Multi-target stereoscopic warehouse goods allocation optimization method based on adaptive genetic algorithm

Abstract

The invention discloses a multi-target stereoscopic warehouse goods allocation optimization method based on an adaptive genetic algorithm. Firstly, three-dimensional digital modeling is carried out on a warehouse and goods shelves in the warehouse; then, constructing a multi-target goods allocation optimization mathematical model according to a material stability principle, a material fastest ex-warehouse principle, a material aggregation principle and a different material relevancy principle; and finally, solving the multi-target goods allocation optimization mathematical model by adopting a self-adaptive genetic algorithm, and obtaining an optimal solution which is an optimal goods allocation scheme of the goods and materials before warehousing. The method provided by the invention can effectively optimize the allocation distribution of the warehousing materials, thereby greatly improving the storage efficiency and operation efficiency of the automatic stereoscopic warehouse, and reducing the operation and maintenance cost of the warehouse.

Description

Technical Field

[0001] The invention belongs to the technical field of automated stereoscopic warehouse storage, and in particular relates to a multi-objective stereoscopic warehouse cargo location allocation optimization method based on an adaptive genetic algorithm. Background Art

[0002] Under the influence of the economic environment and relevant policies, the efficiency and management level of global warehousing are growing rapidly and developing steadily. Due to the reduction of land space, the increase of labor costs and the increase of logistics demand for warehouse efficiency, it is necessary to improve space utilization, reduce labor costs, enhance rapid response capabilities and improve operation accuracy. Warehouse layout, cargo location allocation and warehouse scheduling are important components of warehouse operation and warehouse management, which have attracted the attention of scholars around the world.

[0003] An automated high-bay warehouse is an advanced storage system that usually uses automated equipment such as stackers, conveyor belts, etc. to maximize the utilization of warehouse space and the efficiency of cargo handling. Cargo space allocation is a core issue in high-bay warehouse management, which involves how to reasonably store goods of different types, sizes, and weights in various cargo spaces in the warehouse. The problem of cargo space allocation for incoming goods solves how to place the goods in the appropriate cargo space when they are incoming according to the goals they need to achieve. There are many principles to consider for cargo space allocation, such as the principle of light on top and heavy on the bottom to consider the load-bearing capacity of the shelves, the first-in-first-out principle to improve the efficiency of warehousing and out-of-warehouse, and the principle of storage in different aisles to improve the overall operational efficiency.

[0004] The optimization of cargo location allocation is a multi-objective optimization problem, and intelligent algorithms are needed to solve multi-objective optimization problems. Among various intelligent algorithms, genetic algorithms have the characteristics of good robustness, strong global search ability, and high fault tolerance, and are a powerful tool for solving multi-objective optimization problems. Genetic algorithms are a method that simulates the natural evolution process to find the optimal solution, and many research results have been achieved. In addition, genetic algorithms have been widely used in cargo location allocation and optimization in stereoscopic warehouses. Compared with traditional optimization methods (such as linear, integer, and dynamic programming), genetic algorithms have excellent global search capabilities, good convergence, and high robustness. However, genetic algorithms also have the disadvantages of premature, poor local search capabilities, low efficiency in late search, and sensitivity to variable parameters. Summary of the invention

[0005] The purpose of the present invention is to propose a multi-objective stereoscopic warehouse cargo space allocation optimization method based on an adaptive genetic algorithm to obtain the optimal cargo space configuration plan for the current materials, thereby effectively optimizing the cargo space distribution of the incoming materials.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0007] A multi-objective stereoscopic warehouse cargo space allocation optimization method based on an adaptive genetic algorithm, comprising:

[0008] Carry out 3D digital modeling of the warehouse and the shelves therein; take the pallet exit as the origin of the 3D coordinate system, take the center of the material pallet to represent the position of the material pallet in the 3D coordinate system, and initialize the 3D coordinates of each material pallet;

[0009] The objective function is constructed based on the material stability principle to calculate the shelf stability index;

[0010] The objective function is established based on the principle of the fastest material outbound delivery, which is used to calculate the fastest material outbound delivery index;

[0011] The objective function is established based on the material aggregation principle to calculate the material aggregation index;

[0012] The objective function is established based on the material relevance principle to calculate the material relevance index;

[0013] Based on the shelf local stability index, the fastest material outbound index, the material aggregation index and the material relevance index, a multi-objective cargo location optimization function is constructed, and the multi-objective cargo location optimization function is converted into a single-objective optimization problem;

[0014] The single-objective optimization problem is solved using an adaptive genetic algorithm to obtain the optimal cargo space allocation solution.

[0015] Furthermore, it is assumed that there are a rows of shelves in the warehouse, and each row of shelves has b columns and c layers; the warehouse is equipped with an access platform, which has a pallet outbound port and multiple pallet inbound ports, and the materials are all assembled on the material pallets; the coordinate (0,0,0) represents the location of the pallet outbound port; the three-dimensional coordinates of the center of the i-th material pallet in the warehouse cargo space are (x i ,y i ,z i ), x i =1,2,...,a;y i =1,2,...,b;z i =1,2,...,c; all material pallets will be randomly assigned an initial storage location before entering the warehouse, and the three-dimensional coordinates of the initial storage location of the pallet to be entered will be digitally initialized.

[0016] Furthermore, the objective function is constructed based on the material stability principle to calculate the shelf stability index, including:

[0017] The local stability of the shelf is modeled as follows:

[0018]

[0019]

[0020] Among them, h local Represents the local stability index of the shelf, N tray represents the number of pallets to be put into storage; X represents the coordinate value set of the material pallet, L z is the length of the shelf unit in the layer direction; x i and z i Respectively represent the row coordinate and layer coordinate of the i-th material pallet; t i Indicates the type of materials on the i-th material pallet; For the tth i Unit mass of such materials;

[0021] The global stability of the shelf is modeled as follows:

[0022]

[0023] Among them, h global represents the global stability index of the shelf;

[0024] Considering the local stability and global stability of the shelf, the shelf stability index is expressed as follows:

[0025] f1=w local ·h local +w global ·h global

[0026]

[0027] Among them, f1 represents the shelf stability index, w local and w global They represent the weight coefficients of local stability and global stability of the shelf respectively.

[0028] Furthermore, the objective function is established based on the principle of the fastest material delivery, which is used to calculate the fastest material delivery index, including:

[0029] Assume that the row of shelves closest to the warehouse access platform is recorded as row 1, the column closest to the warehouse access platform is recorded as column 1, and the bottom layer is recorded as layer 1. The specific modeling process of the principle of the fastest material outbound is as follows:

[0030]

[0031] Among them, f2 is the fastest material outbound index. Indicates that the incoming materials i Number of pallets of materials of this type; The tth i The turnover rate of materials; i and jli They represent the time and distance factors for the fastest outbound delivery of the i-th material pallet, respectively, and are calculated as follows:

[0032]

[0033] Among them, L x and L y They represent the length of the shelf unit in the row and column directions respectively; L0 is the distance between two adjacent rows of shelves; y i represents the column coordinates of the i-th material; v x 、v y 、v z They respectively represent the running speed of the pallet transport device in the direction of shelf row, column and layer when transporting material pallets.

[0034] Furthermore, the objective function is established based on the material aggregation principle to calculate the material aggregation index, including:

[0035] Material cohesion is modeled as follows:

[0036]

[0037] Among them, nj represents the material cohesion index, N type Indicates the number of types of materials to be stored, e represents a natural constant, k is the material type index, N k Indicates the number of pallets of the kth type of materials in the warehouse; x k ,y k and z k Respectively represent the row, column and layer coordinates of the material pallet of the k-th type of material;

[0038] The material coupling model is as follows:

[0039]

[0040] Among them, oh represents the material coupling index, and Respectively represent the t i The row, column and layer coordinates of the material pallets of the same type of materials;

[0041] The material aggregation principle is modeled based on material cohesion and coupling as follows:

[0042] f3=w nj ·nj+w oh ·oh

[0043]

[0044] Among them, f3 is the material aggregation index, w nj and woh They represent the weight coefficients of material cohesion and coupling respectively.

[0045] Furthermore, the objective function is established based on the material relevance principle to calculate the material relevance index, including:

[0046] The specific process of establishing the objective function based on the principle of material relevance is as follows:

[0047]

[0048] Among them, f4 represents the material relevance index, k and l are material type indexes; C kl and D kl Respectively represent the material correlation matrix and the material storage area center distance matrix:

[0049]

[0050] Among them, c kl represents the material correlation between the kth category material and the lth category material, c k is the sum of the material correlations between the k-th material and all other types of materials, k,l=1,2,...,N type ;

[0051]

[0052] Among them, d kl represents the distance between the pallets of the kth and lth categories of materials, d k It is the sum of the distances between the kth type of materials and the material pallets of all other types of materials.

[0053] Furthermore, based on the shelf local stability index, the fastest material outbound index, the material aggregation index and the material relevance index, a multi-objective cargo location optimization function is constructed, and the multi-objective cargo location optimization function is converted into a single-objective optimization problem, including:

[0054] The multi-objective cargo location optimization function is expressed as follows:

[0055]

[0056] Converted into a single objective optimization problem:

[0057] minf=w1·minf1+w2·minf2+w3·minf3+w4·minf4

[0058]

[0059] Where w1, w2, w3 and w4 represent the multi-objective optimization weight parameters.

[0060] Furthermore, the use of an adaptive genetic algorithm to solve the single-objective optimization problem to obtain an optimal cargo space allocation solution includes:

[0061] Step 7.1, initialization of genetic algorithm parameters;

[0062] Step 7.2, randomly generate N pop The initial population of individuals, each of which represents a possible optimization solution for cargo space; the encoding length of each individual is 3·N tray , where the first N tray The coding bits represent the x-axis coordinates of all incoming pallets, and the middle N tray The coded bits represent the y-axis coordinates of all incoming pallets. tray Each encoding bit represents the z-axis coordinate of all incoming pallets;

[0063] Step 7.3, fitness evaluation; the specific fitness function is as follows:

[0064]

[0065] Among them, F represents the fitness value;

[0066] Step 7.4, using the roulette wheel selection method to perform the selection operation;

[0067] Step 7.5, using a sequential crossover method to perform a crossover operation, wherein the crossover probability is adaptively adjusted;

[0068] Step 7.6, use the exchange mutation method to perform mutation operation;

[0069] Step 7.7, update the current population by generating offspring individuals through selection, crossover, and mutation;

[0070] Step 7.8, repeat steps 7.4 to 7.7 until the maximum number of iterations is reached. The individual with the highest fitness in the final population is the final cargo space allocation plan.

[0071] Furthermore, during the crossover operation, the crossover probability p c The adaptive adjustment method is as follows:

[0072]

[0073] Among them, f c Represents the fitness of the new individual produced by the crossover operation; fmean represents the average fitness of all individuals in the current population, f max It represents the maximum value of individual fitness, and A is the adaptive constant.

[0074] Furthermore, during the mutation operation, the mutation probability pm The adaptive adjustment method is as follows:

[0075]

[0076] Among them, f m Represents the fitness of the new individual produced by the mutation operation.

[0077] Compared with the prior art, the present invention has the following technical features:

[0078] The method of the present invention, based on the three-dimensional digital modeling of stereoscopic shelves, comprehensively considers four target mathematical models, namely, the principle of material center of gravity stability, the principle of fastest material entry / exit, the principle of aggregation of similar materials, and the principle of correlation of heterogeneous materials, and solves the established mathematical model with the help of an adaptive genetic algorithm to obtain the optimal cargo location configuration plan for the current materials. The method can effectively optimize the cargo location distribution of incoming materials, thereby greatly improving the storage efficiency and operation efficiency of the automated stereoscopic warehouse and reducing the operation and maintenance cost of the warehouse. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Figure 1 Schematic diagram of the overall workflow of the method of the present invention;

[0080] Figure 2 Schematic diagram of warehouse shelf layout after 3D digitization;

[0081] Figure 3 Schematic diagram of random initialization distribution of storage locations for incoming materials in the example;

[0082] Figure 4 A schematic diagram of the distribution of storage locations of incoming materials in an example of the inventive method. DETAILED DESCRIPTION

[0083] Referring to the attached drawings, the present invention provides a multi-objective stereoscopic warehouse cargo space allocation optimization method based on an adaptive genetic algorithm. The method comprehensively considers the mathematical model of the four principles of material center of gravity stability, fastest material delivery, aggregation of similar materials, and relevance of heterogeneous materials on the basis of three-dimensional modeling of stereoscopic shelves, and solves the established mathematical model with the help of an adaptive genetic algorithm. The specific steps are as follows:

[0084] Step 1, perform three-dimensional digital modeling of the warehouse and the shelves therein; take the pallet exit as the origin of the three-dimensional coordinate system, take the center of the material pallet to represent the position of the material pallet in the three-dimensional coordinate system, and initialize the three-dimensional coordinates of each material pallet; wherein the x-axis of the three-dimensional coordinate system is the row direction of the shelf, the y-axis is the column direction of the shelf, and the z-axis is the layer direction of the shelf.

[0085] like Figure 1As shown in the figure, it is assumed that there are a rows of shelves in the warehouse, and each row of shelves has b columns and c layers; the warehouse is equipped with an access platform, which has a pallet outbound port and multiple pallet inbound ports, and the materials are all assembled on the material pallets; the coordinate (0,0,0) represents the location of the pallet outbound port; the three-dimensional coordinates of the center of the i-th material pallet in the warehouse cargo space are (x i ,y i ,z i )(x i =1,2,...,a;y i =1,2,...,b;z i =1,2,...,c), all material pallets will be randomly assigned an initial storage location before entering the warehouse, and the three-dimensional coordinates of the initial storage location of the pallet to be entered will be digitally initialized.

[0086] Step 2: Construct an objective function based on the material stability principle to calculate the shelf stability index.

[0087] Material stability principle: In order to ensure the stability of materials and lower the center of gravity of materials, generally heavy materials are stored on lower shelves, and light materials are placed on higher shelves, taking into account the stability of local single-row shelves and the problems of global shelves.

[0088] The local stability of the shelf is modeled as follows:

[0089]

[0090] Among them, h local Represents the local stability index of the shelf, N tray represents the number of pallets to be put into storage; X represents the coordinate value set of the material pallet, L z is the length of the shelf unit in the layer direction; x i and z i Respectively represent the row coordinate and layer coordinate of the i-th material pallet; t i Indicates the type of materials on the i-th material pallet; For the tth i The unit mass of the material; x represents the specific value of the current row coordinate when traversing the shelves by row; x i = x means that during the traversal process, when the row coordinate x of the i-th material tray i When equal to x.

[0091] The global stability of the shelf is modeled as follows:

[0092]

[0093] Among them, h global represents the global stability index of the shelf;

[0094] Considering the local stability and global stability of the shelf, the shelf stability index is expressed as follows:

[0095] f1=w local ·h local +w global ·h global

[0096]

[0097] Among them, f1 represents the shelf stability index, w local and w global They represent the weight coefficients of local stability and global stability of the shelf respectively.

[0098] Step 3: Establish an objective function based on the principle of fastest material delivery, which is used to calculate the fastest material delivery index.

[0099] In order to shorten the distance of materials with high frequency of shipment, materials with high frequency of shipment are generally placed close to the pallet shipment port (or a storage location that is convenient for shipment), and materials with low frequency of shipment are stored deep in the warehouse (or a storage location with more complicated shipment procedures).

[0100] Assume that the row of shelves closest to the warehouse access platform is recorded as row 1, the column closest to the warehouse access platform is recorded as column 1, and the bottom layer is recorded as layer 1. The specific modeling process of the principle of the fastest material outbound is as follows:

[0101]

[0102] Among them, f2 is the fastest material outbound index. Indicates that the incoming materials i Number of pallets of materials of this type; The tth i The turnover rate of materials; i and jl i They represent the time and distance factors for the fastest outbound delivery of the i-th material pallet, respectively, and are calculated as follows:

[0103]

[0104] Among them, L x and L y They represent the length of the shelf unit in the row and column directions respectively; L0 is the distance between two adjacent rows of shelves; y i represents the column coordinates of the i-th material; v x 、v y 、v z They respectively represent the running speed of the pallet transport device in the direction of shelf row, column and layer when transporting material pallets.

[0105] Step 4: Establish an objective function based on the material aggregation principle to calculate the material aggregation index.

[0106] Material aggregation principle: In order to facilitate the inventory of warehouse managers, similar materials are generally stored in the same area. This principle comprehensively considers the high intrinsic nature of similar materials and the coupling of heterogeneous materials.

[0107] Material cohesion is modeled as follows:

[0108]

[0109] Among them, nj represents the material cohesion index, N type Indicates the number of types of materials to be stored, e represents a natural constant, k is the material type index, N k Indicates the number of pallets of the kth type of materials in the warehouse; x k ,y k and z k They represent the row, column and layer coordinates of the material pallet of the kth type of material, respectively, and are calculated as follows:

[0110]

[0111] The material coupling model is as follows:

[0112]

[0113] Among them, oh represents the material coupling index, and Respectively represent the t i The row, column and layer coordinates of the material pallets of each type of material and the calculation method of the center coordinates of each type of material refer to formula (9).

[0114] The material aggregation principle is modeled based on material cohesion and coupling as follows:

[0115]

[0116] Among them, f3 is the material aggregation index, w nj and w oh They represent the weight coefficients of material cohesion and coupling respectively.

[0117] Step 5: Establish an objective function based on the material relevance principle to calculate the material relevance index.

[0118] Correlation principle: In order to facilitate the picking of transfer orders, the storage areas of two different materials with relatively high correlation are usually close to each other; the material correlation refers to the correlation between two types of materials in a certain process. For example, for material delivery, if the efficiency of two types of materials is high when delivered together, it can be considered that the two materials have a high correlation; the material correlation is a preset value, and the material correlation matrix can be obtained by calculation before allocation optimization, as shown in Table 2 in the embodiment.

[0119] The specific process of establishing the objective function based on the principle of material relevance is as follows:

[0120]

[0121] Among them, f4 represents the material relevance index, l is the material type index; C kl and D kl They represent the material correlation matrix and the material storage area center distance matrix respectively. The specific forms are as follows:

[0122]

[0123] Among them, c kl represents the material correlation between the kth category material and the lth category material, c k is the sum of the material correlations between the k-th material and all other types of materials, k,l=1,2,...,N type For example, in formula 13, c 12 is the material correlation between Class 1 materials and Class 2 materials, and c1 is sum.

[0124]

[0125] Among them, d kl represents the distance between the pallets of the kth and lth categories of materials, d k It is the sum of the distances between the kth type of materials and the material pallets of all other types of materials.

[0126] The distance between the center coordinates of any two types of materials (assuming they are the kth type of materials and the lth type of materials) is calculated as follows:

[0127]

[0128] Among them, x l ,y l and z l They represent the row, column and layer coordinates of the center of the material tray of the lth type of material respectively. The calculation method refers to formula (9). The cumulative sum of the distances from the center of a certain type of material (assuming it is the kth type of material) to the centers of other types of materials is calculated as follows:

[0129]

[0130] Step 6: Based on the shelf local stability index, the fastest material outbound index, the material aggregation index and the material relevance index, a multi-objective cargo location optimization function is constructed, and the multi-objective cargo location optimization function is converted into a single-objective optimization problem.

[0131] According to the analysis of the above different principles, four objective functions are established. Combining these four objective functions into a whole can form the mathematical model of the main research problem. It can also be seen that it is a multi-objective optimization problem. The specific mathematical model is as follows:

[0132]

[0133] It is a multi-objective function optimization problem, which is converted into a single-objective optimization problem:

[0134] minf=w1·minf1+w2·minf2+w3·minf3+w4·minf4

[0135]

[0136] Where w1, w2, w3 and w4 represent the multi-objective optimization weight parameters.

[0137] Step 7, using an adaptive genetic algorithm to solve the single-objective optimization problem and obtain the optimal cargo space allocation solution.

[0138] Step 7.1, parameter initialization

[0139] Set the initial values, including warehouse shelf parameters (a, b, c, L x , L y , L z 、v x 、v y 、v z ), material attribute parameters Material correlation matrix C, cargo space allocation model weight parameter (w local 、w global , w1, w2, w3, w4), genetic algorithm parameters (initial population size N pop , maximum crossover probability Minimum crossover probability Maximum mutation probability Minimum mutation probability The maximum number of iterations N of the algorithm iter )wait.

[0140] Step 7.2, Coding Design

[0141] Randomly generate N popThe initial population of individuals, each of which represents a possible optimization solution for cargo space; the encoding length of each individual is 3 ·N tray , first N tray The encoding bits represent the x-coordinates of all incoming pallets, and the middle N tray The coded bits represent the y coordinates of all incoming pallets, and the last N tray Each encoding bit represents the z coordinate of all incoming pallets.

[0142] Step 7.3, fitness evaluation

[0143] The larger the fitness value, the greater the probability of being inherited to the next generation, and the smaller the objective function value represented. The specific fitness function is as follows:

[0144]

[0145] Among them, F represents the fitness value.

[0146] Step 7.4, Select an action

[0147] The roulette wheel selection method is used to calculate Select individuals with high probability, where N pop Represents the number of populations in each generation; every two individuals form a group, one as the parent and the other as the mother for replication; where f(i) represents the fitness value of the i-th individual in the population.

[0148] Step 7.5, crossover operation

[0149] The crossover operation is performed using the sequential crossover method. The main process is as follows: let a certain gene position of the paternal and maternal chromosomes selected according to the crossover probability be exchanged to produce two different new individuals. Taking the chromosomes composed of the cargo position as an example, first randomly select two different chromosomes according to the crossover probability, that is, select two different cargo positions.

[0150] According to the sigmoid function, the crossover probability (p c ), the adaptive adjustment method is as follows:

[0151]

[0152] Among them, e is a natural constant, f c represents the fitness of the new individual produced by the crossover operation; f mean represents the average fitness of all individuals in the current population, f max It represents the maximum value of individual fitness, A is the adaptive constant, and its value is 9.903.

[0153] Step 7.6, mutation operation

[0154] The mutation operation is performed by using the exchange mutation method. There is genetic information on each chromosome, that is, the row, column, layer and other information of a cargo position.

[0155] Design the cargo location optimization adjustment problem mutation probability (p m ), the adaptive adjustment method is as follows:

[0156]

[0157] Among them, f m Represents the fitness of the new individual produced by the mutation operation.

[0158] Step 7.7, determine the new population

[0159] The current population is updated by generating offspring individuals through selection, crossover and mutation. The first ξ proportion individuals with the highest fitness in the current population are the next generation population, where ξ is the elite proportion parameter.

[0160] Step 7.8: Output the cargo space allocation plan

[0161] Repeat steps 7.4 to 7.7 until the maximum number of iterations N is reached. iter , the individual with the highest fitness in the final population is the final cargo space allocation plan.

[0162] Example:

[0163] The feasibility of the multi-objective optimization method for automated stereoscopic warehouse cargo space allocation based on an adaptive genetic algorithm provided by the present invention is verified and explained in conjunction with specific examples.

[0164] Assume that a company has 30 rows, 4 columns and 3 layers of automated warehouse shelves. The length, width and height of each cargo space are 1 meter, and the distance between two rows of shelves is 1.5 meters. The speeds of the warehouse's material access equipment in the x, y and z directions are 1 meter per second, 1 meter per second and 0.5 meters per second respectively. The current shelf status of the warehouse is empty, and there is a batch of materials containing 5 categories and a total of 30 pallets waiting to be put into storage. In order to verify the effectiveness of the invented method, firstly, the cargo spaces are randomly assigned to the 30 pallets of materials to be put into storage, and then the invented method is used to optimize the initial cargo spaces of the materials to be put into storage, and the minimum value of the objective function and the allocation coordinates of the cargo spaces under the minimum value are obtained. The attribute matrix of the materials to be put into storage is shown in Table 1, and the correlation matrix between the materials is shown in Table 2.

[0165] Table 1 Attribute matrix of incoming materials

[0166]

[0167]

[0168]

[0169] Table 2 Correlation matrix of materials

[0170]

[0171] Step 1: Initialize the model parameters according to the relevant data of the instance shelf, including warehouse shelf parameters a=30, b=4, c=3, L x =1, L y =1, L z =1, L0 = 1.5, v x =1, v y =1, v z =0.5, material attribute parameter N tray =30, N type =5, N 第1类物资 =5, N 第2类物资 =6, N 第3类物资 =6, N 第4类物资 =5, N 第5类物资 =8, t 1～5 =1, t 6～11 =2, t 12～17 =3, t 18～22 =4, t 23～30 =5, p 第1类物资 =0.42, p 第2类物资 =0.88, p 第3类物资 =0.34, p 第4类物资 =0.16, p 第5类物资 =0.66, the value of the material correlation matrix C is shown in Table 1, and the weight parameter w of the cargo space allocation model local =0.5, w global =0.5, w nj =0.5, w oh =0.5, w1=0.4, w2=0.3, w3=0.2, w4=0.1, genetic algorithm parameter N pop =1000, ξ=0.2、N iter =10000.

[0172] Step 2: Randomly generate 1000 initial individual populations, each of which represents a possible cargo location optimization solution. The encoding length of each individual is 90, the first 30 encoding bits represent the x-coordinates of all incoming pallets, the 31st to 60th encoding bits represent the y-coordinates of all incoming pallets, and the last 61st to 90th encoding bits represent the z-coordinates of all incoming pallets.

[0173] Step 3: Use formula (17) to calculate the fitness of all first-generation individuals and use Select the top 200 with the largest fitness (N pop ·ξ) individuals as the initial population for the next generation.

[0174] Step 4: First, calculate the crossover probability (p) of the current round according to formula (19). c ), then let p c By swapping a certain gene position of the selected paternal and maternal chromosomes, two different new individuals can be generated. Taking the chromosome composed of the cargo position as an example, first randomly select two different chromosomes according to the crossover probability, that is, select two different cargo positions. The two gene positions of cargo position 3 (2, 5, 4) and cargo position 4 (1, 2, 4) are randomly swapped in the same position. After crossover, two different new individuals (2, 2, 4) and (1, 5, 4) can be generated.

[0175] Step 5: First, calculate the current round mutation probability (p m ), and then use the exchange mutation method to perform mutation operations. There is genetic information on each chromosome, that is, the row, column, layer, etc. of a cargo position. For example, according to the mutation probability of the cargo position, cargo position 3 (1, 2, 2) is randomly selected, and the selected gene mutation position is the chromosome "1". Exchanging "2" with it will produce (2, 2, 1), and a new individual will be generated.

[0176] Step 6: First, update the current population by selecting, crossover, and mutation to generate offspring individuals, and then execute step 3.

[0177] Step 7, repeat steps 3 to 6 until the maximum number of iterations is 10,000. The individual with the highest fitness in the final population is the final cargo space allocation solution. The effects before and after cargo space allocation optimization can be found in Figure 3 and Figure 4 .

[0178] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A multi-objective warehouse location allocation optimization method based on adaptive genetic algorithm, characterized in that: include: Conduct 3D digital modeling of the warehouse and the shelves in it; The pallet outbound port is used as the origin of the three-dimensional coordinate system, and the center of the material pallet represents the position of the material pallet in the three-dimensional coordinate system, and the three-dimensional coordinates of each material pallet are initialized; The objective function is constructed based on the material stability principle to calculate the shelf stability index; The objective function is established based on the principle of the fastest material outbound delivery, which is used to calculate the fastest material outbound delivery index; The objective function is established based on the material aggregation principle to calculate the material aggregation index; The objective function is established based on the material relevance principle to calculate the material relevance index; Based on the shelf local stability index, the fastest material outbound index, the material aggregation index and the material relevance index, a multi-objective cargo location optimization function is constructed, and the multi-objective cargo location optimization function is converted into a single-objective optimization problem; The single-objective optimization problem is solved using an adaptive genetic algorithm to obtain the optimal cargo space allocation solution.

2. The multi-objective warehouse cargo space allocation optimization method based on adaptive genetic algorithm according to claim 1 is characterized in that: Assume that there are a rows of shelves in the warehouse, and each row of shelves has b columns and c layers; the warehouse is equipped with an access platform, which has a pallet outbound port and multiple pallet inbound ports, and the materials are all assembled on the material pallets; the coordinate (0,0,0) represents the location of the pallet outbound port; the three-dimensional coordinates of the center of the i-th material pallet in the warehouse cargo space are (x i ,y i ,z i ), x i =1,2,...,a;y i =1,2,...,b;z i =1,2,...,c; all material pallets will be randomly assigned an initial storage location before entering the warehouse, and the three-dimensional coordinates of the initial storage location of the pallet to be entered will be digitally initialized.

3. The multi-objective warehouse cargo space allocation optimization method based on adaptive genetic algorithm according to claim 1 is characterized in that: The objective function constructed based on the material stability principle is used to calculate the shelf stability index, including: The local stability of the shelf is modeled as follows: Among them, h local Represents the local stability index of the shelf, N tray represents the number of pallets to be put into storage; X represents the coordinate value set of the material pallet, L z is the length of the shelf unit in the layer direction; x i and z i Respectively represent the row coordinate and layer coordinate of the i-th material pallet; t i Indicates the type of materials on the i-th material pallet; For the tth i Unit mass of such materials; The global stability of the shelf is modeled as follows: Among them, h global represents the global stability index of the shelf; Considering the local stability and global stability of the shelf, the shelf stability index is expressed as follows: Among them, f1 represents the shelf stability index, w local and w global They represent the weight coefficients of local stability and global stability of the shelf respectively.

4. The multi-objective warehouse cargo space allocation optimization method based on adaptive genetic algorithm according to claim 1 is characterized in that: The objective function is established based on the principle of the fastest material outbound delivery, and is used to calculate the fastest material outbound delivery index, including: Assume that the row of shelves closest to the warehouse access platform is recorded as row 1, the column closest to the warehouse access platform is recorded as column 1, and the bottom layer is recorded as layer 1. The specific modeling process of the principle of the fastest material outbound is as follows: Among them, f2 is the fastest material outbound index. Indicates that the incoming materials i Number of pallets of materials of this type; The tth i The turnover rate of materials; i and jl i They represent the time and distance factors for the fastest outbound delivery of the i-th material pallet, respectively, and are calculated as follows: Among them, L x and L y They represent the length of the shelf unit in the row and column directions respectively; L0 is the distance between two adjacent rows of shelves; y i represents the column coordinates of the i-th material; v x 、v y 、v z They respectively represent the running speed of the pallet transport device in the direction of shelf row, column and layer when transporting material pallets.

5. The multi-objective warehouse cargo space allocation optimization method based on adaptive genetic algorithm according to claim 1 is characterized in that: The objective function established based on the material aggregation principle is used to calculate the material aggregation index, including: Material cohesion is modeled as follows: Among them, nj represents the material cohesion index, N type Indicates the number of types of materials to be stored, e represents a natural constant, k is the material type index, N k Indicates the number of pallets of the kth type of materials in the warehouse; x k ,y k and z k Respectively represent the row, column and layer coordinates of the material pallet of the k-th type of material; The material coupling model is as follows: Among them, oh represents the material coupling index, and Respectively represent the t i The row, column and layer coordinates of the material pallets of the same type of materials; The material aggregation principle is modeled based on material cohesion and coupling as follows: Among them, f3 is the material aggregation index, w nj and w oh They represent the weight coefficients of material cohesion and coupling respectively.

6. The multi-objective warehouse cargo space allocation optimization method based on adaptive genetic algorithm according to claim 1 is characterized in that: The objective function is established based on the material relevance principle to calculate the material relevance index, including: The specific process of establishing the objective function based on the principle of material relevance is as follows: Among them, f4 represents the material relevance index, k and l are material type indexes; C kl and D kl Respectively represent the material correlation matrix and the material storage area center distance matrix: Among them, c kl represents the material correlation between the kth category material and the lth category material, c k is the sum of the material correlations between the k-th material and all other types of materials, k,l=1,2,...,N type ; Among them, d kl represents the distance between the pallets of the kth and lth categories of materials, d k It is the sum of the distances between the kth type of materials and the material pallets of all other types of materials.

7. The multi-objective warehouse cargo space allocation optimization method based on adaptive genetic algorithm according to claim 1 is characterized in that: The multi-objective cargo location optimization function is constructed based on the shelf local stability index, the fastest material outbound index, the material aggregation index and the material relevance index, and the multi-objective cargo location optimization function is converted into a single-objective optimization problem, including: The multi-objective cargo location optimization function is expressed as follows: Converted into a single objective optimization problem: Where w1, w2, w3 and w4 represent the multi-objective optimization weight parameters.

8. The multi-objective warehouse cargo space allocation optimization method based on adaptive genetic algorithm according to claim 1 is characterized in that: The method of using an adaptive genetic algorithm to solve the single-objective optimization problem and obtain an optimal cargo space allocation solution includes: Step 7.1, initialization of genetic algorithm parameters; Step 7.2, randomly generate N pop The initial population of individuals, each of which represents a possible optimization solution for cargo space; the encoding length of each individual is 3·N tray , where the first N tray The coding bits represent the x-axis coordinates of all incoming pallets, and the middle N tray The coded bits represent the y-axis coordinates of all incoming pallets. tray Each encoding bit represents the z-axis coordinate of all incoming pallets; Step 7.3, fitness evaluation; the specific fitness function is as follows: Among them, F represents the fitness value; Step 7.4, using the roulette wheel selection method to perform the selection operation; Step 7.5, using a sequential crossover method to perform a crossover operation, wherein the crossover probability is adaptively adjusted; Step 7.6, use the exchange mutation method to perform mutation operation; Step 7.7, update the current population by generating offspring individuals through selection, crossover, and mutation; Step 7.8, repeat steps 7.4 to 7.7 until the maximum number of iterations is reached. The individual with the highest fitness in the final population is the final cargo space allocation plan.

9. The multi-objective warehouse cargo space allocation optimization method based on adaptive genetic algorithm according to claim 8 is characterized in that: During the crossover operation, the crossover probability p c The adaptive adjustment method is as follows: Among them, f c represents the fitness of the new individual produced by the crossover operation; f mean represents the average fitness of all individuals in the current population, f max It represents the maximum value of individual fitness, and A is the adaptive constant.

10. The multi-objective warehouse cargo space allocation optimization method based on adaptive genetic algorithm according to claim 8, characterized in that: During the mutation operation, the mutation probability p m The adaptive adjustment method is as follows: Among them, f m Represents the fitness of the new individual produced by the mutation operation.

Images