CN115587679A - AGV optimization scheduling method for intelligent warehouse - Google Patents

AGV optimization scheduling method for intelligent warehouse Download PDF

Info

Publication number
CN115587679A
CN115587679A CN202211291544.9A CN202211291544A CN115587679A CN 115587679 A CN115587679 A CN 115587679A CN 202211291544 A CN202211291544 A CN 202211291544A CN 115587679 A CN115587679 A CN 115587679A
Authority
CN
China
Prior art keywords
trolley
shelf
lion
agv
path
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.)
Withdrawn
Application number
CN202211291544.9A
Other languages
Chinese (zh)
Inventor
董辉
高彩云
陈云
林文杰
周祥清
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.)
Zhejiang University of Technology ZJUT
Original Assignee
Zhejiang University of Technology ZJUT
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 Zhejiang University of Technology ZJUT filed Critical Zhejiang University of Technology ZJUT
Priority to CN202211291544.9A priority Critical patent/CN115587679A/en
Publication of CN115587679A publication Critical patent/CN115587679A/en
Withdrawn legal-status Critical Current

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"
    • 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
    • G06Q10/06311—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
    • 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

Abstract

The invention discloses an AGV optimization scheduling method of an intelligent warehouse, which comprises the following steps: constructing an AGV dispatching model for the warehouse AGV picking system based on a grid modeling method; solving an AGV scheduling model by utilizing a lion group algorithm to generate an initial feasible path; and detecting path points with conflicts on the basis of the initial feasible path, and then adjusting the path points with conflicts by using the AGV priority to finally obtain an efficient and feasible global AGV scheduling scheme. The AGV optimization scheduling method for the intelligent warehouse provides a two-stage optimization method for solving the problem of scheduling of the trolleys of the warehouse, and improves the efficiency and accuracy of scheduling of the warehouse.

Description

AGV optimization scheduling method for intelligent warehouse
Technical Field
The invention belongs to the technical field of warehouse scheduling, and particularly relates to an AGV optimal scheduling method for an intelligent warehouse.
Background
In the life cycle of products, most of the produced products need to be packaged in a warehouse, and then the products are transported to a specified position through an AGV trolley and are taken out of the warehouse.
The factory uses the warehouse, compared with the traditional warehouse, the biggest characteristic is that the conversion from 'person to goods' to 'goods to person' is realized, and the method can be roughly described as follows: the system distributes the goods shelf moving tasks to be picked to each moving trolley, the trolleys can automatically move to the position below the goods shelf along a near route by walking and steering in a narrow space, then the goods shelf is transported to the operation platform, and the goods shelf is arranged in a queue in a specified area to wait for manual operation after reaching a target place. After the manual operation is finished, the trolley transports the goods shelf back to the original position, and then the trolley automatically returns to the initial parking area or continuously finishes the next scheduling task. In this intelligence warehouse, the staff need not push to choose the freight train and shuttle between goods shelves and choose the goods, only need stand in operation platform department, choose the corresponding quantity of corresponding product on the goods shelves according to the instruction of laser, put into corresponding order container, can realize the high-efficient acquisition in the short time to all order commodities, promoted the efficiency and the rate of accuracy that the warehouse picked the goods, also reduced workman's intensity of labour simultaneously.
The AGV is the most important transport equipment in intelligent warehouse, generally small in size is the puck type, can lift up the goods shelves of 3000 pounds of weight, advances according to the road of regulation through scanning bar code on the ground, takes the goods shelves to freely shuttle between the warehouse, transports the goods shelves of commodity place to operation platform according to wireless instruction. The AGV comprises the following parts: the lifting device, the trolley can lift the goods shelf off the ground through the screw device; the collision detection system is characterized in that the trolley is provided with an infrared sensor, so that other peripheral objects can be rapidly detected, and the trolley can be rapidly stopped to avoid collision if obstacles exist; the navigation system is characterized in that a camera is arranged on a trolley and used for reading a bar code at the bottom of a goods shelf and a bar code on the ground; the driving control system realizes bidirectional controllable driving of the trolley, and receives and executes tasks such as guidance, path selection and the like of the central control system on the robot trolley; and the charging system provides a trolley power support guarantee.
The intelligent AGV scheduling problem of the warehouse is mainly characterized by the following two aspects: path planning of the AGVs and collision avoidance between the AGVs. The path planning problem includes: and determining the position of a target shelf to be accessed by the AGV, arranging the target shelf to a trolley which correspondingly executes a moving task, and generating a feasible path of the trolley according to the road condition requirement. Collision avoidance problems include: and detecting and adjusting the global initial path to generate a collision-free path in a preset state, and adjusting collision avoidance again under the condition of emergencies generated in the system operation.
At present, the mainstream path planning algorithms at home and abroad mainly comprise a local search algorithm, a simulated annealing algorithm, a tabu search algorithm, a wolf colony algorithm, an ant colony algorithm and the like, wherein part of the algorithms have better running time advantages in the algorithms, but the local optimal result is easy to step in.
The traditional lion group algorithm has the advantages of low space complexity, capability of quickly approaching to a local extreme value, simple model, less parameters, high convergence speed and the like, but has the defects of parameter dependence, poor robustness, low local optimization efficiency, central trend and the like.
Disclosure of Invention
The invention aims to provide an AGV optimization scheduling method for an intelligent warehouse, provides a two-stage optimization method for solving the problem of scheduling of a warehouse trolley, and improves the efficiency and accuracy of scheduling of the warehouse.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
an AGV optimal scheduling method of an intelligent warehouse comprises the following steps:
step 1, constructing an AGV dispatching model for an AGV picking system of a warehouse based on a grid modeling method;
step 2, solving an AGV scheduling model by utilizing a lion group algorithm to generate an initial feasible path;
step 3, detecting path points with conflicts on the basis of the initial feasible path, then adjusting the path points with conflicts by using the AGV priority, and finally obtaining an efficient and feasible global AGV dispatching scheme;
wherein, the adult lion in the lion subgroup algorithm is determined as follows:
assuming that L lions form a population in the D-dimensional search space, where the number of adult lions is M, then
Figure BDA0003898032780000021
Wherein M = L β, the comparative example parameter β is adjusted by chaotic mapping:
β＝2r 2 ·sin(π·r)
wherein r is a random number between [0,1 ].
Several alternatives are provided below, but not as an additional limitation to the above general solution, but merely as a further addition or preference, each alternative being combinable individually for the above general solution or among several alternatives without technical or logical contradictions.
Preferably, the building of the AGV dispatching model for the warehouse AGV sorting system based on the grid building method includes:
the warehouse AGV picking system is described as follows: the system is provided with m trolleys and n shelves, the movement area of each trolley is marked as a two-dimensional plane O, a rectangular coordinate system is established for the two-dimensional plane O, the lower left corner is taken as a coordinate origin O, an X axis is established transversely, a Y axis is established longitudinally, the width required by trolley walking is taken as a unit, the plane is divided into grids, the grids are numbered in sequence, and the trolleys and the shelves in the warehouse respectively occupy one grid;
putting the warehouse into a rectangular coordinate system: numbering each shelf and obtaining corresponding coordinate information; numbering each trolley and obtaining corresponding coordinate information; numbering each platform for parking the trolley and obtaining corresponding coordinate information, wherein grids occupied by the three types of entities cannot be listed in the range of the channel, and other grids are free grids and can be used for the trolley to pass through;
therefore, an AGV dispatching model is constructed by taking the minimization of the picking time of the trolley with the longest picking path as a target, and the target function is as follows:
Figure BDA0003898032780000031
where Z represents the pick time for the AGV with the longest pick path, N = {1,2,3, …, N } represents the set of all shelves, K = {1,2,3, …, m } represents the set of all carts,
Figure BDA0003898032780000032
representing a binary decision variable, if the cart k transports the goods shelf j after transporting the goods shelf i, the value is 1, otherwise, the value is 0,
Figure BDA0003898032780000033
representing the time required for cart k to travel the shortest polygonal line path from shelf i to shelf j,
Figure BDA0003898032780000034
representing a binary decision variable, taking the value of 1 if the goods shelf i is transported by the trolley k, or else, taking the value of 0,S i Indicating the required service time for shelf i.
Preferably, the AGV scheduling model has the following constraints:
(1) Cart k starts from the initial stop:
Figure BDA0003898032780000035
in the formula (I), the compound is shown in the specification,
Figure BDA0003898032780000036
representing a binary decision variable, if the trolley k starts from an initial stop station 0 to a goods shelf j, taking the value as 1, otherwise, taking the value as 0;
(2) Car k finally returns to the initial stop:
Figure BDA0003898032780000037
in the formula (I), the compound is shown in the specification,
Figure BDA0003898032780000038
representing a binary decision variable, if the trolley k returns to the initial stop station 0 from the goods shelf i, the value is 1, otherwise, the value is 0;
(3) Each shelf is accessed only once:
Figure BDA0003898032780000039
Figure BDA0003898032780000041
(4) Line balance constraint, the inlet and outlet flow of each point is equal:
Figure BDA0003898032780000042
where a = {0,1,2, …, a } represents the set of all points, including all shelves and all carts, where 0 represents the initial stop point of the cart and a = n + m;
Figure BDA0003898032780000043
representing a binary decision variable, and if the trolley k enters the point h from the point q, taking the value of 1, otherwise, taking the value of 0;
Figure BDA0003898032780000044
representing a binary decision variable, if the trolley k starts from a point h to a point p, taking the value of 1, otherwise, taking the value of 0;
(5) If cart k goes from shelf j to shelf i, then shelf i is serviced by cart k:
Figure BDA0003898032780000045
(6) The time it takes for car k to move from shelf i to shelf j is equal to the fold line distance between shelves divided by the car k travel speed:
Figure BDA0003898032780000046
in the formula (X) i ,Y i ) Is the position coordinate of the shelf i, (X) j ,Y j ) Is the position coordinate, V, of the goods shelf j k The running speed of the trolley k;
(7) The time required for cart k to travel from the initial stop to shelf j is equal to the polyline distance between the initial stop and shelf j divided by the travel speed of cart k:
Figure BDA0003898032780000047
in the formula (AGV _ X) k0 ,AGV_Y k0 ) Coordinates of the initial stop of cart k;
(8) The time required for cart k to return from shelf i to the initial stop point is equal to the polyline distance between shelf i and the initial stop point divided by the travel speed of cart k:
Figure BDA0003898032780000048
(9) The service time required by the goods shelf i is equal to the time required by the goods shelf to move from the storage position to the goods picking platform to walk the shortest broken line path, the time required by the goods shelf to pick goods manually on the goods picking platform is added, and the time required by the goods shelf to move back from the goods picking platform to the storage position to walk the shortest broken line path is added:
Figure BDA0003898032780000049
in the formula (I), the compound is shown in the specification,
Figure BDA0003898032780000051
representing the time required to move the shelf i from the storage location to the picking platform along the shortest polyline path,
Figure BDA0003898032780000052
show goods shelvesi the time of manual picking at the picking platform,
Figure BDA0003898032780000053
representing the time required for moving the goods shelf i from the goods picking platform back to the storage position to walk the shortest broken line path;
(10) The time for the trolley k to complete all transportation tasks and return to the initial stop is equal to the trolley k walking time plus the trolley service shelf time:
Figure BDA0003898032780000054
(11) The time for the shelf i to start moving is equal to the starting moving time of the last shelf served by the trolley k plus the service time of the last shelf plus the walking time between the two shelves:
Figure BDA0003898032780000055
in the formula, A i Indicates the time at which the shelf i starts to move, A j Indicates the time at which the shelf j starts to move, S j Indicating the required service time for the shelf j,
Figure BDA0003898032780000056
representing a binary decision variable, if the cart k transports the goods shelf i after transporting the goods shelf j, the value is 1, otherwise, the value is 0,
Figure BDA0003898032780000057
the time required for the trolley k to move from the shelf j to the shelf i to take the shortest broken line path is represented;
(12) And (3) eliminating the sub-loop:
Figure BDA0003898032780000058
Figure BDA0003898032780000059
in the formula (I), the compound is shown in the specification,
Figure BDA00038980327800000510
representing the order of points visited in the path of cart k from point q to shelf i,
Figure BDA00038980327800000511
indicating the order of points visited in the path of cart k from point q to point q on shelf i,
Figure BDA00038980327800000512
representing the order of points visited in the path of cart k from point q to point p,
Figure BDA00038980327800000513
and (3) representing a binary decision variable, wherein if the trolley k enters a point q from a point p, the value is 1, and if not, the value is 0.
Preferably, the using the lion group algorithm to solve the AGV scheduling model to generate the initial feasible path includes:
step 2.1, initializing lion group algorithm related parameters, including: randomly generating the positions of L lions, taking a goods shelf pick-and-place position and a trolley as the position of each lion, and obtaining an initial coding sequence by adopting a matrix mode for position coding;
2.2, calculating a fitness function value of the initial coding sequence according to an objective function of the AGV scheduling model, and calculating a global optimal solution and an individual historical optimal solution of the lion group;
step 2.3, updating the individual positions of the male lion, the female lion and the young lion in the lion group;
step 2.4, calculating a fitness function value of each individual, and updating a global optimal solution and an individual historical optimal solution;
step 2.5, randomly selecting a candidate solution set in the setting field of the current global optimal solution;
step 2.6, judging whether the current candidate solution exists in a taboo table, if so, jumping to step 2.7, otherwise, jumping out of the current iteration;
step 2.7, calculating a fitness function value of the candidate solution, comparing the fitness function value with the fitness function value of the current global optimal solution, if the fitness function value of the candidate solution is better than the fitness function value of the current global optimal solution, updating the current candidate solution into the current global optimal solution, and updating a tabu table;
step 2.8, if the maximum tabu search times is reached, outputting the current global optimal solution as an initial feasible path, otherwise, returning to the step 2.5;
and 2.9, if the maximum iteration times are reached, outputting the global optimal solution as an initial feasible path, otherwise, returning to the step 2.2.
Preferably, the updating of the individual positions of the male lion, the female lion and the young lion in the lion group comprises:
the M adult lion has 1 lion king, the first adult lion is recorded as the lion king, and the position of the lion king is recorded as x l ＝(x l1 x l2 ... x lD ) Wherein l is more than or equal to 1 and less than or equal to M, and the updating position of the lion king is updated as follows:
Figure BDA0003898032780000061
in the formula (I), the compound is shown in the specification,
Figure BDA0003898032780000062
for the updated position of lion king, gamma is a random number generated according to normal distribution N (0,1), g t For a global optimal position at the t-th generation,
Figure BDA0003898032780000063
the historical optimal position of the ith individual in the t generation;
position memory of the female lion in the number of M-1 and the v
Figure BDA0003898032780000064
Wherein v is more than or equal to 1 and less than or equal to M and v is not equal to l, female lion updates individual position
Figure BDA0003898032780000065
As follows:
Figure BDA0003898032780000066
in the formula (I), the compound is shown in the specification,
Figure BDA0003898032780000067
for the historical optimal position of the v-th individual in the t generation,
Figure BDA0003898032780000068
is the historical optimal position of another female lion randomly selected from the t-th female lion group, i.e. c ≠ v, alpha f A nonlinear function which changes from large to small in a search space;
the number of the young lion in the lion group is L-M, and the position of the w-th young lion is recorded as
Figure BDA0003898032780000069
W is more than or equal to 1 and less than or equal to L-M, and the position of the young lion is updated
Figure BDA00038980327800000610
As shown in the following formula:
Figure BDA0003898032780000071
Figure BDA0003898032780000072
Figure BDA0003898032780000073
wherein q is based on uniform distribution of U [0,1]The random number is generated by the random number generator,
Figure BDA0003898032780000074
for the child lion to follow the tth historical best position of the mother lion,
Figure BDA0003898032780000075
is the historical optimal position of the w-th individual in the t generation, alpha' c Disturbance factor of moving range of young lion, T max Is the maximum iteration number of the algorithm, t is the current iteration number,
Figure BDA0003898032780000076
in order to use the reverse position after the reverse elite algorithm, namely the position after the young lion is driven out of the group
Figure BDA0003898032780000077
And
Figure BDA0003898032780000078
the minimum value mean value and the maximum value mean value of each dimension in the lion movement space range are respectively.
Preferably, the detecting the path points where the conflict exists on the basis of the initial feasible path includes:
with O ir Representing the optimal path of the trolley handling pallet i, i.e. the raster sequence through which the trolley handling pallet i passes in T ir And out T ir respectively representing the entry and exit of a vehicle into and out of the grid sequence O ir According to the total transport path O of the carriages r Expressed as:
Figure BDA0003898032780000079
cart ingress and egress grid sequence O r May be represented as:
Figure BDA00038980327800000710
get
Figure BDA00038980327800000711
Is a trolley k 1 One of the grids in the conveying path,
Figure BDA00038980327800000712
is a trolley k 2 Conveying pathIs (i ∈ N) ^ (k) 1 ,k 2 ∈K)∧(k 1 ≠k 2 ) Get it
Figure BDA00038980327800000713
Is a trolley k 1 Entry grid
Figure BDA00038980327800000714
The time of (a) is,
Figure BDA00038980327800000715
is a trolley k 2 Entry grid
Figure BDA00038980327800000716
The method for identifying the node conflict among the trolleys comprises the following steps:
if it is
Figure BDA00038980327800000717
And is provided with
Figure BDA00038980327800000718
Then the trolley k 1 And a trolley k 2 In that
Figure BDA00038980327800000719
Is composed of
Figure BDA00038980327800000720
Node collisions occur at the moment.
Preferably, the adjusting conflicting path points using AGV priority includes:
judging the priority of the trolleys with node conflict according to a preset rule;
according to the priority evaluation result of the trolley with node conflict, the trolley with the highest priority is allowed to enter the grid with conflict preferentially, and other trolleys are stopped for waiting and the subsequent time window is updated; in order to ensure that the trolley with the highest priority smoothly passes through the grid with conflict, after the trolley completely leaves, other trolleys can enter the grid with conflict, namely the time for the trolley with the low priority to perform one-time parking waiting is the time required for the trolley to normally run through one grid;
and in the conflict solving process, when the trolley executes the parking waiting once, the number of the shelf carried by the trolley at the moment is recorded, the time required by the trolley to carry the shortest broken line path of the next shelf and the service time required by carrying the shelf at the moment are updated, and the fitness function value is calculated again according to the objective function.
Preferably, the preset rule includes:
rule 1: will J c The trolleys in the sequence are sorted in descending order according to the transport completion time, and are sequentially given priority from high to low according to the sorting, J c Representing a trolley set with node conflict;
rule 2: if rule 1 is not satisfied, that is, if the transportation completion times of the plurality of vehicles are equal, J is calculated c The sum of the collision times of each trolley and other trolleys is given priority from high to low to the trolleys with the same transportation completion time in sequence according to the number of the collision times;
rule 3: if rule 1 and rule 2 are not satisfied, that is, if the sum of the transportation completion time and the number of collisions of the plurality of vehicles are equal, J is calculated c The number of the same grids in the transportation paths of each trolley and other trolleys is given priority from high to low to the trolleys with the equal transportation completion time and the equal collision times in sequence according to the number of the same grids;
rule 4: if J c If a plurality of trolleys do not meet the 3 rules, the priority order of the trolleys is randomly determined.
According to the AGV optimal scheduling method for the intelligent warehouse, provided by the invention, the objective of taking and placing tasks of all products is effectively achieved in a short time by applying the improved lion group algorithm and the AGV priority, the task balanced distribution of a multi-AGV picking system is realized, the running efficiency and rationality of the AGV picking system are improved, the transportation cost is reduced, and the enterprise benefit is improved.
Drawings
FIG. 1 is a block diagram of an AGV optimization scheduling method for an intelligent warehouse according to the present invention;
FIG. 2 is a flowchart illustrating an AGV optimization scheduling method for an intelligent warehouse according to the present invention;
FIG. 3 is a diagram of a grid model of the present invention;
FIG. 4 is a flowchart of a two-stage optimization method according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.
In one embodiment, an AGV optimal scheduling method for an intelligent warehouse is provided, as shown in fig. 1, in the method of this embodiment, time is divided into a plurality of intervals, a dynamic demand is a static demand, and an order demand and a replenishment demand in each interval are known. And performing AGV (trolley) path planning on the part of the demands, determining the shelf position and the walking path which need to be accessed by the trolley by adopting mathematical modeling and a path optimization algorithm based on the improved lion group algorithm, and generating an initial feasible path. And on the basis of the initial feasible path, adding a collision avoidance algorithm, actively detecting path points which are possibly collided, and carrying out passive adjustment to finally obtain an efficient and feasible global AGV scheduling scheme.
Specifically, as shown in fig. 2, the AGV optimal scheduling method for an intelligent warehouse according to this embodiment includes the following steps:
step 1, constructing an AGV dispatching model for the AGV picking system of the warehouse based on a grid modeling method.
The warehouse AGV picking system is described as follows: the system is provided with m trolleys and n shelves, the movement area of each trolley is marked as a two-dimensional plane O, a rectangular coordinate system is established for the two-dimensional plane O, the lower left corner is used as a coordinate origin O, an X axis is established transversely, a Y axis is established longitudinally, the width required by trolley walking is used as a unit, the plane is divided into grids, the grids are numbered in sequence, and the trolleys and the shelves in the warehouse respectively occupy one grid.
As shown in FIG. 3, the coordinates of the first grid in the lower left corner are defined as (1,1), each grid having a defined coordinate (x, y) in the coordinate system. Putting the warehouse into a rectangular coordinate system: numbering each shelf and obtaining corresponding coordinate information; numbering each trolley and obtaining corresponding coordinate information; and numbering each platform for parking the trolley and obtaining corresponding coordinate information, wherein grids occupied by the three types of entities cannot be listed in the range of the channel, and other grids are free grids and can be used for the trolley to pass through.
The following provisions are made for AGV operation:
(1) All AGVs in the system run at a constant speed.
(2) The situation that the power of the AGV is insufficient is not considered.
(3) All AGVs are homogeneous, regardless of the weight limitations of the transport rack.
(4) Each AGV carries a maximum of one rack at a time.
(5) The condition that the warehouse has a shortage and can not meet the order requirement is not considered.
(6) The AGV can only walk straight lines and does not walk diagonal lines of the grids.
Therefore, an AGV dispatching model is constructed by taking the minimization of the picking time of the AGV with the longest picking path as an objective, and the objective function is as follows:
Figure BDA0003898032780000101
where Z represents the pick time of the AGV with the longest pick path, N = {1,2,3, …, N } represents the set of all racks, K = {1,2,3, …, m } represents the set of all carts,
Figure BDA0003898032780000102
representing a binary decision variable, if the cart k transports the goods shelf j after transporting the goods shelf i, the value is 1, otherwise, the value is 0,
Figure BDA0003898032780000103
representing the time required for cart k to travel the shortest polygonal line path from shelf i to shelf j,
Figure BDA0003898032780000104
representing a binary decision variable, taking the value of 1 if the goods shelf i is transported by the trolley k, or else, taking the value of 0,S i Indicating the required service time for shelf i.
And the constraint conditions of the AGV scheduling model are as follows:
(1) Cart k starts from the initial stop:
Figure BDA0003898032780000105
in the formula (I), the compound is shown in the specification,
Figure BDA0003898032780000106
representing a binary decision variable, if the trolley k starts from an initial stop station 0 to a goods shelf j, taking the value as 1, otherwise, taking the value as 0;
(2) Car k finally returns to the initial stop:
Figure BDA0003898032780000107
in the formula (I), the compound is shown in the specification,
Figure BDA0003898032780000108
representing a binary decision variable, if the trolley k returns to the initial stop station 0 from the goods shelf i, the value is 1, otherwise, the value is 0;
(3) Each shelf is accessed only once:
Figure BDA0003898032780000109
Figure BDA00038980327800001010
(4) Line balance constraint, the inlet and outlet flow of each point is equal:
Figure BDA00038980327800001011
where a = {0,1,2, …, a } represents the set of all points, including all shelves and all carts, where 0 represents the initial stop point of the cart and a = n + m;
Figure BDA0003898032780000111
representing a binary decision variable, and if the trolley k enters the point h from the point q, taking the value of 1, otherwise, taking the value of 0;
Figure BDA0003898032780000112
representing a binary decision variable, if the trolley k starts from a point h to a point p, taking the value of 1, otherwise, taking the value of 0;
(5) If cart k goes from shelf j to shelf i, then shelf i is serviced by cart k:
Figure BDA0003898032780000113
(6) The time it takes for car k to move from shelf i to shelf j is equal to the fold line distance between shelves divided by the car k travel speed:
Figure BDA0003898032780000114
wherein (X) i ,Y i ) Is the position coordinate of the shelf i, (X) j ,Y j ) Is the position coordinate, V, of the goods shelf j k The running speed of the trolley k;
(7) The time required for cart k to travel from the initial stop to shelf j is equal to the polyline distance between the initial stop and shelf j divided by the travel speed of cart k:
Figure BDA0003898032780000115
in the formula (AGV _ X) k0 ,AGV_Y k0 ) Coordinates of the initial stop of cart k;
(8) The time required for cart k to return from shelf i to the initial stop point is equal to the polyline distance between shelf i and the initial stop point divided by the travel speed of cart k:
Figure BDA0003898032780000116
(9) The service time required by the goods shelf i is equal to the time required by the goods shelf to move from the storage position to the goods picking platform to walk the shortest broken line path, the time required by the goods shelf to pick goods manually on the goods picking platform is added, and the time required by the goods shelf to move back from the goods picking platform to the storage position to walk the shortest broken line path is added:
Figure BDA0003898032780000117
in the formula (I), the compound is shown in the specification,
Figure BDA0003898032780000118
indicating the time required to move the shelf i from the storage location to the picking platform along the shortest polygonal line path,
Figure BDA0003898032780000119
indicating the time of the goods shelf i to pick goods manually on the goods picking platform,
Figure BDA00038980327800001110
representing the time required to move the shelf i from the picking platform back to the storage position to take the shortest broken line path;
(10) The time for the trolley k to complete all transportation tasks and return to the initial stop is equal to the trolley k walking time plus the trolley service shelf time:
Figure BDA0003898032780000121
(11) The time for the shelf i to start moving is equal to the starting moving time of the last shelf served by the trolley k plus the service time of the last shelf plus the walking time between the two shelves:
Figure BDA0003898032780000122
in the formula, A i Indicates the time at which the shelf i starts to move, A j Indicates the time at which the shelf j starts to move, S j Indicating the required service time for the shelf j,
Figure BDA0003898032780000123
representing a binary decision variable, if the cart k transports the goods shelf i after transporting the goods shelf j, the value is 1, otherwise, the value is 0,
Figure BDA0003898032780000124
indicating the time required for cart k to take the shortest polyline path from shelf j to shelf i.
(12) And (3) eliminating the sub-loop:
Figure BDA0003898032780000125
Figure BDA0003898032780000126
in the formula (I), the compound is shown in the specification,
Figure BDA0003898032780000127
indicating the order of points visited in the path of cart k from point q to shelf i, i.e. the cart k visits shelf-related flow variables,
Figure BDA0003898032780000128
indicating the order of points visited by cart k in the path from point q to point q on shelf i,
Figure BDA0003898032780000129
representing the order of points visited in the path of cart k from point q to point p,
Figure BDA00038980327800001210
and (3) representing a binary decision variable, wherein if the trolley k enters a point q from a point p, the value is 1, and if not, the value is 0.
Besides the above constraint conditions, the method also comprises the following steps of constraining the variable values in the AGV scheduling model:
Figure BDA00038980327800001211
Figure BDA00038980327800001212
Figure BDA00038980327800001213
the two-stage optimization method provided by the embodiment for dispatching the warehouse AGVs is shown in fig. 4, and specifically includes the following steps 2 and 3, and the two-stage optimization method is implemented to effectively overcome the defects of the existing warehouse dispatching and to obtain the optimal dispatching and carrying path of each trolley.
And 2, solving an AGV scheduling model by utilizing a lion group algorithm to generate an initial feasible path.
And 2.1, initializing the lion group algorithm related parameters.
For the picking target shelf, firstly, all orders are collected to obtain a list of goods to be picked, and meanwhile, the maximum stock of each single storage position of various required goods is calculated. Then, the list splitting process of the commodities to be picked is carried out, so as to ensure that each demand can be met by at least one goods space. If the demand exceeds the maximum inventory of the single cargo space, the demand needs to be split until the single demand is less than or equal to the maximum inventory of the single cargo space. And performing random initialization processing on the lion group based on the working environment model.
And randomly generating the positions of L D-dimensional lions, wherein L is the population number of the lions, taking and placing positions of a goods shelf and a trolley are used as the positions of each lion, and the position codes adopt a matrix mode to obtain an initial coding sequence.
And 2.2, calculating a fitness function value of the initial coding sequence according to an objective function of the AGV scheduling model, and calculating a global optimal solution and an individual historical optimal solution of the lion group.
And 2.3, updating the individual positions of the male lion (namely the lion king), the female lion and the young lion in the lion group.
Assuming that in the D-dimension search space, L lions form a population, wherein the number of adult lions (including male lions and female lions) is M, then
Figure BDA0003898032780000131
Wherein M = L β, the comparative example parameter β is adjusted by chaotic mapping:
β＝2r 2 ·sin(π·r)
wherein r is a random number between [0,1 ].
The M adult lion has 1 lion king, the first adult lion is recorded as the lion king, and the position of the lion king is recorded as x l ＝(x l1 x l2 ... x lD ) Wherein l is more than or equal to 1 and less than or equal to M, and the updating position of the lion king is updated as follows:
Figure BDA0003898032780000132
in the formula (I), the compound is shown in the specification,
Figure BDA0003898032780000133
for the updated position of lion king, gamma is a random number generated according to normal distribution N (0,1), g t For a global optimal position at the t-th generation,
Figure BDA0003898032780000134
the historical optimal position of the ith individual in the t generation; the position of the lion king is updated by the global optimal position g t Determining and updating the weight
Figure BDA0003898032780000135
Is the difference between the lion Wang Lishi optimal position and the global optimal position.
Position record of the shared female lion M-1 and the v th female lion
Figure BDA0003898032780000136
Wherein v is more than or equal to 1 and less than or equal to M and v is not equal to l, female lion updates individual position
Figure BDA0003898032780000137
As follows:
Figure BDA0003898032780000138
in the formula (I), the compound is shown in the specification,
Figure BDA0003898032780000139
for the historical optimal position of the v-th individual in the t generation,
Figure BDA00038980327800001310
is the historical optimal position of another female lion randomly selected from the t-th female lion group, i.e. c ≠ v, alpha f The time search space is a nonlinear function which changes from large to small, and the corresponding biological characteristics are that the female lion firstly searches a wider range in the hunting process, and narrows a circle around the hunting when approaching the hunting object for hunting. When a female lion is hunting, the location update is determined by the cooperation of its current location and a random other female lion.
The number of the young lion in the lion group is L-M, and the position of the w-th young lion is recorded as
Figure BDA0003898032780000141
W is more than or equal to 1 and less than or equal to L-M, and the position of the young lion is updated
Figure BDA0003898032780000142
As shown in the following formula:
Figure BDA0003898032780000143
Figure BDA0003898032780000144
Figure BDA0003898032780000145
wherein q is based on uniform distribution of U [0,1]The random number is generated by the random number generator,
Figure BDA0003898032780000146
for the child lion to follow the historical best position of the parent lion in the t th generation,
Figure BDA0003898032780000147
is the historical optimal position of the w-th individual in the t generation, alpha' c The disturbance factor of the moving range of the young lion is improved into a nonlinear disturbance factor, T max Is the maximum iteration number of the algorithm, t is the current iteration number,
Figure BDA0003898032780000148
in order to use the reverse elite algorithm to reverse the position, i.e. after the young lion is expelled from the group
Figure BDA0003898032780000149
And
Figure BDA00038980327800001410
the minimum value mean value and the maximum value mean value of each dimension in the lion movement space range are respectively.
And 2.4, calculating the fitness function value of each individual, and updating the global optimal solution and the individual historical optimal solution.
And 2.5, randomly selecting a candidate solution set in the setting field of the current global optimal solution.
And 2.6, judging whether the current candidate solution exists in the tabu table, if so, jumping to the step 2.7, and otherwise, jumping out of the current iteration.
And 2.7, calculating a fitness function value of the candidate solution, comparing the fitness function value with the fitness function value of the current global optimal solution, if the fitness function value of the candidate solution is better than the fitness function value of the current global optimal solution, updating the current candidate solution into the current global optimal solution, and updating a tabu table.
And 2.8, if the maximum tabu search times are reached, outputting the current global optimal solution as an initial feasible path, and otherwise, returning to the step 2.5.
And 2.9, if the maximum iteration times are reached, outputting the global optimal solution as an initial feasible path, otherwise, returning to the step 2.2.
And 3, detecting path points with conflicts on the basis of the initial feasible path, and then adjusting the path points with conflicts by using the AGV priority to finally obtain an efficient and feasible global AGV scheduling scheme.
Step 3.1, in the obtained path planning results of all the trolleys, calculating the time window when the trolleys pass through the path, judging whether conflicts occur among the trolleys, and providing a method for determining the priority of the conflict AGV, modifying the time window of the trolleys with lower priority to solve the conflicts among the AGV, wherein the specific steps of executing the collision avoidance algorithm on the basis of the initial feasible path are as follows:
with O ir Representing the optimal path of the trolley handling pallet i, i.e. the raster sequence through which the trolley handling pallet i passes in T ir And out T ir respectively representing the entry and exit of the car into and out of the grid sequence O ir According to the transport path of each shelf, the total transport path O of the carriages r Expressed as:
Figure BDA0003898032780000151
cart ingress and egress grid sequence O r May be represented as:
Figure BDA0003898032780000152
get
Figure BDA0003898032780000153
Is a trolley k 1 One of the grids in the conveying path,
Figure BDA0003898032780000154
is a trolley k 2 A grid in the conveying path, (i ∈ N) ^ (k) 1 ,k 2 ∈K)∧(k 1 ≠k 2 ) Get it
Figure BDA0003898032780000155
Is a trolley k 1 Entry grid
Figure BDA0003898032780000156
The time of (a) is,
Figure BDA0003898032780000157
is a trolley k 2 Entry grid
Figure BDA0003898032780000158
The method for identifying the node conflict among the trolleys comprises the following steps:
if it is
Figure BDA0003898032780000159
And is
Figure BDA00038980327800001510
Then the trolley k 1 And a trolley k 2 In that
Figure BDA00038980327800001511
Is composed of
Figure BDA00038980327800001512
Node collisions occur at the moment.
And 3.2, judging the priority of the AGV with the node conflict according to a preset rule.
The preset rules in this embodiment include:
rule 1: will J c The trolleys in the sequence are sorted in descending order according to the transport completion time, and are sequentially given priority from high to low according to the sorting, J c Representing a trolley set with node conflict;
rule 2: if rule 1 is not satisfied, that is, if the transportation completion times of the plurality of vehicles are equal, J is calculated c The sum of the collision times of each trolley and other trolleys is given priority from high to low to the trolleys with the same transportation completion time in sequence according to the number of the collision times;
rule 3: if rule 1 and rule 2 are not satisfied, that is, if the sum of the transportation completion time and the number of collisions of the plurality of vehicles are equal, J is calculated c The number of the same grids in the conveying paths of each trolley and other trolleys is given priority from high to low to the trolleys with equal conveying completion time and conflict times in sequence from most to least according to the number of the same grids;
rule 4: if J c If a plurality of trolleys do not meet the 3 rules, the priority order of the trolleys is randomly determined.
3.3, according to the priority evaluation result of the trolley with the node conflict, enabling the trolley with the highest priority to enter a grid with the conflict, and enabling other trolleys to stop for waiting and updating a subsequent time window; in order to ensure that the trolley with the highest priority smoothly passes through the grid with conflict, after the trolley completely leaves, other trolleys can enter the grid with conflict, namely the time for the trolley with the low priority to perform one-time parking waiting is the time required for the trolley to normally run through one grid;
and recording the number of the shelf carried by the trolley at the moment and updating the time required by the shortest broken line path from the trolley to the next shelf and the service time required by the trolley carrying the shelf at the moment when the trolley carries out parking waiting once in the conflict resolving process, and calculating the fitness function value according to the objective function again.
The technical features of the embodiments described above may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments described above are not described, but should be considered as being within the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above examples are merely illustrative of several embodiments of the present invention, and the description thereof is more specific and detailed, but not to be construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present invention should be subject to the appended claims.

Claims (8)

1. An AGV optimal scheduling method of an intelligent warehouse is characterized by comprising the following steps:
step 1, constructing an AGV dispatching model for an AGV picking system of a warehouse based on a grid modeling method;
step 2, solving an AGV scheduling model by utilizing a lion group algorithm to generate an initial feasible path;
step 3, detecting path points with conflicts on the basis of the initial feasible path, then adjusting the path points with conflicts by using the AGV priority, and finally obtaining an efficient and feasible global AGV dispatching scheme;
wherein, the adult lion in the lion subgroup algorithm is determined as follows:
assuming that L lions form a population in the D-dimensional search space, where the number of adult lions is M, then
Figure FDA0003898032770000011
Wherein M = L β, the comparative example parameter β is adjusted by chaotic mapping:
β＝2r 2 ·sin(π·r)
wherein r is a random number between [0,1 ].
2. The AGV optimized scheduling method of claim 1, wherein the building of the AGV scheduling model for the AGV picking system of the warehouse based on the grid building method comprises:
the warehouse AGV picking system is described as follows: the system is provided with m trolleys and n shelves, the movement area of each trolley is marked as a two-dimensional plane O, a rectangular coordinate system is established for the two-dimensional plane O, the lower left corner is taken as a coordinate origin O, an X axis is established transversely, a Y axis is established longitudinally, the width required by trolley walking is taken as a unit, the plane is divided into grids, the grids are numbered in sequence, and the trolleys and the shelves in the warehouse respectively occupy one grid;
putting the warehouse into a rectangular coordinate system: numbering each shelf and obtaining corresponding coordinate information; numbering each trolley and obtaining corresponding coordinate information; numbering each platform for parking the trolley and obtaining corresponding coordinate information, wherein grids occupied by the three types of entities cannot be listed in the range of the channel, and other grids are free grids and can be used for the trolley to pass through;
therefore, an AGV dispatching model is constructed by taking the minimization of the picking time of the trolley with the longest picking path as a target, and the target function is as follows:
Figure FDA0003898032770000012
where Z represents the pick time for the AGV with the longest pick path, N = {1,2,3, …, N } represents the set of all shelves, K = {1,2,3, …, m } represents the set of all carts,
Figure FDA0003898032770000013
representing a binary decision variable, if the cart k transports the goods shelf j after transporting the goods shelf i, the value is 1, otherwise, the value is 0,
Figure FDA0003898032770000014
representing the time required for cart k to travel the shortest polygonal line path from shelf i to shelf j,
Figure FDA0003898032770000021
representing a binary decision variable, taking the value of 1 if the goods shelf i is transported by the trolley k, otherwise, taking the value of 0,S i Indicating the required service time for shelf i.
3. The AGV optimal scheduling method according to claim 2, wherein the constraints of said AGV scheduling model are as follows:
(1) Cart k starts from the initial stop:
Figure FDA0003898032770000022
in the formula (I), the compound is shown in the specification,
Figure FDA0003898032770000023
representing a binary decision variable, if the trolley k starts from an initial stop station 0 to a goods shelf j, taking the value as 1, otherwise, taking the value as 0;
(2) Car k finally returns to the initial stop:
Figure FDA0003898032770000024
in the formula (I), the compound is shown in the specification,
Figure FDA0003898032770000025
representing a binary decision variable, if the trolley k returns to the initial stop station 0 from the goods shelf i, the value is 1, otherwise, the value is 0;
(3) Each shelf is accessed only once:
Figure FDA0003898032770000026
Figure FDA0003898032770000027
(4) Line balance constraint, the inlet and outlet flow of each point is equal:
Figure FDA0003898032770000028
where a = {0,1,2, …, a } represents the set of all points, including all shelves and all carts, where 0 represents the initial stop point of the cart and a = n + m;
Figure FDA0003898032770000029
representing a binary decision variable, and if the trolley k enters a point h from a point q, taking the value of the decision variable as 1, otherwise, taking the value of the decision variable as 0;
Figure FDA00038980327700000210
representing a binary decision variable, if the trolley k starts from a point h to a point p, taking the value of 1, otherwise, taking the value of 0;
(5) If cart k goes from shelf j to shelf i, then shelf i is serviced by cart k:
Figure FDA00038980327700000211
(6) The time it takes for car k to move from shelf i to shelf j is equal to the fold line distance between shelves divided by the car k travel speed:
Figure FDA0003898032770000031
in the formula (X) i ,Y i ) Is the position coordinate of the shelf i, (X) j ,Y j ) Is the position coordinate, V, of the goods shelf j k The running speed of the trolley k;
(7) The time required for cart k to travel from the initial stop to shelf j is equal to the polyline distance between the initial stop and shelf j divided by the travel speed of cart k:
Figure FDA0003898032770000032
in the formula (AGV _ X) k0 ,AGV_Y k0 ) Coordinates of the initial stop of cart k;
(8) The time required for cart k to return from shelf i to the initial stop point is equal to the polyline distance between shelf i and the initial stop point divided by the travel speed of cart k:
Figure FDA0003898032770000033
(9) The service time required by the goods shelf i is equal to the time required by the goods shelf to move from the storage position to the goods picking platform to walk the shortest broken line path, the time required by the goods shelf to manually pick goods on the goods picking platform is added, and the time required by the goods shelf to move back from the goods picking platform to the storage position to walk the shortest broken line path is added:
Figure FDA0003898032770000034
in the formula (I), the compound is shown in the specification,
Figure FDA0003898032770000035
indicating the time required to move the shelf i from the storage location to the picking platform along the shortest polygonal line path,
Figure FDA0003898032770000036
indicating the time of the goods shelf i to pick goods manually on the goods picking platform,
Figure FDA0003898032770000037
representing the time required to move the shelf i from the picking platform back to the storage position to take the shortest broken line path;
(10) The time for the trolley k to complete all transportation tasks and return to the initial stop is equal to the trolley k walking time plus the trolley service shelf time:
Figure FDA0003898032770000038
(11) The time for the shelf i to start moving is equal to the starting moving time of the last shelf served by the trolley k plus the service time of the last shelf, plus the walking time between the two shelves:
Figure FDA0003898032770000039
in the formula, A i Indicates the time at which the shelf i starts to move, A j Indicates the time at which the shelf j starts to move, S j Indicating the required service time for the shelf j,
Figure FDA00038980327700000310
representing a binary decision variable, if the trolley k transports the goods shelf i after transporting the goods shelf j, the value is 1, otherwise the value is 0,
Figure FDA0003898032770000041
the time required for the trolley k to move from the shelf j to the shelf i to take the shortest broken line path is represented;
(12) Cancellation of the sub-loop:
Figure FDA0003898032770000042
Figure FDA0003898032770000043
in the formula (I), the compound is shown in the specification,
Figure FDA0003898032770000044
representing the order of points visited in the path of cart k from point q to shelf i,
Figure FDA0003898032770000045
indicating the order of points visited by cart k in the path from point q to point q on shelf i,
Figure FDA0003898032770000046
representing the order of points visited in the path of cart k from point q to point p,
Figure FDA0003898032770000047
and (3) representing a binary decision variable, wherein if the trolley k enters a point q from a point p, the value is 1, and if not, the value is 0.
4. The AGV optimal scheduling method according to claim 1, wherein the generating an initial feasible path by solving the AGV scheduling model using the lion group algorithm includes:
step 2.1, initializing lion group algorithm related parameters, including: randomly generating the positions of L lions, taking a goods shelf pick-and-place position and a trolley as the position of each lion, and obtaining an initial coding sequence by adopting a matrix mode for position coding;
2.2, calculating a fitness function value of the initial coding sequence according to an objective function of the AGV scheduling model, and calculating a global optimal solution and an individual historical optimal solution of the lion group;
step 2.3, updating the individual positions of the male lion, the female lion and the young lion in the lion group;
step 2.4, calculating a fitness function value of each individual, and updating a global optimal solution and an individual historical optimal solution;
step 2.5, randomly selecting a candidate solution set in the setting field of the current global optimal solution;
step 2.6, judging whether the current candidate solution exists in a taboo table, if so, jumping to step 2.7, otherwise, jumping out of the current iteration;
step 2.7, calculating a fitness function value of the candidate solution, comparing the fitness function value with the fitness function value of the current global optimal solution, if the fitness function value of the candidate solution is better than the fitness function value of the current global optimal solution, updating the current candidate solution into the current global optimal solution, and updating a tabu table;
step 2.8, if the maximum tabu search times is reached, outputting the current global optimal solution as an initial feasible path, otherwise, returning to the step 2.5;
and 2.9, if the maximum iteration times are reached, outputting the global optimal solution as an initial feasible path, otherwise, returning to the step 2.2.
5. The AGV optimal scheduling method according to claim 4, wherein said updating the individual positions of the male lion, the female lion and the young lion of the lion group comprises:
the M adult lion has 1 lion king, the first adult lion is recorded as the lion king, and the position of the lion king is recorded as x l ＝(x l1 x l2 ... x lD ) Wherein l is more than or equal to 1 and less than or equal to M, and the updating position of the lion king is updated as follows:
Figure FDA0003898032770000051
in the formula (I), the compound is shown in the specification,
Figure FDA0003898032770000052
for the updated position of lion king, gamma is a random number generated according to normal distribution N (0,1), g t For a global optimal position at the t-th generation,
Figure FDA0003898032770000053
the historical optimal position of the ith individual in the t generation;
position record of the shared female lion M-1 and the v th female lion
Figure FDA0003898032770000054
Wherein v is more than or equal to 1 and less than or equal to M and v is not equal to l, female lion updates individual position
Figure FDA0003898032770000055
As follows:
Figure FDA0003898032770000056
in the formula (I), the compound is shown in the specification,
Figure FDA0003898032770000057
for the historical optimal position of the v-th individual in the t generation,
Figure FDA0003898032770000058
is the historical optimal position of another female lion randomly selected from the t-th female lion group, i.e. c ≠ v, alpha f A nonlinear function which changes from large to small in a search space;
the small lion in the lion group has L-M, and the position of the w-th small lion is recorded as
Figure FDA0003898032770000059
W is more than or equal to 1 and less than or equal to L-M, and the position of the young lion is updated
Figure FDA00038980327700000510
As shown in the following formula:
Figure FDA00038980327700000511
Figure FDA00038980327700000512
Figure FDA00038980327700000513
wherein q is based on uniform distribution of U [0,1]The random number is generated by the random number generator,
Figure FDA00038980327700000514
for the child lion to follow the historical best position of the parent lion in the t th generation,
Figure FDA00038980327700000515
is the historical optimal position of the w-th individual in the t generation, alpha' c Disturbance factor of moving range of young lion, T max Is the maximum iteration number of the algorithm, t is the current iteration number,
Figure FDA00038980327700000516
in order to use the reverse position after the reverse elite algorithm, namely the position after the young lion is driven out of the group
Figure FDA0003898032770000061
And
Figure FDA0003898032770000062
the minimum mean value and the maximum mean value of each dimension in the lion movement space range are respectively.
6. The AGV optimized scheduling method for intelligent warehouses according to claim 2, wherein the detecting conflicting path points based on the initial feasible path includes:
with O ir Representing the optimal path of the trolley handling pallet i, i.e. the raster sequence through which the trolley handling pallet i passes in T ir And out T ir respectively representing the entry and exit of the car into and out of the grid sequence O ir According to the transport path of each shelf, the total transport path O of the carriages r Expressed as: o is r ＝{O 1r ,O 2r ,...,O ir ,...,O nr },
Figure FDA0003898032770000063
Cart ingress and egress grid sequence O r May be represented as: in T r ＝{ in T 1r , in T 2r ,..., in T kr ,..., in T mr },
Figure FDA0003898032770000064
out T r ＝{ out T 1r , out T 2r ,..., out T kr ,..., out T mr },
Figure FDA0003898032770000065
get the
Figure FDA0003898032770000066
Is a trolley k 1 One of the grids in the conveying path,
Figure FDA0003898032770000067
is a trolley k 2 A grid in the conveying path, (i ∈ N) ^ (k) 1 ,k 2 ∈K)∧(k 1 ≠k 2 ) Get it
Figure FDA0003898032770000068
Is a trolley k 1 Entry grid
Figure FDA0003898032770000069
The time of (a) is,
Figure FDA00038980327700000610
is a trolley k 2 Entry grid
Figure FDA00038980327700000611
The method for identifying the node conflict among the trolleys comprises the following steps:
if it is
Figure FDA00038980327700000612
And is
Figure FDA00038980327700000613
Then the trolley k 1 And a trolley k 2 In that
Figure FDA00038980327700000614
Is composed of
Figure FDA00038980327700000615
Node collisions occur at the moment.
7. The AGV optimized scheduling method for intelligent warehouses according to claim 1, wherein the adjusting conflicting path points using AGV priorities includes:
judging the priority of the trolley with node conflict according to a preset rule;
according to the priority evaluation result of the trolley with node conflict, the trolley with the highest priority is allowed to enter the grid with conflict preferentially, and other trolleys are stopped for waiting and the subsequent time window is updated; in order to ensure that the trolley with the highest priority smoothly passes through the grid with conflict, after the trolley completely leaves, other trolleys can enter the grid with conflict, namely the time for the trolley with the low priority to perform one-time parking waiting is the time required for the trolley to normally run through one grid;
and recording the number of the shelf carried by the trolley at the moment and updating the time required by the shortest broken line path from the trolley to the next shelf and the service time required by the trolley carrying the shelf at the moment when the trolley carries out parking waiting once in the conflict resolving process, and calculating the fitness function value according to the objective function again.
8. The AGV optimized scheduling method according to claim 7, wherein said preset rules include:
rule 1: will J c The trolleys in the sequence are sorted in descending order according to the transport completion time, and are sequentially given priority from high to low according to the sorting, J c Representing a trolley set with node conflict;
rule 2: if rule 1 is not satisfied, that is, if the transportation completion times of the plurality of vehicles are equal, J is calculated c The sum of the number of times of conflict between each dolly and other dolliesGiving high-to-low priority to trolleys with equal transportation completion time according to the sum of the collision times from small to large;
rule 3: if rule 1 and rule 2 are not satisfied, that is, if the sum of the transportation completion time and the number of collisions of the plurality of vehicles are equal, J is calculated c The number of the same grids in the transportation paths of each trolley and other trolleys is given priority from high to low to the trolleys with the equal transportation completion time and the equal collision times in sequence according to the number of the same grids;
rule 4: if J c If a plurality of trolleys do not meet the 3 rules, the priority order of the trolleys is randomly determined.
CN202211291544.9A 2022-10-19 2022-10-19 AGV optimization scheduling method for intelligent warehouse Withdrawn CN115587679A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211291544.9A CN115587679A (en) 2022-10-19 2022-10-19 AGV optimization scheduling method for intelligent warehouse

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211291544.9A CN115587679A (en) 2022-10-19 2022-10-19 AGV optimization scheduling method for intelligent warehouse

Publications (1)

Publication Number Publication Date
CN115587679A true CN115587679A (en) 2023-01-10

Family

ID=84780191

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211291544.9A Withdrawn CN115587679A (en) 2022-10-19 2022-10-19 AGV optimization scheduling method for intelligent warehouse

Country Status (1)

Country Link
CN (1) CN115587679A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116187594A (en) * 2023-04-27 2023-05-30 北京玻色量子科技有限公司 Multi-target multi-task path scheduling cost optimization method, device, medium and equipment
CN116187595A (en) * 2023-04-27 2023-05-30 北京玻色量子科技有限公司 Multi-target multi-task path scheduling efficiency optimization method, device, medium and equipment
CN117875189A (en) * 2024-03-06 2024-04-12 安徽建筑大学 Three-dimensional warehouse space layout method based on GA optimization GRO
CN117875189B (en) * 2024-03-06 2024-05-14 安徽建筑大学 Three-dimensional warehouse space layout method based on GA optimization GRO

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116187594A (en) * 2023-04-27 2023-05-30 北京玻色量子科技有限公司 Multi-target multi-task path scheduling cost optimization method, device, medium and equipment
CN116187595A (en) * 2023-04-27 2023-05-30 北京玻色量子科技有限公司 Multi-target multi-task path scheduling efficiency optimization method, device, medium and equipment
CN116187594B (en) * 2023-04-27 2023-09-08 北京玻色量子科技有限公司 Multi-target multi-task path scheduling cost optimization method, device, medium and equipment
CN117875189A (en) * 2024-03-06 2024-04-12 安徽建筑大学 Three-dimensional warehouse space layout method based on GA optimization GRO
CN117875189B (en) * 2024-03-06 2024-05-14 安徽建筑大学 Three-dimensional warehouse space layout method based on GA optimization GRO

Similar Documents

Publication Publication Date Title
Lee et al. Robotics in order picking: evaluating warehouse layouts for pick, place, and transport vehicle routing systems
CN115587679A (en) AGV optimization scheduling method for intelligent warehouse
Fragapane et al. Planning and control of autonomous mobile robots for intralogistics: Literature review and research agenda
CN112036773B (en) AGV trolley task allocation method, equipment, storage medium and device
CN205230118U (en) Intelligence warehouse system based on multirobot
CN107628404B (en) Order-to-person-based sorting system and method for logistics storage center
JP2022533784A (en) Warehousing task processing method and apparatus, warehousing system and storage medium
US20190177086A1 (en) A picking system having a transport robot for moving underneath individualshelves and transporting vehicle
Wang et al. Research on logistics distribution vehicle scheduling based on heuristic genetic algorithm
CN104346658B (en) System dynamic dispatching method is accessed based on the automatic vehicle for improving banker's algorithm
CN109969989B (en) Driving strategy determination method, intelligent forklift and storage medium
CN108897316B (en) Cluster warehousing robot system control method based on pheromone navigation
Li et al. A simulation study on the robotic mobile fulfillment system in high-density storage warehouses
CN113387098A (en) Cargo conveying method, cargo conveying device, electronic equipment and storage medium
CN110597263A (en) Automatic meal delivery path planning method for unmanned restaurant
CN114964253A (en) Path planning method, electronic device, storage medium and program product
Shi et al. Task allocation and path planning of many robots with motion uncertainty in a warehouse environment
CN117234214A (en) Automatic shuttle for stacking industrial goods
Le et al. Integrating both routing and scheduling into motion planner for multivehicle system
US20230211508A1 (en) Loading and unloading by an autonomous mobile robot
CN114757591B (en) Multi-vehicle type collaborative sorting scheduling method based on behavior dependency graph
Zhou et al. AGV Path planning based on improved adaptive genetic algorithm
US20220162001A1 (en) Predicting a path of material handling equipment and determining an obstacle-free path
CN115393003A (en) Similarity-based multi-sorting-table-oriented order batching, sorting and sorting method and system
Yuan et al. A task scheduling problem in mobile robot fulfillment systems

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
WW01 Invention patent application withdrawn after publication

Application publication date: 20230110

WW01 Invention patent application withdrawn after publication