Discrete sectional type conveyor line system path optimization method based on improved genetic algorithm

Abstract

The invention discloses a discrete sectional type conveyor line system path optimization method based on an improved genetic algorithm. The method comprises the following steps: establishing a task scheduling model; the method comprises the following steps: modeling a conveying network of a tray conveying line, modeling the conveying network into an adjacent matrix of a weight graph G = (V, E), and calculating the shortest path of an upper and lower line port path of any conveying line through a Floyd algorithm; a multi-material collision avoidance path is planned, and collision of multiple materials on a conveying line is avoided by adopting a heuristic avoidance algorithm; a genetic simulated annealing algorithm is used to carry out iterative optimization on an initial population online task order sequence, sampling determination is carried out on a result obtained by genetic algorithm evolution through a Metropis sampling process of the simulated annealing algorithm, and then a result obtained by sampling is used as an initial population for the genetic algorithm to carry out next generation evolution operation until an optimal solution is found. The problem that the conveying line system capability and path optimization scheme cannot be accurately evaluated and evaluated in the prior art is solved.

Description

Technical Field

[0001] The invention relates to the technical field of discrete segmented conveyor path planning, and in particular to a discrete segmented conveyor line system path optimization method based on an improved genetic algorithm. Background Art

[0002] In the logistics warehousing system, the discrete segmented conveyor has a relatively strong conveying capacity. In the long-distance conveying process, the discrete segmented conveyor can guarantee the stability and reliability of the whole process to the greatest extent. The discrete segmented conveyor can support the conveying of a wide range of materials, most of which are medium and large materials with pallets. It is widely used in logistics warehousing systems. Due to the complex and changeable logistics process of the conveyor line, the working conditions are different, and there is no standard method to calculate and simulate the flow capacity and path optimization of the discrete segmented conveyor line. Therefore, the establishment and simulation of the discrete segmented conveyor line system model requires high professional and technical capabilities of engineers, resulting in the inability to accurately evaluate and assess the conveyor line system capacity and path optimization plan. Summary of the invention

[0003] The purpose of the present invention is to provide a method for optimizing the path of a discrete segmented conveyor line system based on an improved genetic algorithm in order to solve the above-mentioned problems.

[0004] The technical solution of the present invention is as follows:

[0005] A discrete segmented conveyor line system path optimization method based on an improved genetic algorithm comprises the following steps:

[0006] Establish a task scheduling model and describe the task scheduling problem of discrete segmented conveyor line system with a mathematical model;

[0007] Model the conveying network of the pallet conveyor line as an adjacency matrix of weighted graph G = (V, E), and calculate the shortest path of any conveyor line entrance and exit path through the Floyd algorithm;

[0008] Multi-material collision avoidance path planning, using heuristic avoidance algorithm to avoid collision of multiple materials on the conveyor line;

[0009] The genetic simulated annealing algorithm is used to iteratively optimize the order sequence of the initial population's online task orders. The Metroplis sampling process of the simulated annealing algorithm is used to sample and determine the results of the genetic algorithm evolution. The sampled results are then used as the initial population for the next generation of evolutionary operations of the genetic algorithm until the optimal solution is found.

[0010] Furthermore, the genetic simulated annealing algorithm specifically includes the following steps:

[0011] Coding: Segment coding is adopted. The first segment is the material sequence code, the second segment is the equipment code, and the third segment is the joint code of the upper and lower line port numbers;

[0012] Fitness function: Select the inverse of the objective function as the individual's fitness value. The longer the time to complete all tasks, the smaller the fitness function value. The fitness function is:

[0013]

[0014] Among them, T max It represents the maximum completion time of the equipment to perform the task through the conveyor line system, k represents the equipment number, and K represents the equipment number set;

[0015] Selection operator, selection operation: Roulette method is used for selection operation. According to the fitness function, the selection probability of each individual is:

[0016]

[0017] Among them, f i represents the fitness of the ith individual, Represents the fitness of all individuals, N p is the population size;

[0018] Crossover: Two-point crossover is used for material sequence coding and equipment coding, and the joint code for the upper and lower line port numbers remains unchanged;

[0019] Mutation: Encode the material sequence and use a multi-point movement mutation method based on the material order sequence to randomly select n in the parent material order sequence. l Genes are randomly inserted into other positions of the offspring to generate new individuals; for device coding, a multi-point mutation method based on the device coding genome is used to randomly insert n genes into the parent generation. m The group gene is randomly selected from its optional set to generate new individuals; for the joint code, a multi-point mutation mutation method based on the joint code gene is adopted to randomly select n in the parent generation. e genes, and randomly selects others from its selectable set to produce new individuals;

[0020] Initial temperature settings for control parameters:

[0021]

[0022] Among them, f max and f min is the fitness of the best and worst individuals in the current group; P r is the initial acceptance probability of the control parameter tf;

[0023] Generate a new solution S′ based on the current solution S mutation and calculate the increment Δt s =C(S′)-C(S), where C(S) is the evaluation function; if Δt s If it meets the Metropolis criterion, then accept the new solution, otherwise Accept S′ as the new current solution, thereby generating a new population P′ after simulated annealing, and evaluate P′;

[0024] Use the cooling function to cool down the control parameter tf:

[0025] tf (k+1) =λtf k ,

[0026] Among them, λ is the temperature reduction coefficient, which is a constant, λ∈(0,1);

[0027] Continue to iterate until the optimal solution is output after completing G rounds.

[0028] Furthermore, the following adaptive formula is used in the mutation:

[0029]

[0030] Among them, f max is the maximum fitness value in the group, f avg is the average fitness value of each generation, f′ is the larger fitness value of the two individuals to be crossed, f is the fitness value of the individual to be mutated, P c is the crossover probability, P m is the mutation probability,

[0031] Furthermore, the material sequence coding specifically includes the following steps:

[0032] The material sequence code Loads uses an integer code based on the maximum number of materials to be substituted n, specifically:

[0033] Loads={L1,L2,…,L n},

[0034] Among them, each gene position L of the material sequence encoding Loads i (i=1,2,…,n) represents the code of all the items to be picked up at this location. The order of the gene bits represents the order of picking up all the items to be picked up. The code length is equal to the total number of items to be picked up. The code value L i ∈[1,n],L i It can be randomly generated in the integer range from 1 to n, and there are no repeated values.

[0035] Furthermore, the device encoding specifically includes the following steps:

[0036] For the number of devices k, the device encoding Mac adopts the heuristic integer full permutation encoding method, taking out all possible permutations from k different devices, the number of which is k factorial, and each permutation order corresponds to a group of genes, so there are a total of k groups of genes, each group of genes has a total of k unit genes, and each unit gene in the group does not have duplication;

[0037] If the maximum number of materials to be picked up is n, and n is divisible by k!, the Mac code length The device code Each genome position M i (i=1,2,…,N m ) represents the arrangement order of a group of k devices, that is, a unit gene position in each group corresponds to a material sequence coding gene position.

[0038] Furthermore, the combined code of the upper and lower line port numbers is encoded in fixed-length hexadecimal format and is divided into two parts from left to right:

[0039] The first part on the left is the online port code O = {O1, O2, ..., O j}, the encoding length is:

[0040]

[0041] Where j represents the number of online ports, and the encoding value is O j ∈[1,j]; Note that when j is an odd number, the mutation may contain illegal codes and needs to be processed;

[0042] The second part on the right is the downline port code F = {F1, F2, ..., F k}, the encoding length is:

[0043]

[0044] Where k represents the number of offline ports, and the encoding value is F k ∈[1,k], note that when j is an odd number, the mutation may contain illegal codes, which need to be processed; the total encoding length of the entire joint code is N o +N f .

[0045] Furthermore, the establishment of the task scheduling model specifically includes the following steps:

[0046] A discrete segmented conveyor line system with k upstream devices, where k devices are assigned n k There are N online tasks in total, which are assigned to the r material online platform and then transported to the q material offline platform for execution;

[0047] Minimize the maximum completion time T of all equipment conveyor line systems:

[0048] T=min(max k∈K (T k )),

[0049] Among them, max k∈K (T k ) implements the maximum completion time for all equipment conveyor line systems;

[0050] I represents the online task set, K represents the equipment number set; R represents the material online port set; Q represents the material offline port set; i represents the online task number, i∈I; o represents the virtual start task number; f represents the virtual end task number; k represents the equipment number, k∈K; r represents the material online port number, r∈R; q represents the material offline port number, q∈Q;

[0051]

[0052] in, Constrain the decision variable value to determine whether device k executes task i;

[0053]

[0054] in, Constraints on the decision variable value whether device k executes task i and then executes task i′;

[0055] formula This means that each task can only be executed by one device;

[0056] formula Represents the variable relationship restriction constraint, that is, after device k completes task i, it must also perform other tasks;

[0057] formula: It means that the initial task of each device is virtual task o, the final task is virtual task f, and the time taken to execute the virtual task is 0;

[0058] formula Indicates the job order of each task performed by device k;

[0059]

[0060] Indicates whether the decision variable value constraint is that the material launch port of task i is r;

[0061] formula Indicates that each task can only be launched from one material launch port;

[0062] formula Indicates that at a certain Δtper During the cycle time, the maximum number of online tasks for the material online port r must be less than or equal to m.

[0063] Furthermore, the establishment of the task scheduling model also includes the following steps:

[0064] formula Indicates whether the decision variable value constraint is that the material outlet for task i is q;

[0065] formula Indicates that each task can only be offline from one material offline port;

[0066] formula It means that on the same device, the completion time of the previous task is guaranteed to be earlier than the start time of the next task. M is a large number.

[0067] formula:

[0068]

[0069] Indicates the execution time of equipment k processing the i-th material, time It obeys the uniform distribution F(x) with lower limit a and upper limit b;

[0070] formula It represents the comprehensive transportation time of material i from port r to port q. Represents the conveyor line edge e (u,v) The comprehensive delivery time;

[0071] V represents the vertex set of the conveyor line graph G;

[0072] E represents the edge set of the transmission line graph G;

[0073] u|v represents the vertex in the graph G, v∈V, u∈V;

[0074] uv,e (u,v) represents the edge in the graph G,

[0075] formula Indicates that the current material i enters the conveyor line side e (u,v) The time when the next material i' enters the conveyor line side e (u,v) moment;

[0076] formula Indicates that the current material i leaves the conveyor line edge e (u,v) The time when the next material i' leaves the conveyor line e (u,v) moment;

[0077] formula Indicates that the current material i enters the conveyor line side e (u,v) The time when material i leaves the conveyor line e (u,v) moment, The conveyor line side (u,v) The comprehensive delivery time;

[0078] formula It indicates the time it takes for the equipment numbered k to complete the last task i and the material on the conveyor line;

[0079] formula Indicates that the equipment numbered k completes n k The operation time of each task, Indicates the time when equipment k processes the i-th material;

[0080] formula Indicates that the total number of completed tasks of k devices is N.

[0081] Furthermore, the modeling of the conveying network of the pallet conveying line specifically includes the following steps:

[0082] The transportation network is modeled as an adjacency matrix of a weighted graph G = (V, E), and the intersection points and material entry and exit points are set as vertices of the graph G, and the two vertices of the existing path are set as the edges of the graph G.

[0083] Set graph G edge e (u,v) The weight is the length w of the conveyor line, the direction is from vertex u to v, and the speed of the conveyor line is vel hr , calculated according to the uniform linear motion formula, the conveyor line side e (u,v) The direct delivery time is At the same time, the priority method of path passing is set to determine whether lifting and conveying is required, as well as the priority passing order of two materials under different paths at the same time; e (u,v).level The smaller the value, the higher the priority, that is, the lifting and lowering conveying cycle time between the main line and the branch line is set:

[0084] Then the conveyor line edge e (u,v) The comprehensive delivery time is

[0085] By using the Floyd shortest path algorithm, the shortest path from node r to node q can be found, set as E rq , then the comprehensive transport time is:

[0086] Furthermore, the collision avoidance path planning of multiple materials specifically includes the following steps:

[0087] Create and maintain a w Changing transmission line capacity matrix A, Δt w is the matrix scanning interval, that is, every Δt w Long time, scan the matrix A;

[0088] Each element in matrix A is a (u，v) ,u,v are the corresponding vertices in graph G, uv is the corresponding edge in graph G, and here it also represents the row and column numbers of the elements of matrix A;

[0089] For the conveyor line transporting the i-th material, in the process of calculating the shortest path from node r to node q, time, For the i-th material entering the conveyor line edge e (u,v) At the moment, if a (u，v) <e (u,v).capacity , that is, the number of materials a on the current conveyor line (u，v) Less than the material capacity e at the current conveyor line (u,v).capacity , while e (u,v).level＝0 , that is, the highest priority is given to the transmission, then a (u，v) =a (u，v) +1, that is, the number of materials on the current conveyor line a (u，v) Add 1, Moment, a (u，v) =a (u，v) -1, that is, the number of materials on the current conveyor line a (u，v) minus 1;

[0090] exist moment, if a (u，v) =e (u,v).capacity , that is, the number of materials a on the current conveyor line (u，v) Equal to the material capacity e on the current conveyor line side (u,v).capacity ,but is the waiting time for the material, and a (u，v) =a (u，v) +1, that is, the material waits for a certain collision avoidance waiting time before entering the conveyor line side a (u，v) , the number of materials on the current conveyor line a (u，v) Add 1; Moment, a (u，v) =a (u，v) -1, that is, the number of materials on the current conveyor line a (u，v) Minus 1.

[0091] Compared with the prior art, the present invention has the following beneficial effects:

[0092] A method for optimizing the path of a discrete segmented conveyor system based on an improved genetic algorithm is proposed. First, an algorithm model is established through a discrete segmented conveyor system, initialization parameters are set, and genetic encoding is performed according to the characteristics of the conveyor line. Then, the conveyor line path is optimized through a genetic simulated annealing algorithm. In the process, the Floyd shortest path algorithm is used to determine the shortest path between the material conveyor line entry r and the conveyor line exit q. A heuristic path priority method is used to determine the shortest path comprehensive transportation time between the entry r and the conveyor line exit q. A heuristic avoidance algorithm is used to deal with the collision problem of multiple materials on the conveyor line. After multiple generations of iterative path optimization, the output objective function minimizes the maximum completion time of all equipment conveyors. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] Figure 1 Schematic diagram of the process of this application. DETAILED DESCRIPTION

[0094] It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0095] The features and performance of the present invention are further described in detail below in conjunction with the embodiments.

[0096] See also Figure 1 , a discrete segmented conveyor line system path optimization method based on an improved genetic algorithm, comprising the following steps:

[0097] Establish a task scheduling model and describe the task scheduling problem of discrete segmented conveyor line system with a mathematical model;

[0098] Model the conveying network of the pallet conveyor line as an adjacency matrix of weighted graph G = (V, E), and calculate the shortest path of any conveyor line entrance and exit path through the Floyd algorithm;

[0099] Multi-material collision avoidance path planning uses a heuristic avoidance algorithm to avoid collisions between multiple materials on the conveyor line; since the speed of the materials on the conveyor line is constant, the path planning problem can be converted into the shortest path problem with a known starting point and end point. A heuristic avoidance algorithm is used to determine whether the running trajectories of the materials overlap. If there is overlap in time and space, and the number of materials on the same conveyor line has reached the maximum capacity, it means that the materials will collide during the conveying process. For the collision problems that may occur during the operation of the conveyor line, a collision waiting processing strategy is used to obtain the shortest path with the least waiting time and the lowest collision probability.

[0100] The genetic simulated annealing algorithm is used to iteratively optimize the order sequence of the initial population's online task orders. The Metroplis sampling process of the simulated annealing algorithm is used to sample and determine the results of the genetic algorithm evolution. The sampled results are then used as the initial population for the next generation of evolutionary operations of the genetic algorithm until the optimal solution is found.

[0101] The genetic simulated annealing algorithm specifically includes the following steps:

[0102] Coding: Segment coding is adopted. The first segment is the material sequence code, the second segment is the equipment code, and the third segment is the joint code of the upper and lower line port numbers;

[0103] Material serial coding specifically includes the following steps:

[0104] The material sequence code Loads uses an integer code based on the maximum number of materials to be substituted n, specifically:

[0105] Loads={L1,L2,…,L n},

[0106] Among them, each gene position L of the material sequence encoding Loads i (i=1,2,…,n) represents the code of all the items to be picked up at this location. The order of the gene bits represents the order of picking up all the items to be picked up. The code length is equal to the total number of items to be picked up. The code value L i ∈[1,n],L i It can be randomly generated in the integer range from 1 to n, and there are no repeated values.

[0107] Device coding specifically includes the following steps:

[0108] For the number of devices k, the device encoding Mac adopts the heuristic integer full permutation encoding method, taking out all possible permutations from k different devices, the number of which is k factorial, and each permutation order corresponds to a group of genes, so there are a total of k groups of genes, each group of genes has a total of k unit genes, and each unit gene in the group does not have duplication;

[0109] If the maximum number of materials to be picked up is n, and n is divisible by k!, the Mac code length The device code Each genome position M i (i=1,2,…,N m ) represents the arrangement order of a group of k devices, that is, a unit gene position in each group corresponds to a material sequence coding gene position.

[0110] The combined code for the upper and lower line port numbers is a fixed-length hexadecimal code, which is divided into two parts from left to right:

[0111] The first part on the left is the online port code O = {O1, O2, ..., O j}, the encoding length is:

[0112]

[0113] Where j represents the number of online ports, and the encoding value is O j ∈[1,j]; Note that when j is an odd number, the mutation may contain illegal codes and needs to be processed;

[0114] The second part on the right is the downline port code F = {F1, F2, ..., F k}, the encoding length is:

[0115]

[0116] Where k represents the number of offline ports, and the encoding value is F k ∈[1,k], note that when j is an odd number, the mutation may contain illegal codes, which need to be processed; the total encoding length of the entire joint code is N o +N f .

[0117] Fitness function: Select the inverse of the objective function as the individual's fitness value. The longer the time to complete all tasks, the smaller the fitness function value. The fitness function is:

[0118]

[0119] Among them, T max It represents the maximum completion time of the equipment to perform the task through the conveyor line system, k represents the equipment number, and K represents the equipment number set;

[0120] Selection operator, selection operation: Roulette method is used for selection operation. According to the fitness function, the selection probability of each individual is:

[0121]

[0122] Among them, fi represents the fitness of the ith individual, Represents the fitness of all individuals, N p is the population size;

[0123] Crossover: Two-point crossover is used for material sequence coding and equipment coding, and the joint code for the upper and lower line port numbers remains unchanged;

[0124] Mutation: In order to diversify the population and prevent the algorithm from falling into a local optimum, it is necessary to perform mutation operations on the genes. To encode the material sequence, a multi-point moving mutation method based on the material order sequence is adopted, and n random l Genes are randomly inserted into other positions of the offspring to generate new individuals; for device coding, a multi-point mutation method based on the device coding genome is used to randomly insert n genes into the parent generation. m The group gene is randomly selected from its optional set to generate new individuals; for the joint code, a multi-point mutation mutation method based on the joint code gene is adopted to randomly select n in the parent generation. e genes, and randomly selects others from its selectable set to produce new individuals;

[0125] The following adaptive formula is used in the mutation:

[0126]

[0127] Among them, f max is the maximum fitness value in the group, f avg is the average fitness value of each generation, f′ is the larger fitness value of the two individuals to be crossed, f is the fitness value of the individual to be mutated, P c is the crossover probability, P m is the mutation probability,

[0128] Initial temperature settings for control parameters:

[0129]

[0130] Among them, f max and f min is the fitness of the best and worst individuals in the current group; P r is the initial acceptance probability of the control parameter tf;

[0131] Generate a new solution S′ based on the current solution S mutation and calculate the increment Δt s =C(S′)-C(S), where C(S) is the evaluation function; if Δt s If it meets the Metropolis criterion, then accept the new solution, otherwise Accept S′ as the new current solution, thereby generating a new population P′ after simulated annealing, and evaluate P′;

[0132] Use the cooling function to cool down the control parameter tf:

[0133] tf (k+1) =λtf k ,

[0134] Among them, λ is the temperature reduction coefficient, which is a constant, λ∈(0,1);

[0135] Continue to iterate until the optimal solution is output after completing G rounds.

[0136] Establishing a task scheduling model specifically includes the following steps:

[0137] A discrete segmented conveyor line system with k upstream devices, where k devices are assigned n k There are N online tasks in total, which are assigned to the r material online platform and then transported to the q material offline platform for execution;

[0138] Minimize the maximum completion time T of all equipment conveyor line systems:

[0139] T=min(max k∈K (T k )),

[0140] Among them, max k∈K (T k ) implements the maximum completion time for all equipment conveyor line systems;

[0141] I represents the online task set, K represents the equipment number set; R represents the material online port set; Q represents the material offline port set; i represents the online task number, i∈I; o represents the virtual start task number; f represents the virtual end task number; k represents the equipment number, k∈K; r represents the material online port number, r∈R; q represents the material offline port number, q∈Q;

[0142]

[0143] in, Constrain the decision variable value to determine whether device k executes task i;

[0144]

[0145] in, Constraints on the decision variable value whether device k executes task i and then executes task i′;

[0146] formula This means that each task can only be executed by one device;

[0147] formula Represents the variable relationship restriction constraint, that is, after device k completes task i, it must also perform other tasks;

[0148] formula: It means that the initial task of each device is virtual task o, the final task is virtual task f, and the time taken to execute the virtual task is 0;

[0149] formula Indicates the job order of each task performed by device k;

[0150]

[0151] Indicates whether the decision variable value constraint is that the material launch port of task i is r;

[0152] formula Indicates that each task can only be launched from one material launch port;

[0153] formula Indicates that at a certain Δt per During the cycle time, the maximum number of online tasks for the material online port r must be less than or equal to m.

[0154] formula Indicates whether the decision variable value constraint is that the material outlet for task i is q;

[0155] formula Indicates that each task can only be offline from one material offline port;

[0156] formula It means that on the same device, the completion time of the previous task is guaranteed to be earlier than the start time of the next task. M is a large number.

[0157] formula:

[0158]

[0159] Indicates the execution time of equipment k processing the i-th material, time It obeys the uniform distribution F(x) with lower limit a and upper limit b;

[0160] formula It represents the comprehensive transportation time of material i from port r to port q. Represents the conveyor line edge e (u,v) The comprehensive delivery time;

[0161] V represents the vertex set of the conveyor line graph G;

[0162] E represents the edge set of the transmission line graph G;

[0163] u|v represents the vertex in the graph G, v∈V, u∈V;

[0164] uv,e (u,v) represents the edge in graph G,

[0165] formula Indicates that the current material i enters the conveyor line side e (u,v) The time when the next material i' enters the conveyor line side e (u,v) moment;

[0166] formula Indicates that the current material i leaves the conveyor line edge e (u,v) The time when the next material i' leaves the conveyor line e (u,v) moment;

[0167] formula Indicates that the current material i enters the conveyor line side e (u,v) The time when material i leaves the conveyor line e (u,v) moment, The conveyor line side (u,v) The comprehensive delivery time;

[0168] formula It indicates the time it takes for the equipment numbered k to complete the last task i and the material on the conveyor line;

[0169] formula Indicates that the equipment numbered k completes n k The operation time of each task, Indicates the time when equipment k processes the i-th material;

[0170] formula Indicates that the total number of completed tasks of k devices is N.

[0171] Modeling the conveyor network of a pallet conveyor line includes the following steps:

[0172] The transportation network is modeled as an adjacency matrix of a weighted graph G = (V, E), and the intersection points and material entry and exit points are set as vertices of the graph G, and the two vertices of the existing path are set as the edges of the graph G.

[0173] Set graph G edge e (u,v) The weight is the length w of the conveyor line, the direction is from vertex u to v, and the speed of the conveyor line is vel hr , calculated according to the uniform linear motion formula, the conveyor line side e (u,v)The direct delivery time is At the same time, the priority method of path passing is set to determine whether lifting and conveying is required, as well as the priority passing order of two materials under different paths at the same time; e (u,v).level The smaller the value, the higher the priority, that is, the lifting and lowering conveying cycle time between the main line and the branch line is set:

[0174] Then the conveyor line edge e (u,v) The comprehensive delivery time is

[0175] By using the Floyd shortest path algorithm, the shortest path from node r to node q can be found, set as E rq , then the comprehensive transport time is:

[0176] The collision avoidance path planning of multiple materials specifically includes the following steps:

[0177] Create and maintain a w Changing transmission line capacity matrix A, Δt w is the matrix scanning interval, that is, every Δt w Long time, scan the matrix A;

[0178] Each element in matrix A is a (u，v) ,u,v are the corresponding vertices in graph G, uv is the corresponding edge in graph G, and here it also represents the row and column numbers of the elements of matrix A;

[0179] For the conveyor line transporting the i-th material, in the process of calculating the shortest path from node r to node q, time, For the i-th material entering the conveyor line edge e (u,v) At the moment, if a (u，v) <e (u,v).capacity , that is, the number of materials a on the current conveyor line (u，v) Less than the material capacity e at the current conveyor line (u,v).capacity , while e (u,v).level＝0 , that is, the highest priority is given to the transmission, then a (u，v) =a (u，v) +1, that is, the number of materials on the current conveyor line a (u，v) Add 1, Moment, a (u，v) =a (u，v) -1, that is, the number of materials on the current conveyor line a (u，v) minus 1;

[0180] exist moment, if a (u，v) =e(u,v).capacity , that is, the number of materials a on the current conveyor line (u，v) Equal to the material capacity e on the current conveyor line side (u,v).capacity ,but Δt wait is the waiting time for the material, and a (u，v) =a (u，v) +1, that is, the material waits for a certain collision avoidance waiting time before entering the conveyor line side a (u，v) , the number of materials on the current conveyor line a (u，v) Add 1; Moment, a (u，v) =a (u，v) -1, that is, the number of materials on the current conveyor line a (u，v) Minus 1.

[0181] In another specific embodiment, 1: after the task scheduling model is constructed, the transportation network is modeled, and the collision avoidance path of multiple materials is planned;

[0182] 2: Initialize a transmission line capacity matrix A, for the edge e in the graph G 9u，v) ∈E, element a in A (u，v) = 0, edge e does not exist in graph G 9u’，v’) ∈E, element a in A (u，v) =∞;

[0183] 3: Establish the corresponding coding gene sequence according to the constraints and the actual delivery system characteristics;

[0184] 4: Solve the objective function T through the algorithm model of genetic simulated annealing algorithm:

[0185] 4.1: According to the problem to be solved, set relevant parameters, including: population size N p , the population size represents the online material order list. According to the constraints, the tasks in the task set I are randomly divided into k parts, and the number of each part is n k , randomly assigned to each device in the device number set K, the generation scale is N p =N's initial population P;

[0186] 4.2: According to the algorithm model, the maximum completion time T′=max for each individual equipment conveyor system under this initial population P k∈K (T k ), and determine the maximum number of iterations G and the cooling control value P r , cooling rate λ, end temperature t e , current generation gen=1;

[0187] 4.3: According to T', the probability of each individual selection P is obtained i, the individual fitness function f, f max , f avg , and sort and evaluate;

[0188] 4.4: Determine whether the algorithm convergence condition is satisfied? If so, P = P′, go to 4.9, otherwise go to 4.5.

[0189] 4.5: Generate a new population P′:

[0190] (1) Select the parents. Through roulette wheel selection, select the individuals with high fitness in the population to enter the next operation;

[0191] (2) Adopt the adaptive crossover operator P c , the adaptive mutation operator P m , perform crossover and mutation operations on the parent individuals to generate offspring individuals;

[0192] (3) Add the offspring individuals after crossover and mutation to the population P′, and determine whether the number of offspring individuals reaches the population size N p , if so, go to 4.6, otherwise go to (1);

[0193] 4.6: Perform simulated annealing on all newly generated offspring individuals:

[0194] For each offspring individual, set the initial temperature of the control parameter to tf0, and the initial state to S, S = T′;

[0195] Generate a new solution S′ through mutation.

[0196] According to the Metroplis criterion, calculate the increment Δt s = C(S′) - C(S), where C(S) is the evaluation function.

[0197] If Δt s < 0, then accept S′ as the new current solution, otherwise with probability accept S′ as the new current solution.

[0198] If the termination condition tf ≤ t e is satisfied, then output the current solution as the optimal solution, and then go to 4.7, otherwise go to [6].

[0199] tf gradually decreases, and tf > 0, then go to [2].

[0200] 4.7: Through the new population P′ after simulated annealing, then use the algorithm model established in steps 1, 2, and 3 to obtain the maximum completion time T′ of all equipment conveyor line systems of the gene individuals. If T′ < T, then T = T′;

[0201] 4.8: Recalculate the selection probability P of each individual in the newly generated populationi , each individual fitness function f, f max , f avg , f′, current algebra gen=gen+1, go to 4.4.

[0202] 4.9: Output the results of the genetic simulated annealing algorithm, including the final population P, which minimizes the maximum completion time T executed by all equipment conveyor line systems.

[0203] The above-mentioned embodiments only express the specific implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the protection scope of the present application. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the technical solution concept of the present application, and these all belong to the protection scope of the present application.

Claims

1. A discrete segmented conveyor line system path optimization method based on an improved genetic algorithm, characterized in that: The following steps are involved: Establish a task scheduling model and describe the task scheduling problem of discrete segmented conveyor line system with a mathematical model; Model the conveying network of the pallet conveyor line as an adjacency matrix of weighted graph G = (V, E), and calculate the shortest path of any conveyor line entrance and exit path through the Floyd algorithm; Multi-material collision avoidance path planning, using heuristic avoidance algorithm to avoid collision of multiple materials on the conveyor line; The genetic simulated annealing algorithm is used to iteratively optimize the order sequence of the initial population's online task orders. The Metroplis sampling process of the simulated annealing algorithm is used to sample and determine the results of the genetic algorithm evolution. The sampled results are then used as the initial population for the next generation of evolutionary operations of the genetic algorithm until the optimal solution is found.

2. The method for optimizing the path of a discrete segmented conveyor line system based on an improved genetic algorithm according to claim 1, characterized in that: The genetic simulated annealing algorithm specifically comprises the following steps: Coding: Segment coding is adopted. The first segment is the material sequence code, the second segment is the equipment code, and the third segment is the joint code of the upper and lower line port numbers; Fitness function: Select the inverse of the objective function as the individual's fitness value. The longer the time to complete all tasks, the smaller the fitness function value. The fitness function is: Among them, T max It represents the maximum completion time of the equipment to perform the task through the conveyor line system, k represents the equipment number, and K represents the equipment number set; Selection operator, selection operation: Roulette method is used for selection operation. According to the fitness function, the selection probability of each individual is: Among them, f i represents the fitness of the ith individual, Represents the fitness of all individuals, N p is the population size; Crossover: Two-point crossover is used for material sequence coding and equipment coding, and the joint code for the upper and lower line port numbers remains unchanged; Mutation: Encode the material sequence and use a multi-point movement mutation method based on the material order sequence to randomly select n in the parent material order sequence. l Genes are randomly inserted into other positions of the offspring to generate new individuals; for device coding, a multi-point mutation method based on the device coding genome is used to randomly insert n genes into the parent generation. m The group gene is randomly selected from its optional set to generate new individuals; for the joint code, a multi-point mutation mutation method based on the joint code gene is adopted to randomly select n in the parent generation. e genes, and randomly selects others from its selectable set to produce new individuals; Initial temperature settings for control parameters: Among them, f max and f min is the fitness of the best and worst individuals in the current group; P r is the initial acceptance probability of the control parameter tf; Generate a new solution S′ based on the current solution S mutation and calculate the increment Δt s =C(S′)-C(S), where C(S) is the evaluation function; if Δt s If it meets the Metropolis criterion, then accept the new solution, otherwise Accept S′ as the new current solution, thereby generating a new population P′ after simulated annealing, and evaluate P′; Use the cooling function to cool down the control parameter tf: tf (k+1) =λtf k , Among them, λ is the temperature reduction coefficient, which is a constant, λ∈(0,1); Continue to iterate until the optimal solution is output after completing G rounds.

3. The method for optimizing the path of a discrete segmented conveyor line system based on an improved genetic algorithm according to claim 2, characterized in that: The following adaptive formula is used in the mutation: Among them, f max is the maximum fitness value in the group, f avg is the average fitness value of each generation, f′ is the larger fitness value of the two individuals to be crossed, f is the fitness value of the individual to be mutated, P c is the crossover probability, P m is the mutation probability, 4. A method for optimizing the path of a discrete segmented conveyor line system based on an improved genetic algorithm according to claim 2 or 3, characterized in that: The material sequence coding specifically includes the following steps: The material sequence code Loads uses an integer code based on the maximum number of materials to be substituted n, specifically: Loads={L1,L2,…,L n }, Among them, each gene position L of the material sequence encoding Loads i (i=1,2,…,n) represents the code of all the items to be picked up at this location. The order of the gene bits represents the order of picking up all the items to be picked up. The code length is equal to the total number of items to be picked up. The code value L i ∈[1,n],L i It can be randomly generated in the integer range from 1 to n, and there are no repeated values.

5. A method for optimizing the path of a discrete segmented conveyor line system based on an improved genetic algorithm according to claim 2 or 3, characterized in that: The device coding specifically comprises the following steps: For the number of devices k, the device encoding Mac adopts the heuristic integer full permutation encoding method, taking out all possible permutations from k different devices, the number of which is k factorial, and each permutation order corresponds to a group of genes, so there are a total of k groups of genes, each group of genes has a total of k unit genes, and each unit gene in the group does not have duplication; If the maximum number of materials to be picked up is n, and n is divisible by k!, the Mac code length The device code Each genome position M i (i=1,2,…,N m ) represents the arrangement order of a group of k devices, that is, a unit gene position in each group corresponds to a material sequence coding gene position.

6. A method for optimizing the path of a discrete segmented conveyor line system based on an improved genetic algorithm according to claim 2 or 3, characterized in that: The combined code for the upper and lower line port numbers is encoded in fixed-length hexadecimal format and is divided into two parts from left to right: The first part on the left is the online port code O = {O1, O2, ..., O j }, the encoding length is: Where j represents the number of online ports, and the encoding value is O j ∈[1,j]; Note that when j is an odd number, the mutation may contain illegal codes and needs to be processed; The second part on the right is the downline port code F = {F1, F2, ..., F k }, the encoding length is: Where k represents the number of offline ports, and the encoding value is F k ∈[1,k], note that when j is an odd number, the mutation may contain illegal codes, which need to be processed; the total encoding length of the entire joint code is N o +N f .

7. The method for optimizing the path of a discrete segmented conveyor line system based on an improved genetic algorithm according to claim 1, characterized in that: The task scheduling model is established by: A discrete segmented conveyor line system with k upstream devices, where k devices are assigned n k There are N online tasks in total, which are assigned to the r material online platform and then transported to the q material offline platform for execution; Minimize the maximum completion time T of all equipment conveyor line systems: T=min(max k∈K (T k )), Among them, max k∈K (T k ) implements the maximum completion time for all equipment conveyor line systems; I represents the online task set, K represents the equipment number set; R represents the material online port set; Q represents the material offline port set; i represents the online task number, i∈O; o is the virtual start task number; f is the virtual end task number; k is the equipment number, k∈K; r is the material online port number, r∈R; q is the material offline port number, q∈Q; in, Constrain the decision variable value to determine whether device k executes task i; in, Constrain the decision variable value to determine whether device k executes task i and then executes task i′; formula This means that each task can only be executed by one device; formula Represents the variable relationship restriction constraint, that is, after device k completes task i, it must also perform other tasks; formula: It means that the initial task of each device is virtual task o, the final task is virtual task f, and the time taken to execute the virtual task is 0; formula Indicates the job order of each task performed by device k; Indicates whether the decision variable value constraint is that the material launch port of task i is r; formula Indicates that each task can only be launched from one material launch port; formula Indicates that at a certain Δt per During the cycle time, the maximum number of online tasks for the material online port r must be less than or equal to m.

8. The method for optimizing the path of a discrete segmented conveyor line system based on an improved genetic algorithm according to claim 7, characterized in that: The task scheduling model establishment further comprises the following steps: formula Indicates whether the decision variable value constraint is that the material outlet for task i is q; formula Indicates that each task can only be offline from one material offline port; formula It means that on the same device, the completion time of the previous task is guaranteed to be earlier than the start time of the next task. M is a large number. official: Indicates the execution time of equipment k processing the i-th material, time It obeys the uniform distribution F(x) with lower limit a and upper limit b; formula It represents the comprehensive transportation time of material i from port r to port q. Represents the conveyor line edge e (u,v) The comprehensive delivery time; V represents the vertex set of the conveyor line graph G; E represents the edge set of the transmission line graph G; u|v represents the vertex in the graph G, v∈V, u∈V; uv,e (u,v) represents the edge in graph G, formula Indicates that the current material i enters the conveyor line side e (u,v) The time when the next material i' enters the conveyor line side e (u,v) moment; formula Indicates that the current material i leaves the conveyor line edge e (u,v) The time when the next material i' leaves the conveyor line e (u,v) moment; formula Indicates that the current material i enters the conveyor line side e (u,v) The time when material i leaves the conveyor line is earlier than (u,v) moment, The conveyor line side (u,v) The comprehensive delivery time; formula It indicates the time it takes for the equipment numbered k to complete the last task i and the material on the conveyor line; formula Indicates that the equipment numbered k completes n k The operation time of each task, Indicates the time when equipment k processes the i-th material; formula Indicates that the total number of completed tasks of k devices is N.

9. The method for optimizing the path of a discrete segmented conveyor line system based on an improved genetic algorithm according to claim 8, characterized in that: The modeling of the conveying network of the pallet conveying line specifically includes the following steps: The transportation network is modeled as an adjacency matrix of a weighted graph G = (V, E), with intersections and material entry and exit points set as vertices of the graph G, and the two vertices of the existing path set as the edges of the graph G. Set graph G edge e (u,v) The weight is the length w of the conveyor line, the direction is from vertex u to v, and the speed of the conveyor line is vel hr , calculated according to the uniform linear motion formula, the conveyor line side e (u,v) The direct delivery time is At the same time, the priority method of path passing is set to determine whether lifting and conveying is required, as well as the priority passing order of two materials under different paths at the same time; e (u,v).level The smaller the value, the higher the priority, that is, the lifting and lowering conveying cycle time between the main line and the branch line is set: Then the conveyor line edge e (u,v) The comprehensive delivery time is By using the Floyd shortest path algorithm, the shortest path from node r to node q can be found, set as E rq , then the comprehensive transport time is:

10. The method for optimizing the path of a discrete segmented conveyor line system based on an improved genetic algorithm according to claim 1, characterized in that: The multi-material collision avoidance path planning specifically includes the following steps: Create and maintain a w Changing transmission line capacity matrix A, Δt w is the matrix scanning interval, that is, every Δt w Long time, scan the matrix A; Each element in matrix A is a (u，v) ,u,v are the corresponding vertices in graph G, uv is the corresponding edge in graph G, and here it also represents the row and column numbers of the elements of matrix A; For the conveyor line transporting the i-th material, in the process of calculating the shortest path from node r to node q, time, For the i-th material entering the conveyor line edge e (u,v) At the moment, if a (u，v) <e (u,v).capacity , that is, the number of materials a on the current conveyor line (u，v) Less than the material capacity e at the current conveyor line (u,v).capacity , while e (u,v).level＝0 , that is, the highest priority is given to the transmission, then a (u，v) =a (u，v) +1, that is, the number of materials on the current conveyor line a (u，v) Add 1, Moment, a (u，v) =a (u，v) -1, that is, the number of materials on the current conveyor line a (u，v) minus 1; exist moment, if a (u，v) =e (u,v).capacity , that is, the number of materials a on the current conveyor line (u，v) Equal to the material capacity e on the current conveyor line side (u,v).capacity ,but Δt wait is the waiting time for the material, and a (u，v) =a (u，v) +1, that is, the material waits for the collision avoidance waiting time before entering the conveyor line side a (u，v) , the number of materials on the current conveyor line a (u，v) Add 1; Moment, a (u，v) =a (u，v) -1, that is, the number of materials on the current conveyor line a (u，v) Minus 1.

Images