directions of the two AGV carts are opposite, a multipoint collision will occur between the two AGV carts.
As shown in FIG. 10, the multipoint collision avoidance scheme is that multipoint collision occurs between two busy AGV carts, and the meeting position Lmeet of the two AGV carts is obtained according to the running speed of the two AGV carts, and the multipoint collision avoidance scheme is that the AGV carts AF with lower priority is obtained 1 A shortest path bypassing the meeting position Lmeet is planned and sent to the AGV trolley AF with low priority 1 。
Example 4
As shown in fig. 1 to fig. 4, a multi-AGV dynamic path planning method based on an ant colony optimization algorithm is applied to a task system and a path planning system, wherein the task system comprises a task arrangement system and a task scheduling system; the path planning system comprises a logistics storage environment, a multi-AGV trolley system and a conflict resolution system; the task system and the path planning system perform information interaction; the method comprises the following steps:
S1-S5, see specifically the steps S1-S5 in example 1, which are not described in detail herein.
In step S334, an embodiment of the conflict type being a single-sided conflict is as follows:
as shown in FIG. 11, the single-side collision is the AF of the AGV trolley 1 About to pass through node and be AGV dolly AF 2 Just pass node and AGV dolly AF 2 The node to be passed is also AGV car AF 1 The single-side collision between two AGV trolleys occurs at the node which just passes through;
as shown in FIG. 12, the single-side collision avoidance scheme is that single-side collision occurs between two busy AGV carts, and the cart AF with lower priority is broken 1 Connectivity of the edge is changed into AF of the AGV trolley 1 A shortest path is planned to the start or end of the task.
Example 5
As shown in fig. 1 to fig. 4, a multi-AGV dynamic path planning method based on an ant colony optimization algorithm is applied to a task system and a path planning system, wherein the task system comprises a task arrangement system and a task scheduling system; the path planning system comprises a logistics storage environment, a multi-AGV trolley system and a conflict resolution system; the task system and the path planning system perform information interaction; the method comprises the following steps:
S1-S5, see specifically the steps S1-S5 in example 1, which are not described in detail herein.
In step S334, an embodiment that the conflict type is a busy-idle conflict is as follows:
as shown in FIG. 13, the busy-idle collision is an idle AGV car AF 2 Appear in busy AGV dolly AF 1 Remaining path remaininPath 1 On, busy AGV car AF 1 AGV trolley AF to be free 2 Generating busy and idle conflicts;
as shown in FIG. 14, the busy/idle collision avoidance scheme is a busy AGV car AF 1 And an idle AGV car AF 2 Generates busy and idle conflict between the AGV and the car AF 2 Tuning away the location where the conflict occurred to the nearest free point.
Variations and modifications to the above would be obvious to persons skilled in the art to which the invention pertains from the foregoing description and teachings. Therefore, the invention is not limited to the specific embodiments disclosed and described above, but some modifications and changes of the invention should be also included in the scope of the claims of the invention. In addition, although specific terms are used in the present specification, these terms are for convenience of description only and do not limit the present invention in any way.

Claims (7)

1. A multi-AGV dynamic path planning method based on an ant colony optimization algorithm is applied to a task system and a path planning system, wherein the task system comprises a task arrangement system and a task scheduling system; the path planning system comprises a logistics storage environment, a multi-AGV trolley system and a conflict resolution system; the task system and the path planning system perform information interaction; the method is characterized by comprising the following steps of:
S1, a task arrangement system issues a new cargo scheduling request, and issues the cargo scheduling request to a task scheduling system;
s2, acquiring all idle AGV trolleys in the warehouse by using a logistics warehouse environment and a multi-AGV trolley system;
s3, acquiring the positions of all the idle AGV trolleys, calculating the time cost from the position of each idle AGV trolley to the task starting point through the logistics storage environment, the multi-AGV trolley system and the conflict resolution system, and selecting one idle AGV trolley with the minimum time cost to respond to the cargo scheduling request; the method specifically comprises the following steps:
s31, acquiring the position information C of the idle AGV AFj in the warehouse road network at the current moment AFj J=1, 2, … … f, AFj is any one idle trolley, f is the total number of all idle AGV trolleys at the current moment;
the storage road network is G= (V, E) and is graph data formed by a node set V and an edge set E;
s32, planning a position C of the idle AGV AFj for the idle AGV according to an ant colony optimization shortest path algorithm AFj Optimal path ArriveSPath for scheduling to task Start Point S j ；
The ant colony optimization shortest path algorithm comprises the following steps:
s321, initializing pheromones with the same size for each road network node according to the node connection relation of the storage road network, and initializing a population containing m ants, wherein the departure position is C AFj ；
S322, selecting an ant, and calculating the probability of the ant for neighboring nodes with different trend according to the intensity and the visibility of the pheromone of the neighbors of the current node, namely, the transition probability;
the smaller the distance between two adjacent nodes, the higher the visibility; the larger the pheromone value of the neighbor of the node is, the higher the probability that the subsequent ants select the node is;
s323, iteratively finding a position C of the AGV trolley by adopting a roulette method according to the transfer probability AFj Calculating the pheromone value left by each node of the ant on the path to the path of the task starting point S, and dead the selected ant;
s324, iteratively selecting other ants, calculating the transition probabilities of the ants from the current node to different neighbor nodes, searching paths, and calculating the remained pheromone value until m ants all reach the task starting point S and die;
s325, updating the pheromone values of all nodes of the road network according to the pheromone values left by all ants; updating the slave position C obtained by once exploring m ants AFj The shortest path to the task start S;
s326, meeting the ant colony optimization termination condition, and outputting the position C of the idle AGV AFj from the position C AFj Current optimal path ArriveSPath for task Start S j The method comprises the steps of carrying out a first treatment on the surface of the If the ant colony optimization termination condition is not met, all the populations die, the populations of m ants are regenerated, and the positions C are on the road network for updating the pheromone values AFj Re-iterating to find an optimal path;
according to the average running speed V of the idle AGV AFj j Calculating the position C of the trolley from the position C AFj Time cost ArriveSTime scheduled to task Start Point S j Wherein, arriveSTime j =ArriveSPath j /V j ；
S33, the conflict resolution system reads the idle AGV AFj from the position C of the system according to a conflict resolution algorithm and a deadlock resolution algorithm AFj The collision type with the busy AGV in the process of reaching the task starting point S, and calculating extra time cost ExtraTime according to the collision type j Then the idle AGV AFj is from its own position C AFj Reach any oneThe time cost of the service origin S is ArriveSTime j +ExtraTime j Selecting a time cost ArriveSTime from all idle AGV carts j +ExtraTime j The smallest free AGV truck AFj responds to the load dispatch request with an optimal path to the task start S of OptArriveSPath j* The method comprises the steps of carrying out a first treatment on the surface of the The method specifically comprises the following steps:
s331, acquiring current state information of all trolleys and executing task conditions: obtaining current position information C of busy AGV ABi from logistics storage environment and multi-AGV system ABi Remaining path information remaininpath of executing task i All the position information C where the idle AGV AFj is currently located AFj The method comprises the steps of carrying out a first treatment on the surface of the Wherein i=1, 2, … … b, ABi is any one busy car, b is the total number of all busy AGV cars at the current time;
s332, setting avoidance order rules for all AGV trolleys: when collision occurs in the AGV small workshops, the AGV small trolley with low priority avoids the AGV small trolley with high priority;
s333, reading the AGV trolley with the deadlock according to the position information of the AGV trolley, calculating the extra time cost generated by each AGV trolley in the deadlock for resolving the deadlock, scheduling the AGV trolley with the minimum extra time cost to an idle node for avoiding, resolving the deadlock, and acquiring a deadlock resolving evasion path;
s334, according to the position information C of the busy AGV ABi ABi And remaining path information remaininpath i Position information C of idle AGV AFj The AGV trolley generating the conflict is read, and the conflict type comprises the conflict generated between the busy AGV trolley and the conflict generated between the busy AGV trolley and the idle AGV trolley;
executing a corresponding conflict avoidance scheme according to the conflict type to be generated, and acquiring a conflict avoidance path;
s335, adding all deadlock resolution avoidance paths and conflict avoidance paths to an optimal path ArriveSPath j In (a) and (b);
s336, according to a time cost calculation formula in the step S32, according to the average running speed V of the idle AGV AFj j Computing deadlock resolutionThe sum of time costs respectively generated by the avoidance path and the collision avoidance path is extra time cost extraTime j ；
S4, acquiring an optimal path of the idle AGV responding to the cargo scheduling request from the self-position to a task starting point according to the idle AGV responding to the cargo scheduling request in the step S3, and calculating an optimal path of the idle AGV responding to the cargo scheduling request from the task starting point to the task ending point according to a calculation method of time cost of scheduling all the idle AGVs from the self-position to the task starting point in the step S3;
s5, sending the optimal path from the self-position to the task start point and the optimal path from the task start point to the task end point in the step S4 to the idle AGV trolley responding to the cargo scheduling request, executing the cargo scheduling task, and changing the running state of the idle AGV trolley responding to the cargo scheduling request into busy.
2. The multi-AGV dynamic path planning method according to claim 1 wherein the visibility is the inverse of the distance between the current node of the ant and its neighbors.
3. The multi-AGV dynamic path planning method based on the ant colony optimization algorithm according to claim 1, wherein the deadlock is a closed avoidance loop generated by avoiding a plurality of AGV trolleys mutually, and all the AGV trolleys are in a pause state.
4. The multi-AGV dynamic path planning method according to claim 1, wherein the priority of the avoidance order rule is set as follows: the priority of all the idle AGV trolleys is lower than that of the busy AGV trolleys; the earlier all busy AGV carts begin to perform their scheduled tasks, the higher the priority of the busy AGV cart.
5. The multi-AGV dynamic path planning method according to claim 1, wherein the collision generated between the busy AGV trolley ABm and the busy AGV trolley ABn includes a single point opposite collision, a single point T-type collision, a single side collision and a multi-point collision; the collision generated between the busy AGV cart ABn and the idle AGV cart AFj includes a busy-idle collision; wherein m and n are two busy AGV carts that generate a conflict;
the conflict avoidance scheme comprises a single-point opposite conflict avoidance scheme, a single-point T-shaped conflict avoidance scheme, a single-side conflict avoidance scheme, a multi-point conflict avoidance scheme and a busy and idle conflict avoidance scheme.
6. The multi-AGV dynamic path planning method according to claim 5 wherein said single point of opposite collision is a remaining path of busy AGV car ABm m And the remaining path remaininPath of busy AGV dolly ABn n A common path section starting from a first position exists, the node number of the common path section is equal to 1, and the running directions of ABm and ABn are opposite, so that single-point opposite collision occurs between two AGV carts;
the single point T-type conflict is the remaining path remain Path of the busy AGV car ABm m And the remaining path remaininPath of busy AGV dolly ABn n A common path section starting from a first position exists, the node number of the common path section is equal to 1, and the running directions of ABm and ABn are mutually perpendicular, so that a single-point T-shaped conflict occurs between two AGV trolleys;
the multipoint conflict is the remaining path remaininPath of the busy AGV car ABm m And the remaining path remaininPath of busy AGV dolly ABn n A common path section starting from the first-position exists, the node number of the common path section is larger than 1, and the running directions of ABm and ABn are opposite, so that multi-point conflict occurs between two AGV carts;
the unilateral collision is that the node to be passed by the busy AGV trolley ABm is the node just passed by the busy AGV trolley ABn, and the node to be passed by the busy AGV trolley ABn is the node just passed by the busy AGV trolley ABm, so that the unilateral collision is generated between the two AGV trolleys;
The busy-idle conflict is that the idle AGV AFj appears in the busy AGVRemaining path remaininPath for ABn n On this, the busy AGV ABn will generate a busy collision with the idle AGV AFj.
7. The multi-AGV dynamic path planning method based on the ant colony optimization algorithm according to claim 5, wherein the single-point opposite collision avoidance scheme is that single-point opposite collision occurs between two busy AGV carts, connectivity of the side where the cart with lower priority is located is broken, and a shortest path reaching a task starting point or a task end point is planned for the AGV carts again;
the single-point T-shaped collision avoidance scheme is that single-point T-shaped collision occurs between two busy AGV trolleys, the AGV trolley with lower priority pauses first, and whether the AGV trolley with higher priority can pass through a T-shaped intersection or not; if the vehicle can pass through, the vehicle with lower priority pauses, so as to avoid the conflict, and if the vehicle can not pass through, the vehicle with higher priority pauses, so as to avoid the conflict;
the multipoint collision avoidance scheme is characterized in that multipoint collision occurs between two busy AGV trolleys, the meeting position Lmeet of the two AGV trolleys is obtained according to the running speed of the two AGV trolleys, a shortest path which bypasses the meeting position Lmeet is planned for the AGV trolley with lower priority, and the shortest path is sent to the AGV trolley with lower priority;
The unilateral collision avoidance scheme is that unilateral collision occurs between two busy AGV trolleys, connectivity of the side where the trolley with lower priority is located is broken, and a shortest path reaching a task starting point or a task end point is planned for the AGV trolley again;
the busy-idle collision avoidance scheme is that busy-idle collision is generated between a busy AGV trolley and an idle AGV trolley, and the idle AGV trolley is adjusted away from a collision generating position to the nearest idle point.
CN202311189442.0A 2023-09-15 2023-09-15 Multi-AGV dynamic path planning method based on ant colony optimization algorithm Active CN116911483B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311189442.0A CN116911483B (en) 2023-09-15 2023-09-15 Multi-AGV dynamic path planning method based on ant colony optimization algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311189442.0A CN116911483B (en) 2023-09-15 2023-09-15 Multi-AGV dynamic path planning method based on ant colony optimization algorithm

Publications (2)

Publication Number Publication Date
CN116911483A CN116911483A (en) 2023-10-20
CN116911483B true CN116911483B (en) 2024-02-23

Family

ID=88353553

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311189442.0A Active CN116911483B (en) 2023-09-15 2023-09-15 Multi-AGV dynamic path planning method based on ant colony optimization algorithm

Country Status (1)

Country Link
CN (1) CN116911483B (en)

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114692939A (en) * 2021-12-06 2022-07-01 西安电子科技大学广州研究院 Multi-AGV task scheduling method based on double-layer strategy

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20210103286A1 (en) * 2019-10-04 2021-04-08 Hong Kong Applied Science And Technology Research Institute Co., Ltd. Systems and methods for adaptive path planning

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114692939A (en) * 2021-12-06 2022-07-01 西安电子科技大学广州研究院 Multi-AGV task scheduling method based on double-layer strategy

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
多自动导引车路径规划的诱导蚁群粒子群算法;李军军;许波桅;杨勇生;吴华锋;;计算机集成制造系统(第12期);第201-210页 *
改进蚁群算法在AGV路径规划中的应用;胡春阳 等;计算机工程与应用;第56卷(第8期);第270-278页 *
李腾.《系统工程》.2023,第1-13页. *

Also Published As

Publication number Publication date
CN116911483A (en) 2023-10-20

Similar Documents

Publication Publication Date Title
CN109991977B (en) Path planning method and device for robot
CN110209485B (en) Dynamic avoidance method for multiple robots during cooperative operation
RU2589869C2 (en) Method and system for efficient scheduling for plurality of automated nonholonomic vehicles using scheduler of coordinated routes
CN112833905B (en) Distributed multi-AGV collision-free path planning method based on improved A-x algorithm
CN111633655B (en) Traffic scheduling method for distributed autonomous mobile robot
CN113074728B (en) Multi-AGV path planning method based on jumping point routing and collaborative obstacle avoidance
CN111596658A (en) Multi-AGV collision-free operation path planning method and scheduling system
CN113780633B (en) Complex environment-oriented multi-AGV intelligent cooperative scheduling method with real-time conflict resolution function
CN111638717A (en) Design method of distributed autonomous robot traffic coordination mechanism
CN114489062B (en) Workshop logistics-oriented multi-automatic guided vehicle distributed dynamic path planning method
CN113075927A (en) Storage latent type multi-AGV path planning method based on reservation table
CN114692939A (en) Multi-AGV task scheduling method based on double-layer strategy
CN112016811A (en) AGV intelligent scheduling system and method based on reinforcement learning
Fan et al. Time window based path planning of multi-AGVs in logistics center
CN115981264A (en) AGV scheduling and quantity combined optimization method considering conflicts
Lian et al. A probabilistic time-constrained based heuristic path planning algorithm in warehouse multi-AGV systems
CN114840001A (en) Multi-vehicle collaborative track planning method in closed environment
CN116755401A (en) Multi-unmanned forklift scheduling control method comprising path planning and vehicle passing strategy
CN116911483B (en) Multi-AGV dynamic path planning method based on ant colony optimization algorithm
CN114326610A (en) AGV operation optimization system and method based on double-layer space-time network structure
CN114326621A (en) Group intelligent airport trolley dispatching method and system based on layered architecture
Tai et al. A time-efficient approach to solve conflicts and deadlocks for scheduling AGVs in warehousing applications
CN112598316B (en) Material distribution and cooperative scheduling method for co-track double AGVs
CN114415675B (en) Multi-AGV planning method and system based on intelligent prediction and error recognition
Guo et al. Warehouse AGV path planning based on Improved A* algorithm

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