CN114358233A - Multi-AGV path planning problem optimization method and system based on double-hybrid particle swarm - Google Patents
Multi-AGV path planning problem optimization method and system based on double-hybrid particle swarm Download PDFInfo
- Publication number
- CN114358233A CN114358233A CN202111454010.9A CN202111454010A CN114358233A CN 114358233 A CN114358233 A CN 114358233A CN 202111454010 A CN202111454010 A CN 202111454010A CN 114358233 A CN114358233 A CN 114358233A
- Authority
- CN
- China
- Prior art keywords
- population
- point
- individual
- condition
- individuals
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 69
- 239000002245 particle Substances 0.000 title claims abstract description 26
- 238000013439 planning Methods 0.000 title claims abstract description 24
- 238000005457 optimization Methods 0.000 title claims description 9
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 42
- 230000035772 mutation Effects 0.000 claims abstract description 12
- 230000000295 complement effect Effects 0.000 claims description 10
- 230000009977 dual effect Effects 0.000 claims description 8
- 239000012634 fragment Substances 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 2
- 239000000463 material Substances 0.000 abstract description 5
- 230000002068 genetic effect Effects 0.000 abstract description 4
- 238000002156 mixing Methods 0.000 abstract description 3
- 238000004519 manufacturing process Methods 0.000 description 10
- 230000032258 transport Effects 0.000 description 8
- 238000011160 research Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 229910000831 Steel Inorganic materials 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 4
- 239000010959 steel Substances 0.000 description 4
- 239000002994 raw material Substances 0.000 description 3
- 241000531116 Blitum bonus-henricus Species 0.000 description 2
- 235000008645 Chenopodium bonus henricus Nutrition 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000003032 molecular docking Methods 0.000 description 2
- 230000032683 aging Effects 0.000 description 1
- 238000003490 calendering Methods 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000010152 pollination Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000010074 rubber mixing Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000012384 transportation and delivery Methods 0.000 description 1
- 238000004073 vulcanization Methods 0.000 description 1
Images
Abstract
The invention provides a method and a system for optimizing multiple AGV path planning problems based on double hybrid particle swarm, and belongs to the technical field of dispatching AGV of warehouse material handling. The double-hybrid particle swarm algorithm is based on the mixing of the particle swarm algorithm and the A-star algorithm, and the crossing and the variation of the genetic algorithm are introduced. The method comprises the following steps: encoding a sequence of cargo sites of an automated sorting warehouse to produce an initial population; generating an optimal path by adopting a path generation algorithm, namely an A algorithm; calculating the number of parking waiting times of each individual according to the optimal path and each individual in the population; calculating the fitness value of each individual of the population; performing crossover and mutation operations on the population; judging whether the current end condition is met; under the condition that the termination condition is not met currently, returning to the step of calculating the fitness value of each individual of the population until the termination condition is met currently; and outputting the optimized scheme under the condition of judging that the termination condition is currently met. The method and the system can improve the efficiency of warehouse scheduling.
Description
Technical Field
The invention relates to the technical field of AGV (automatic guided vehicle) scheduling of warehouse material handling, in particular to a method and a system for optimizing multiple AGV path planning problems based on double hybrid particle swarm.
Background
The tire manufacturing industry is the basic industry of national economy, and the development of the tire directly influences the development of each department of the national economy and also influences the enhancement of the national civilization and national defense strength. Whether the country can go on a prosperous and prosperous road is a key circle in the tire manufacturing industry.
The mode of semi-steel and all steel tire mill's current production prepareeing material and transport mainly is with fork truck, conveyer belt as the owner, and conveyer belt occupation space is big in the mill, and the flexibility is poor, carries the overall arrangement that can optimize tire production line workshop through introducing AGV, promotes the flexibility to promote the work efficiency in mill's workshop. The semi-steel and all-steel tire factory processes comprise rubber mixing, calendering, component forming, vulcanization and the like, the production flow is complex, the temperature of the production environment is high, the labor intensity is high, and the AGV can perfectly complete the tasks in the working environment. The environmental and safety problems caused by manpower can be eliminated.
The problem of AGV scheduling is an important research problem in the relevant problems of AGV, and is a hotspot problem in the research of scholars at home and abroad at present. The reasonable vehicle dispatching scheme can save the transportation cost and time for the tire factory, improve the efficiency of logistics service, and improve the competitiveness for the tire factory, so that the method has important significance for the research of the problem.
The AGV scheduling problems include a single AGV scheduling problem and a multiple AGV scheduling problem. For the single AGV scheduling problem, the high-efficiency real-time scheduling rule is excavated based on the gene expression programming of Chenglin and the like; CAUMOND A and the like research the scheduling problem of an automatic guided vehicle in a flexible manufacturing system, establish a mixed integer linear programming model, provide an optimal solution and research the influence of management rules and some classical assumptions. For the multiple AGV scheduling problem, ZOU et al studied a new automated guided vehicle scheduling problem relating to pick-up and delivery from goods handling in matrix manufacturing plants with multiple varieties and small volume production, which established a multi-objective mixed integer programming model and developed an efficient multi-objective evolutionary algorithm to solve the problem; ZOU, which aims to determine a solution to minimize transportation costs, by first making a mixed integer linear programming model, and proposing a discrete artificial bee colony algorithm and some novel and advanced techniques to solve the problem; the aging and the like provide an automatic guided vehicle scheduling rule based on a soft time window; an improved differential evolution algorithm, a mixed taboo bat algorithm, an optimized fuzzy decision algorithm, an improved flower pollination algorithm, an improved mixed genetic algorithm and a particle swarm algorithm are respectively proposed for Monana, Weiyongyai, Luoxin, Liubiehui, Yueyou Xiao and the like; CHAWLA V K and the like research the performance of a dynamic job selection scheduling rule for simultaneously scheduling multi-load automatic guided vehicles in two flexible manufacturing systems with different scales.
In an automatic sorting warehouse of a tire factory, not only the scheduling of the AGVs but also the path planning of a plurality of AGVs are considered, and especially the problem of collision prevention is considered. In AGV path planning, the Yang Zhou and the like provide a global dynamic path planning algorithm combining an improved ant colony method and a dynamic window method; li Kunpeng et al proposed a two-stage optimization algorithm; in order to solve the anti-collision problem, a loop deadlock search method is provided on Xiaohaining and the like; zhang Danlu et al propose a dynamic weighting map based on the improved A-x algorithm under traffic rules and reservation tables.
The research shows that the researched model and algorithm are only in consideration of solving the problem of multi-AGV path planning of the automatic sorting warehouse for storing the tire raw materials and the components, so that the problem that the algorithm is suitable for the automatic sorting warehouse for storing the tire raw materials and the components is difficult to overcome in actual implementation, the transportation time is reduced, and the efficiency of the automatic sorting warehouse is improved; the invention provides a double-hybrid particle swarm algorithm, which is based on the mixing of the particle swarm algorithm and the A-x algorithm, introduces the intersection and variation of the genetic algorithm, can be well suitable for an automatic sorting warehouse for storing tire raw materials and parts, and improves the dispatching efficiency of the material handling AGV.
Disclosure of Invention
The embodiment of the invention aims to provide a method and a system for optimizing multiple AGV path planning problems based on double-hybrid particle swarm, and the method and the system can improve the dispatching efficiency of AGV for warehouse material handling.
In order to achieve the above object, an embodiment of the present invention provides a method for optimizing multiple AGV path planning problems based on dual hybrid particle swarm, including:
encoding a sequence of goods locations of an automated sorting warehouse to produce an initial population;
generating an optimal path by adopting a preset path generation algorithm, namely an A algorithm;
calculating the number of times of parking waiting of each individual according to the optimal path and each individual in the population;
calculating a fitness value for each individual of the population;
performing individual optimal crossover operations on the population;
performing population optimal cross operation on the population after the individual optimal cross operation;
performing variation operation on the population after the population optimal cross operation;
judging whether the current end condition is met;
under the condition that the termination condition is not met currently, returning to the step of calculating the fitness value of each individual of the population until the termination condition is met currently;
and outputting the optimized scheme under the condition of judging that the termination condition is currently met.
Optionally, encoding the positions and paths of the automated sorting warehouse to produce the initial population comprises:
and the number 0 is adopted to represent the stopping area of the AGV trolley, and an integer is adopted as a goods position to be carried.
Optionally, the path generation algorithm includes:
determining all points to be driven, wherein the points to be driven comprise a starting point, a terminal point or a cargo space point of an AGV trolley;
initializing a path set and a set to be traveled, wherein the path set comprises the starting point, and the set to be traveled comprises all the points to be traveled except the starting point in the set of all the freight sites;
traversing and calculating the f value of each point to be driven in the set to be driven;
selecting the point to be driven with the minimum f value as a current selection point;
adding the current selection point into the path set;
judging whether the current selected point is the end point;
under the condition that the current selection point is judged not to be the terminal point, judging whether the set to be driven contains at least one point to be driven adjacent to the current selection point;
adding the point to be traveled adjacent to the current selection point into the path set under the condition that the set to be traveled comprises at least one point to be traveled adjacent to the current selection point;
taking a point to be driven adjacent to the current selection point as a child node of the current selection point, and calculating f values of the child nodes according to formulas (1) to (3),
f(j)＝g(j)+h(j) (1)
g(j)＝g(i)+|xj-xi|+|yj-yi| (2)
wherein f (j) is the f value of the child node, g (j) is the cost value of the child node from the starting point, h (j) is the predicted cost value of the child node to the end point, g (i) is the cost value of the parent node from the starting point, the parent node is the current selection point, i is the number of the parent node, u (j) is the cost value of the parent node from the starting point, and0is the number of the end point and is,j is the number of the child node, x is the abscissa of the point to be traveled, y is the ordinate of the point to be traveled, and D is the cost value of each travel unit distance of the AGV trolley;
under the condition that the set to be driven does not contain the point to be driven adjacent to the current selection point, judging whether the predicted cost values of the points to be driven adjacent to the current selection point are all smaller than the predicted cost value in the previous iteration;
under the condition that the estimated cost value of the point to be driven adjacent to the current selected point is judged to be smaller than the estimated cost value in the previous iteration, the current selected point is taken as a father node of the point to be driven adjacent to the current selected point, and the f value of the father node, the cost value of the father node from the starting point and the estimated cost value of the father node to the terminal point are updated according to formulas (1) to (3);
returning to the step of selecting the freight site with the minimum f value until the current selection point is judged to be the terminal point;
and under the condition that the current selection point is judged to be the end point, outputting the path set as the optimal path.
Optionally, the calculating the fitness value of each individual comprises:
calculating the fitness value according to formula (4) and formula (5),
Fitness(i)＝max(t1,t2,…,tk,…,tm) (4)
wherein, t1,t2,…,tk,…,tmTotal transport time, t, for AGV carts numbered 1, 2, …, k, …, m, respectivelykIs a plurality of TkAnd, TkAGV with designation number kkThe time required for carrying the goods once,
wherein d isi1The first Manhattan distance of the task i, namely the distance between the starting point of the task i and the shelfManhattan distance of di2The second Manhattan distance representing task i, i.e. the Manhattan distance from the shelf to the picking station, di3A third segment Manhattan distance representing task i, i.e., the Manhattan distance between the pick table to the destination, and di1＝|xs-xi|+|ys-yi|，di2＝|xi-1|+|yi-yp|，di3＝|1-xe|+|yp-ye|，xsIs the abscissa of the start of task i, ysIs the ordinate, x, of the start of task iiAbscissa of shelf for task i, yiOrdinate, y, of the shelf for task ipOrdinate, x, of picking station for task ieIs the abscissa, y, of the end point of task ieIs the ordinate, x, of the end point of task iikFor indicating whether task i is carried by AGVkIndicating variable of transport, xikThe value of 1 indicates that the task i is carried out by an AGVkCarrying, otherwise, taking the value as 0; t is twIndicating the AGV cart one-time parking waiting time.
Optionally, the individual optimal crossover operation comprises:
randomly selecting an unselected individual from the population and an individual with the minimum fitness in the current population to form a parent individual;
randomly generating two natural numbers r1And r2；
Two individuals exchanging parents are in the natural number r1And r2A fragment in between;
judging whether the two exchanged individuals have repeated cargo space point numbers or not;
under the condition that the two exchanged individuals have repeated goods location point numbers, taking a complementary set of the exchanged segments to rearrange the complementary set of the exchanged segments to the non-exchanged segments randomly;
judging whether unselected individuals exist in the population;
under the condition that the unselected individuals exist in the population, the step of randomly selecting one unselected individual from the population and forming a parent individual with the smallest fitness in the current population is returned again until the unselected individual does not exist in the population;
and under the condition that the unselected individuals do not exist in the population, outputting the population with the optimal cross operation of the individuals.
Optionally, the population-optimal crossover operation includes:
randomly selecting an unselected individual from the population and an individual with the minimum fitness in the historically generated population to form a parent individual;
randomly generating two natural numbers r1And r2；
Two individuals exchanging parents are in the natural number r1And r2A fragment in between;
judging whether the two exchanged individuals have repeated cargo space point numbers or not;
under the condition that the two exchanged individuals have repeated goods location point numbers, taking a complementary set of the exchanged segments to rearrange the complementary set of the exchanged segments to the non-exchanged segments randomly;
judging whether unselected individuals exist in the population;
under the condition that the unselected individuals exist in the population, the step of randomly selecting one unselected individual from the population and forming a parent individual with the smallest fitness in the historically generated population is returned again until the unselected individual does not exist in the population;
and under the condition that the unselected individuals do not exist in the population, outputting the population with the optimal population cross operation completed.
Optionally, the mutation operation comprises:
randomly generating two natural numbers r3And r4；
Randomly selecting two unselected individuals from the population;
exchanging selected individual natural numbers r3And r4The cargo site number of the location;
judging whether the population has unselected individuals;
under the condition that the unselected individuals exist in the population, the step of randomly selecting two unselected individuals from the population is returned again until the unselected individuals do not exist in the population;
and under the condition that the unselected individuals do not exist in the population, outputting the population after the mutation operation is finished.
According to the invention, the attributes of the AGV are divided into three attributes of loading in the picking process, loading in the returning process and empty according to different loading conditions of the AGV. According to the method, different priorities are given according to different attributes of the AGVs, and the collision avoidance rule is determined according to the priorities of the AGVs. A comparison of the priorities of the different AGV attributes is set forth in the following table.
TABLE 1
When the attributes of two AGV which generate conflict are different, the priority of the two AGV is judged, the picking trolley of the loaded goods (picking) has the highest priority, the loaded goods (returning) has the second priority, and the empty goods has the lowest priority.
When the attributes of two AGV which generate conflict are the same, judging the priority of the goods under the condition of loading, wherein the AGV with high priority of the goods is high in priority; for the case of empty vehicles, the priority is the same.
In another aspect, the present invention further provides a system for optimizing multiple AGV path planning problems based on dual hybrid particle swarm, wherein the system includes a processor configured to execute any one of the methods described above.
In yet another aspect, the invention also provides a computer readable storage medium having stored thereon instructions for reading by a machine to cause the machine to perform a method as described in any one of the above.
According to the technical scheme, the double-hybrid particle swarm-based multi-AGV path planning problem optimization method and system optimize the working sequence of the AGV by simultaneously combining the sequence of the goods sites and the optimal path of the goods sites, so that the AGV can efficiently complete all goods picking tasks on the optimal path on the premise of adopting the optimal path.
Additional features and advantages of embodiments of the invention will be set forth in the detailed description which follows.
Drawings
The accompanying drawings, which are included to provide a further understanding of the embodiments of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the embodiments of the invention without limiting the embodiments of the invention. In the drawings:
FIG. 1 is a flow chart of a method for optimizing a multiple AGV path planning problem based on dual hybrid particle swarm in accordance with one embodiment of the present invention;
FIG. 2 is a shelf profile of an automated sorting warehouse according to one embodiment of the present invention;
FIG. 3 is a flow diagram of a method of determining an optimal path according to one embodiment of the invention;
FIG. 4 is a flow diagram of a method of individual optimal crossover operation according to one embodiment of the present invention;
FIG. 5 is a flow diagram of a method of population-optimized crossover operation according to one embodiment of the present invention;
FIG. 6 is a flow diagram of a method of mutation operations in accordance with one embodiment of the present invention;
FIG. 7 is one of a comparison graph of convergence curves for an HGA algorithm and a method provided by the present invention, according to one embodiment of the present invention;
FIG. 8 is a second comparison graph of convergence curves of the HGA algorithm and the method provided by the present invention, according to one embodiment of the present invention;
figure 9 is a third comparison of convergence curves for the HGA algorithm and the method provided by the present invention, according to one embodiment of the present invention.
Detailed Description
The following detailed description of embodiments of the invention refers to the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating embodiments of the invention, are given by way of illustration and explanation only, not limitation.
Fig. 1 is a flowchart illustrating a method for optimizing a multiple AGV path planning problem based on dual hybrid particle swarm according to an embodiment of the present invention. In this fig. 1, the method may include:
in step S10, encoding a sequence of cargo sites of an automated sorting warehouse to produce an initial population;
in step S11, generating an optimal path by using a preset path generation algorithm, i.e., a × algorithm;
in step S12, calculating the number of times of parking waiting of each individual from the optimal path and each individual in the population;
in step S13, a fitness value of each individual of the population is calculated;
in step S14, performing individual optimal crossover operation on the population;
in step S15, performing population optimal crossover operation on the population after the individual optimal crossover operation;
in step S16, performing a mutation operation on the population after the population optimal crossover operation;
in step S17, it is determined whether a termination condition is currently satisfied;
in step S18, in the case where it is judged that the termination condition is not currently satisfied, returning again to the step of performing calculation of the fitness value of each individual of the population until it is judged that the termination condition is currently satisfied;
in step S19, in the case where it is judged that the termination condition is currently satisfied, the optimized recipe is output.
In the method shown in fig. 1, step S10 may be used to encode each stock location in the automated sorting warehouse, resulting in a corresponding sequence of stock locations. In an embodiment of the present invention, the encoding method may be a real number encoding method, that is, the number 0 may be used to represent the parking area of the AGV, and an integer is used as the cargo space to be transported. Taking the shelf distribution diagram as shown in fig. 2 as an example, in step S10, it is first necessary to encode all the shelves in the automated sorting warehouse with the numbers 1, 2, 3, …, etc. (real numbers), and to distinguish the docking area, the number 0 is introduced to represent the docking area, so as to obtain particles with the shapes of (0, 64, 25, 6, 3, 18, 0, 12, 8, 27, 55, 34, 0.). Wherein, (0, 64, 25, 6, 3, 18, 0) may indicate a path that the first AGV cart starts from the parking area 0, sequentially transports the rack 64, the rack 25, the rack 6, the rack 33, and the rack 18, and finally returns to the parking area 0 after the transport is completed. Based on the encoding principle, it can be seen that the number of the codes 0 in each particle is m +1, and m is the number of the transport paths of the AGV carts in the particle.
Step S11 may be used to generate an optimal path. The optimal path generation method may be in various forms known to those skilled in the art. However, in a preferred example of the present invention, the generating method may also include the steps as shown in fig. 3. In this fig. 3, the generating method may include:
in step S20, all points to be traveled are determined. Wherein, each point to be driven can comprise a starting point, a terminal point or a cargo position point of the AGV;
in step S21, a route set and a set to be traveled are initialized. Wherein the set of paths may wrap the start of an AGV cart. The set to be driven may include all points to be driven except the starting point in the set of all the freight sites;
in step S22, traversing and calculating the f value of each point to be driven in the set to be driven;
in step S23, the point to be traveled with the smallest f-number is selected as the current selection point;
in step S24, add the current selection point to the path set;
in step S25, it is determined whether the current selected point is an end point;
in step S26, when it is determined that the current selected point is not the end point, it is determined whether the set to be driven includes at least one point to be driven adjacent to the current selected point;
in step S27, in a case where it is determined that the set to be traveled includes at least one point to be traveled that is adjacent to the current selection point, adding the point to be traveled that is adjacent to the current selection point to the route set;
in step S28, the point to be traveled adjacent to the current selection point is taken as a child node of the current selection point, and the f-value of the child node is calculated according to equations (1) to (3),
f(j)＝g(j)+h(j) (1)
g(j)＝g(i)+|xj-xi|+|yj-yi| (2)
wherein f (j) is the f value of the child node, g (j) is the cost value of the child node from the starting point, h (j) is the predicted cost value from the child node to the terminal point, g (i) is the cost value of the father node from the starting point, the father node is the current selection point, i is the number of the father node, u (j) is the cost value of the father node from the starting point0The number of the terminal point is j, the number of the child node is j, x is the abscissa of the point to be driven, y is the ordinate of the point to be driven, and D is the cost value of each unit distance of the AGV trolley;
in step S29, in a case where it is determined that the set to be driven does not include a point to be driven adjacent to the current selection point, it is determined whether the predicted cost values of the points to be driven adjacent to the current selection point are all less than the predicted cost value in the previous iteration;
in step S30, in a case where it is determined that the predicted cost values of the to-be-driven points adjacent to the current selection point are all smaller than the predicted cost value in the previous iteration, the current selection point is taken as a parent node of the to-be-driven point adjacent to the current selection point, and the f value of the parent node, the cost value of the parent node from the start point, and the predicted cost value from the parent node to the end point are updated according to formulas (1) to (3);
returning to the step of selecting the freight site with the minimum f value until the current selection point is judged as the terminal point;
in step S31, when it is determined that the current selection point is the end point, the output path set is the optimal path.
In the method as shown in fig. 3, step S20 may be used to obtain information of all points to be traveled. The information of each point to be traveled may include corresponding position coordinates, and the point to be traveled may be, for example, a start point, an end point, or a cargo location point of the AGV. Step S21 may be used to determine the set of paths that the AGV may need to travel and the set to travel that is not. In step S21, since the AGV can start from the starting point only in the initial state, there is only one starting point in the path set. And all the points to be driven except the starting point can be included in the corresponding set to be driven which is not driven. Step S22 may be used to traverse and calculate the f-value for each point to be traveled. The specific manner of calculating the f value may be a variety of methods known to those skilled in the art. For example, the f value in the initial state may be determined based on the distance between each point to be traveled and the starting point. And the steps S23 to S30 are used for sequentially selecting points to be traveled from the set to be traveled and adding the points to be traveled into the path set on the basis of the optimal path, so that the planning of the optimal path is completed. Specifically, step S23 may be used to select the point to be traveled closest to the current position of the AGV car to join the set to be traveled. Since the current position of the AGV cart will obviously by default become the position of the current selection point in the case where the current selection point is added to the set of paths. Starting from the current selection point, the nearest neighboring point to be traveled would obviously be the closest path. Therefore, step S26 determines whether at least one point to be traveled adjacent to the currently selected point is included in the set to be traveled. When it is determined that the to-be-driven set includes at least one to-be-driven point adjacent to the currently selected point, the adjacent to-be-driven point may be directly added to the to-be-driven set at this time, so as to reduce the number of iterations from step S23 to step S30. In addition, since there may be a plurality of adjacent points to be traveled, one of the plurality of adjacent points to be traveled may be randomly selected to join the route set. In step S26, if the set to be traveled does not include any point to be traveled adjacent to the currently selected point, this indicates that the point to be traveled cannot be added directly to the route set through steps S27 and S28, and a new round of iteration can be performed only by updating information such as the f-number of the point to be traveled adjacent to the currently selected point, and steps S29 and S30. In the case that it is determined in step S25 that the current selected point is the end point, this indicates that all the points to be traveled have been added to the route set, that is, the optimal route planning has been completed, so step S31 may be executed to output the optimal route.
After the planning of the optimal path is completed in step S11, a calculation may be performed for each individual (particle) in the population generated in step S10 to determine the fitness of each particle (steps S12 and S13). Although the calculation method of the fitness may be various methods known to those skilled in the art, in consideration of the combination with the encoding method and the optimal path generation method adopted in step S10 and step S11, in a preferred example of the present invention, the calculation method may be to calculate the fitness value according to formula (4) and formula (5),
Fitness(i)＝max(t1,t2,…,tk,…,tm) (4)
wherein, t1,t2,…,tk,…,tmTotal transport time, t, for AGV carts numbered 1, 2, …, k, …, m, respectivelykIs a plurality of TkAnd, TkAGV with designation number kkThe time required for carrying the goods once,
wherein d isi1Indicating the first Manhattan distance of task i, i.e. the Manhattan distance from the start of task i to the shelf, di2The second Manhattan distance representing task i, i.e. the Manhattan distance from the shelf to the picking station, di3A third segment Manhattan distance representing task i, i.e., the Manhattan distance between the pick table to the destination, and di1＝|xs-xi|+|ys-yi|，di2＝|xi-1|+|yi-yp|，di3＝|1-xe|+|yp-ye|，xsIs the abscissa of the start of task i, ysIs the ordinate, x, of the start of task iiAbscissa of shelf for task i, yiOrdinate, y, of the shelf for task ipOrdinate, x, of picking station for task ieIs the abscissa, y, of the end point of task ieIs the ordinate, x, of the end point of task iikFor indicating whether task i is carried by AGVkIndicating variable of transport, xikThe value of 1 indicates that the task i is carried out by an AGVkCarrying, otherwise, taking the value as 0; t is twIndicating the AGV cart one-time parking waiting time.
After the fitness of each individual is calculated in step S13, since there is generally no optimal individual in the population formed in the initial state, the individuals may be crossed and mutated based on the calculated fitness value, so as to obtain better individuals. In the embodiment of the present invention, the crossing and mutation operations of the formed individuals, that is, the individual optimal crossing operation in step S14, the population optimal crossing operation in step S15, and the mutation operation in step S16, may be performed using steps S14 to S16.
Specifically, in step S14, the individual optimal crossover operation may include a method as illustrated in fig. 4. In fig. 4, the method for the individual optimal crossover operation may include:
in step S40, an unselected individual is randomly selected from the population to form a parent individual with the smallest fitness with the current population;
in step S41, two natural numbers r are randomly generated1And r2；
In step S42, the two individuals of the parent are exchanged for the natural number r1And r2A fragment in between;
in step S43, it is determined whether there are duplicate cargo space point numbers for both exchanged individuals;
in step S44, when it is determined that there are duplicate cargo space point numbers for the two exchanged individuals, the complementary sets of the exchanged segments are re-arranged randomly on the non-exchanged segments;
in step S45, it is determined whether there are unselected individuals in the population;
under the condition that the unselected individuals exist in the population, the step of randomly selecting one unselected individual from the population and forming a parent individual with the minimum fitness in the current population is executed again (namely the step of executing the step S40 is executed again) until the unselected individual does not exist in the population;
in step S46, in the case where it is determined that there is no unselected individual in the population, a population for which the individual optimum crossover operation is completed is output.
In step S15, the population optimal crossing operation may include steps as shown in fig. 5. In this fig. 5, the population optimal crossover operation may include:
in step S50, an unselected individual is randomly selected from the population, and an individual with the minimum fitness in the historically generated population is selected to form a parent individual;
in step S51, two natural numbers r are randomly generated1And r2；
In step S52, the two individuals of the parent are exchanged for the natural number r1And r2A fragment in between;
in step S53, it is determined whether there are duplicate cargo space point numbers for both exchanged individuals;
in step S54, when it is determined that there are duplicate cargo space point numbers for the two exchanged individuals, the complementary sets of the exchanged segments are re-arranged randomly on the non-exchanged segments;
in step S55, it is determined whether there are unselected individuals in the population;
under the condition that the unselected individuals exist in the population, the step of randomly selecting one unselected individual from the population and forming a parent individual with the smallest fitness in the historically generated population is executed again (namely the step of executing the step S50 is executed again) until the unselected individual does not exist in the population;
in step S56, in the case where it is determined that there is no unselected individual in the population, a population for which the population-optimal crossover operation is completed is output.
For the mutation operation in this step S16, a step as shown in fig. 6 may be included. In this fig. 6, the mutation operation may include:
in step S60, two natural numbers r are randomly generated3And r4；
In step S61, two unselected individuals are randomly selected from the population;
in step S62, the selected individual natural number r is exchanged3And r4The cargo site number of the location;
in step S63, it is determined whether there are unselected individuals in the population;
under the condition that the unselected individuals exist in the judged population, the step of randomly selecting two unselected individuals from the population is returned to be executed again (namely, the step of S61 is returned to be executed) until the unselected individuals do not exist in the judged population;
in step S64, when it is determined that there is no unselected individual in the population, the population after the mutation operation is completed is output.
According to the technical scheme, the double-hybrid particle swarm-based multi-AGV path planning problem optimization method and system optimize the working sequence of the AGV by simultaneously combining the sequence of the goods sites and the optimal path of the goods sites, so that the AGV can efficiently complete all goods picking tasks on the optimal path on the premise of adopting the optimal path.
In order to further verify the technical effects of the method and the system for optimizing the multiple AGV path planning problem based on the dual Hybrid particle swarm algorithm, the inventor optimizes the multiple AGV path planning problem based on the automated sorting warehouse shown in fig. 2 by respectively adopting a Hybrid Genetic Algorithm (HGA) commonly used in the prior art and the method for optimizing the multiple AGV path planning problem based on the dual Hybrid particle swarm provided by the invention, wherein the HGA algorithm is an algorithm for mixing GA and a.
The simulation experiment operating environment is Intel (R) core (TM) i5-5200U CPU @2.20GHz, the operating system is Windows10, and the simulation software is Matlab R2016 a. The specific parameter settings of the algorithm are shown in table 2:
TABLE 2
In table 2, n represents the number of tasks, i.e., the problem size; v represents the running speed of the AGV trolley and has the unit of m/s; m represents the number of AGV carts; p represents the number of sorting tables; popSize indicates population size; numIter represents the number of iterations of the algorithm; a represents the number of individuals; i denotes the number of iterations.
Based on the encoding method in the technical content, the HGA and the method provided by the present invention are optimized, and the optimization results are shown in tables 3 to 5.
TABLE 3
TABLE 4
TABLE 5
In tables 3 to 5, the fitness value (maximum transportation completion time) of the method provided by the invention under the scale of three tasks is obviously superior to that of the HGA, so that the method for optimizing the multiple AGV path planning problem based on the dual hybrid particle swarm provided by the invention has stronger searching capability and can obtain the solution of the problem more accurately.
The convergence curves of the HGA algorithm and the method provided by the present invention are shown in fig. 7 to fig. 9, and it can be seen from the results shown in the diagrams that, under the scale of three task numbers, the fitness value of the method provided by the present invention is better than that of the HGA algorithm, and the time for calculating the optimized path by the method provided by the present invention can be saved by 18.73% to the maximum compared with the path before optimization.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in the process, method, article, or apparatus that comprises the element.
The above are merely examples of the present application and are not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.
Claims (9)
1. A multi-AGV path planning problem optimization method based on double hybrid particle swarm is characterized by comprising the following steps:
encoding a sequence of goods locations of an automated sorting warehouse to produce an initial population;
generating an optimal path by adopting a preset path generation algorithm, namely an A algorithm;
calculating the number of times of parking waiting of each individual according to the optimal path and each individual in the population;
calculating a fitness value for each individual of the population;
performing individual optimal crossover operations on the population;
performing population optimal cross operation on the population after the individual optimal cross operation;
performing variation operation on the population after the population optimal cross operation;
judging whether the current end condition is met;
under the condition that the termination condition is not met currently, returning to the step of calculating the fitness value of each individual of the population until the termination condition is met currently;
and outputting the optimized scheme under the condition of judging that the termination condition is currently met.
2. The method of claim 1, wherein encoding the positions and paths of the automated sorting warehouse to generate the initial population comprises:
and the number 0 is adopted to represent the stopping area of the AGV trolley, and an integer is adopted as a goods position to be carried.
3. The method of claim 1, wherein the path generation algorithm comprises:
determining all points to be driven, wherein the points to be driven comprise a starting point, a terminal point or a cargo space point of an AGV trolley;
initializing a path set and a set to be traveled, wherein the path set comprises the starting point, and the set to be traveled comprises all the points to be traveled except the starting point in the set of all the freight sites;
traversing and calculating the f value of each point to be driven in the set to be driven;
selecting the point to be driven with the minimum f value as a current selection point;
adding the current selection point into the path set;
judging whether the current selected point is the end point;
under the condition that the current selection point is judged not to be the terminal point, judging whether the set to be driven contains at least one point to be driven adjacent to the current selection point;
adding the point to be traveled adjacent to the current selection point into the path set under the condition that the set to be traveled comprises at least one point to be traveled adjacent to the current selection point;
taking a point to be driven adjacent to the current selection point as a child node of the current selection point, and calculating f values of the child nodes according to formulas (1) to (3),
f(j)＝g(j)+h(j) (1)
g(j)＝g(i)+|xj-xi|+|yj-yi| (2)
wherein f (j) is the f value of the child node, g (j) is the cost value of the child node from the starting point, h (j) is the predicted cost value of the child node to the end point, g (i) is the cost value of the parent node from the starting point, the parent node is the current selection point, i is the number of the parent node, u (j) is the cost value of the parent node from the starting point, and0the number of the terminal point is j, the number of the child node is x, the abscissa of the point to be traveled is x, the ordinate of the point to be traveled is y, and D is a cost value of each unit distance traveled by the AGV;
under the condition that the set to be driven does not contain the point to be driven adjacent to the current selection point, judging whether the predicted cost values of the points to be driven adjacent to the current selection point are all smaller than the predicted cost value in the previous iteration;
under the condition that the estimated cost value of the point to be driven adjacent to the current selected point is judged to be smaller than the estimated cost value in the previous iteration, the current selected point is taken as a father node of the point to be driven adjacent to the current selected point, and the f value of the father node, the cost value of the father node from the starting point and the estimated cost value of the father node to the terminal point are updated according to formulas (1) to (3);
returning to the step of selecting the freight site with the minimum f value until the current selection point is judged to be the terminal point;
and under the condition that the current selection point is judged to be the end point, outputting the path set as the optimal path.
4. The method of claim 1, wherein calculating the fitness value for each individual comprises:
calculating the fitness value according to formula (4) and formula (5),
Fitness(i)＝max(t1,t2,…,tk,…,tm) (4)
wherein, t1,t2,…,tk,…,tmTotal transport time, t, for AGV carts numbered 1, 2, …, k, …, m, respectivelykIs a plurality of TkAnd, TkAGV with designation number kkThe time required for carrying the goods once,
wherein d isi1Indicating the first Manhattan distance of task i, i.e. the Manhattan distance from the start of task i to the shelf, di2The second Manhattan distance representing task i, i.e. the Manhattan distance from the shelf to the picking station, di3A third segment Manhattan distance representing task i, i.e., the Manhattan distance between the pick table to the destination, and di1＝|xs-xi|+|ys-yi|，di2＝|xi-1|+|yi-yp|，di3＝|1-xe|+|yp-ye|，xsIs the abscissa of the start of task i, ysIs the ordinate, x, of the start of task iiAbscissa of shelf for task i, yiOrdinate, y, of the shelf for task ipOrdinate, x, of picking station for task ieIs the abscissa, y, of the end point of task ieIs the ordinate, x, of the end point of task iikFor indicating that task i isWhether or not to be controlled by AGVkIndicating variable of transport, xikThe value of 1 indicates that the task i is carried out by an AGVkCarrying, otherwise, taking the value as 0; t is twIndicating the AGV cart one-time parking waiting time.
5. The method of claim 1, wherein the individual optimal crossover operation comprises:
randomly selecting an unselected individual from the population and an individual with the minimum fitness in the current population to form a parent individual;
randomly generating two natural numbers r1And r2；
Two individuals exchanging parents are in the natural number r1And r2A fragment in between;
judging whether the two exchanged individuals have repeated cargo space point numbers or not;
under the condition that the two exchanged individuals have repeated goods location point numbers, taking a complementary set of the exchanged segments to rearrange the complementary set of the exchanged segments to the non-exchanged segments randomly;
judging whether unselected individuals exist in the population;
under the condition that the unselected individuals exist in the population, the step of randomly selecting one unselected individual from the population and forming a parent individual with the smallest fitness in the current population is returned again until the unselected individual does not exist in the population;
and under the condition that the unselected individuals do not exist in the population, outputting the population with the optimal cross operation of the individuals.
6. The method of claim 1, wherein the population-optimal crossover operation comprises:
randomly selecting an unselected individual from the population and an individual with the minimum fitness in the historically generated population to form a parent individual;
randomly generating two natural numbers r1And r2；
Two individuals exchanging parents are in the natural number r1And r2A fragment in between;
judging whether the two exchanged individuals have repeated cargo space point numbers or not;
under the condition that the two exchanged individuals have repeated goods location point numbers, taking a complementary set of the exchanged segments to rearrange the complementary set of the exchanged segments to the non-exchanged segments randomly;
judging whether unselected individuals exist in the population;
under the condition that the unselected individuals exist in the population, the step of randomly selecting one unselected individual from the population and forming a parent individual with the smallest fitness in the historically generated population is returned again until the unselected individual does not exist in the population;
and under the condition that the unselected individuals do not exist in the population, outputting the population with the optimal population cross operation completed.
7. The method of claim 1, wherein the mutation operation comprises:
randomly generating two natural numbers r3And r4；
Randomly selecting two unselected individuals from the population;
exchanging selected individual natural numbers r3And r4The cargo site number of the location;
judging whether the population has unselected individuals;
under the condition that the unselected individuals exist in the population, the step of randomly selecting two unselected individuals from the population is returned again until the unselected individuals do not exist in the population;
and under the condition that the unselected individuals do not exist in the population, outputting the population after the mutation operation is finished.
8. A dual hybrid particle swarm-based multiple AGV path planning problem optimization system, the system comprising a processor configured to perform the method of any of claims 1 to 7.
9. A computer-readable storage medium having stored thereon instructions for reading by a machine to cause the machine to perform the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111454010.9A CN114358233A (en) | 2021-12-01 | 2021-12-01 | Multi-AGV path planning problem optimization method and system based on double-hybrid particle swarm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111454010.9A CN114358233A (en) | 2021-12-01 | 2021-12-01 | Multi-AGV path planning problem optimization method and system based on double-hybrid particle swarm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114358233A true CN114358233A (en) | 2022-04-15 |
Family
ID=81097246
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111454010.9A Pending CN114358233A (en) | 2021-12-01 | 2021-12-01 | Multi-AGV path planning problem optimization method and system based on double-hybrid particle swarm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114358233A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114913708A (en) * | 2022-07-18 | 2022-08-16 | 深圳市华睿智兴信息科技有限公司 | Parking path guiding system and method for intelligent parking lot |
| CN115759407A (en) * | 2022-11-18 | 2023-03-07 | 宝开(上海)智能物流科技有限公司 | Sorting system-based multidimensional scheduling scheme optimization method, system and storage medium |
| CN116540743A (en) * | 2023-07-04 | 2023-08-04 | 北京邮电大学 | Centralized scheduling real-time path planning method and device for high-speed sorting robot |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104408589A (en) * | 2014-10-24 | 2015-03-11 | 陕西科技大学 | AGV optimization scheduling method based on mixed particle swarm optimization |
| CN107727099A (en) * | 2017-09-29 | 2018-02-23 | 山东大学 | The more AGV scheduling of material transportation and paths planning method in a kind of factory |
| CN107808215A (en) * | 2017-10-23 | 2018-03-16 | 南昌大学 | A kind of goods yard distribution optimization method applied to the non-traditional layout warehouse of Flying V-types |
| CN109974711A (en) * | 2019-04-12 | 2019-07-05 | 重庆渝博创智能装备研究院有限公司 | A kind of AGV multiple target point autonomous navigation method towards wisdom factory |
| CN111596658A (en) * | 2020-05-11 | 2020-08-28 | 东莞理工学院 | Multi-AGV collision-free operation path planning method and scheduling system |
| CN112666949A (en) * | 2020-12-24 | 2021-04-16 | 武汉理工大学 | Ship path planning method, system and storage medium |
| CN112734324A (en) * | 2020-12-25 | 2021-04-30 | 哈尔滨工业大学 | Storage logistics AGV path planning algorithm based on ant colony algorithm and improved genetic algorithm |
| CN113673922A (en) * | 2021-07-09 | 2021-11-19 | 合肥工业大学 | Fishbone type warehouse layout-based multi-vehicle picking path problem optimization method and system |
-
2021
- 2021-12-01 CN CN202111454010.9A patent/CN114358233A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104408589A (en) * | 2014-10-24 | 2015-03-11 | 陕西科技大学 | AGV optimization scheduling method based on mixed particle swarm optimization |
| CN107727099A (en) * | 2017-09-29 | 2018-02-23 | 山东大学 | The more AGV scheduling of material transportation and paths planning method in a kind of factory |
| CN107808215A (en) * | 2017-10-23 | 2018-03-16 | 南昌大学 | A kind of goods yard distribution optimization method applied to the non-traditional layout warehouse of Flying V-types |
| CN109974711A (en) * | 2019-04-12 | 2019-07-05 | 重庆渝博创智能装备研究院有限公司 | A kind of AGV multiple target point autonomous navigation method towards wisdom factory |
| CN111596658A (en) * | 2020-05-11 | 2020-08-28 | 东莞理工学院 | Multi-AGV collision-free operation path planning method and scheduling system |
| CN112666949A (en) * | 2020-12-24 | 2021-04-16 | 武汉理工大学 | Ship path planning method, system and storage medium |
| CN112734324A (en) * | 2020-12-25 | 2021-04-30 | 哈尔滨工业大学 | Storage logistics AGV path planning algorithm based on ant colony algorithm and improved genetic algorithm |
| CN113673922A (en) * | 2021-07-09 | 2021-11-19 | 合肥工业大学 | Fishbone type warehouse layout-based multi-vehicle picking path problem optimization method and system |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114913708A (en) * | 2022-07-18 | 2022-08-16 | 深圳市华睿智兴信息科技有限公司 | Parking path guiding system and method for intelligent parking lot |
| CN115759407A (en) * | 2022-11-18 | 2023-03-07 | 宝开(上海)智能物流科技有限公司 | Sorting system-based multidimensional scheduling scheme optimization method, system and storage medium |
| CN116540743A (en) * | 2023-07-04 | 2023-08-04 | 北京邮电大学 | Centralized scheduling real-time path planning method and device for high-speed sorting robot |
| CN116540743B (en) * | 2023-07-04 | 2023-09-29 | 北京邮电大学 | Centralized scheduling real-time path planning method and device for high-speed sorting robot |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108764777B (en) | Electric logistics vehicle scheduling method and system with time window | |
| CN114358233A (en) | Multi-AGV path planning problem optimization method and system based on double-hybrid particle swarm | |
| CN112833905B (en) | Distributed multi-AGV collision-free path planning method based on improved A-x algorithm | |
| CN111091238A (en) | Automatic container terminal AGV intelligent scheduling method | |
| CN113074728B (en) | Multi-AGV path planning method based on jumping point routing and collaborative obstacle avoidance | |
| CN111860754A (en) | AGV scheduling method based on ant colony and genetic algorithm | |
| Jun et al. | Pickup and delivery problem with recharging for material handling systems utilising autonomous mobile robots | |
| CN110245890A (en) | Kinds of goods method for sorting and kinds of goods sorting system | |
| Liu et al. | Novel multi-objective resource allocation and activity scheduling for fourth party logistics | |
| CN112001541B (en) | Path optimization method based on improved genetic algorithm | |
| CN112561194B (en) | Integrated scheduling method and system for production and logistics of mixed flow shop | |
| CN113408988B (en) | Scheduling optimization method for warehouse system, electronic equipment and storage medium | |
| US20220292413A1 (en) | Bike sharing rebalancing optimization method based on adaptive neighborhood search algorithm | |
| CN113359702B (en) | Intelligent warehouse AGV operation optimization scheduling method based on water wave optimization-tabu search | |
| CN113222462B (en) | Strip mine multi-energy truck scheduling optimization method based on co-evolution | |
| Subbaiah et al. | Scheduling of AGVs and machines in FMS with makespan criteria using sheep flock heredity algorithm | |
| Takahashi et al. | Online optimization of AGV transport systems using deep reinforcement learning | |
| CN112508481A (en) | Intelligent storage multi-AGV scheduling method | |
| Wang et al. | Research and application of genetic algorithm in path planning of logistics distribution vehicle | |
| Shen et al. | Intelligent material distribution and optimization in the assembly process of large offshore crane lifting equipment | |
| CN114444239B (en) | Operation workshop movement track path guiding optimization method based on hybrid genetic algorithm | |
| CN115358455A (en) | Electric vehicle path optimization method considering battery replacement and hybrid time window constraints | |
| CN114707707A (en) | Method and system for scheduling AGV task based on improved genetic algorithm | |
| Chen et al. | Multi-AGVs scheduling with vehicle conflict consideration in ship outfitting items warehouse | |
| Gan et al. | Scheduling problems of automated guided vehicles in automated container terminals using a genetic 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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20220415 |
|
| RJ01 | Rejection of invention patent application after publication |