A warehouse location allocation optimization method based on crayfish optimization algorithm

Abstract

A kind of based on crayfish optimization algorithm's warehouse location allocation optimization method, contain steps: S1, obtain stereoscopic warehouse information and cargo information;S2, establish multi-objective warehouse location allocation mathematical model;Respectively establish single-objective warehouse location allocation mathematical model;Through linear weighting method, single-objective warehouse location allocation mathematical model is converted into multi-objective warehouse location allocation mathematical model;S3, calculate optimal warehouse location allocation result;Multi-objective warehouse location allocation mathematical model is solved by crayfish optimization algorithm, and optimal warehouse location allocation result is obtained.The present application can more efficiently find the relatively optimal warehouse location allocation scheme under multi-objective constraint.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent logistics, and in particular to a warehouse location allocation optimization method based on crayfish optimization algorithm. BACKGROUND

[0002] The stacker stereoscopic warehouse is an automated warehousing system that realizes the storage and management of goods through stackers. The hydraulic cylinder, as a heavy and usually large industrial component, is one of the items that need to be stored and managed in the stacker stereoscopic warehouse. In the application scenario of intelligent logistics, the stacker stereoscopic warehouse stores and manages large mechanical transmission elements such as hydraulic cylinders through vertical storage and automated access. Reasonable warehouse location allocation of the stacker stereoscopic warehouse not only improves system efficiency, but also reduces warehousing costs and shortens the access time of goods, which is an important link in warehousing logistics.

[0003] The existing patent with application number 202210611361.4 discloses an automatic stereoscopic warehouse location allocation optimization method based on genetic algorithm. First, the in-out warehouse frequency of goods is obtained according to the historical order information of the parts, and a mathematical model for reducing the in-out warehouse handling time is established. Then, the mass of the parts is counted, and a calculation formula for reducing the center of gravity of the shelves is established. On this basis, clustering is performed based on the correlation degree, and a mathematical model is established to improve the spatial aggregation degree of related products. The invention optimizes and solves the comprehensive mathematical model through an improved genetic algorithm, and the optimal solution is the current optimal location allocation scheme.

[0004] The warehouse location allocation optimization method mainly focuses on the setting of optimization objectives and the selection of heuristic algorithms. The optimization objectives are usually focused on improving the in-out warehouse efficiency of the stereoscopic warehouse and enhancing the stability of the shelves. The defect of the existing technology is that the existing warehouse location allocation optimization method considers a single and incomplete optimization objective for the stereoscopic warehouse of special items such as hydraulic cylinders with large size and heavy weight. In terms of algorithm selection, although genetic algorithms are widely used, there is still room for improvement in search capability and computational efficiency, which cannot meet the efficient demand of goods storage and management in the stereoscopic warehouse. SUMMARY

[0005] In view of the above defects of the prior art, the present application provides a warehouse location allocation optimization method based on crayfish optimization algorithm, which can more efficiently find the relatively optimal warehouse location allocation scheme under multi-objective constraints.

[0006] To achieve the above purpose, the technical solution adopted by the present application is:

[0007] A warehouse location allocation optimization method based on crayfish optimization algorithm, comprising the following steps:

[0008] S1, obtaining stereoscopic warehouse information and goods information;

[0009] S2, a multi-objective storage location allocation mathematical model is established;

[0010] A single-objective storage location allocation mathematical model is respectively established; the single-objective storage location allocation mathematical model is converted into the multi-objective storage location allocation mathematical model through a linear weighting method;

[0011] S3, an optimal storage location allocation result is calculated;

[0012] The multi-objective storage location allocation mathematical model is solved through a crayfish optimization algorithm, and the optimal storage location allocation result is obtained.

[0013] In step S1, collecting the information of the stereoscopic warehouse and the goods is the basis for subsequent model establishment and algorithm application. The information of the stereoscopic warehouse includes the size, structure, storage location layout, and storage location quantity of the stereoscopic warehouse, and the information of the goods includes the size, weight, type, and in-out frequency of the goods.

[0014] In step S2, there are multiple single-objective storage location allocation mathematical models, and all the objectives that have an impact on the storage location allocation result are included. If F1, F2, F3, and F4 represent a single-objective storage location allocation mathematical model, respectively, and multiple single-objective storage location allocation mathematical models are linearly weighted, the multi-objective storage location allocation mathematical model is represented as:

[0015] min F = a1xF1 + a2xF2 + a3xF3 + a4xF4

[0016] Wherein, a1, a2, a3, and a4 represent the weights of the single objectives F1, F2, F3, and F4, respectively, and the selection of the weights reflects the relative importance of different single objectives. The distribution of the weights should be determined according to the importance of each objective in the actual problem. The multi-objective function can more comprehensively consider different factors in the optimization problem, because the storage location allocation optimization problem often involves multiple objectives, and there may be conflicts or trade-offs between these objectives. The multi-objective function can reflect this complexity and allow the algorithm to consider multiple objectives simultaneously. At the same time, in many cases, there may be conflicts between different objectives. For example, improving the in-out efficiency may conflict with reducing the stability of the shelf center of gravity. The multi-objective function allows the algorithm to find a balance point between these objectives.

[0017] In step S3, the crayfish optimization algorithm is a meta-heuristic optimization algorithm inspired by the social behavior of crayfish in nature. Crayfish search for food resources within their perception range and move towards areas with higher food concentration, similar to the algorithm searching for better objective function values. The crayfish optimization algorithm is a prior art, but there is no technical solution to use the crayfish optimization algorithm to solve the storage and management of goods in the stereoscopic warehouse.

[0018] As a preferred, the single-target warehouse location allocation mathematical model comprises an in-out warehouse efficiency model F1.

[0019] The in-out warehouse efficiency model F1 is a mathematical model for evaluating and optimizing the efficiency of goods in-out warehouse operations in a stereoscopic warehouse. The purpose of this model is to reduce the access time of goods in the stereoscopic warehouse and improve the efficiency of stereoscopic warehouse operations. By optimizing this single objective function, we can find the best storage location of goods, thereby improving the in-out warehouse efficiency of the stereoscopic warehouse. The specific expression of the in-out warehouse efficiency model F1 is as follows:

[0020]

[0021] Where m, n, l are the number of warehouse locations in X, Y, Z axes of the stereoscopic warehouse, v x , v y , v z are the speeds of the stacker running in X, Y, Z coordinate axes, l t is the length of the warehouse location, w t is the width of the warehouse location, i, j, k are the coordinates of the warehouse X, Y, Z axes, P ijk represents the in-out warehouse frequency penalty of the goods in the warehouse, represents whether there is an in-out warehouse task, and the in-out warehouse efficiency model F1 is the sum of the product of the running time of all in-out warehouse tasks and the in-out warehouse frequency penalty coefficient.

[0022] As a preferred, the single-target warehouse location allocation mathematical model comprises a low overall center of gravity of the shelf model F2.

[0023] The low overall center of gravity of the shelf model F2 is used to optimize the stability of the shelves in the stereoscopic warehouse, especially for storing heavy or large goods. The core goal is to reduce the overall center of gravity of all goods on the shelf to enhance the stability and safety of the shelf. By reducing the center of gravity, the risk of the shelf overturning due to uneven distribution of goods is reduced, ensuring that the shelf can remain stable when subjected to external forces (such as earthquakes, impacts). The specific expression of the low overall center of gravity of the shelf model F2 is as follows:

[0024]

[0025] Where M ijk is the weight of the goods in the warehouse, h t is the height of the goods in the warehouse, and the sum of the product of all goods weight and its height in the warehouse is calculated, and the overall shelf center of gravity is minimized to improve the stability of the shelf.

[0026] As a preferred, the single-target warehouse location allocation mathematical model comprises a single-sided shelf anti-overturning model F3.

[0027] The single-sided shelf anti-overturning model F3 is used to ensure the stability of the shelf, and the main goal is to reduce the risk of shelf overturning due to uneven weight distribution by reasonably distributing the weight of the goods. The single-sided shelf anti-overturning model F3 reduces the possibility of overturning due to uneven weight distribution by balancing the weight on both sides of the shelf, and by reasonably distributing the goods, it makes full use of the storage space of the shelf while maintaining the stability of the shelf. The specific expression of the single-sided shelf anti-overturning model F3 is as follows:

[0028]

[0029] This function calculates the moment of the force generated by the goods on both sides of the three-dimensional warehouse on the Y-axis relative to the center of the respective shelves, and prevents high-level shelf overturning caused by excessive weight by minimizing this function.

[0030] As a preferred embodiment, the single-objective storage location allocation mathematical model comprises a double-stretch location storage location priority same kind of goods model F4.

[0031] The double-stretch location storage location priority same kind of goods model F4 is used to optimize the storage location allocation of the three-dimensional warehouse, aiming to store the same kind of goods in double-stretch locations that can be accessed from both sides simultaneously. By storing the same kind of goods in double-stretch locations, the access efficiency of the goods can be improved because these goods can be accessed from both sides simultaneously. Reducing the need to move goods due to improper storage location reduces labor intensity and operating costs. This storage method can improve the efficiency of the three-dimensional warehouse operation, reduce the operation of reversing the warehouse, and reduce errors caused by confusion of goods. The specific expression of the double-stretch location storage location priority same kind of goods model F4 is as follows:

[0032]

[0033] where δ is an indicator function that returns 1 when the condition in the parentheses is met, and 0 otherwise, and D is the set of all double-stretch goods. This function compares all the goods on the shelves, and when the types of goods on the two shelves of the same double-stretch location are different, it is penalized, and by reducing the value of this function, the operation of reversing the warehouse is reduced.

[0034] As a preferred embodiment, the step S2 comprises adding a constraint condition to the multi-objective storage location allocation mathematical model.

[0035] The multi-objective storage location allocation mathematical model is obtained by linear weighting of the four single-objective storage location allocation mathematical model functions described above, and the formula is as follows:

[0036]

[0037] minF = α1 × F1 + α2 × F2 + α3 × F3 + α4 × F4

[0038] Where 0 < α1< 1, 0 < α2< 1, 0 < α3< 1, 0 < α4< 1, α1+ α2+ α3+ α4= 1; the constraint condition is:

[0039] A. Each cargo can only be assigned to one location:

[0040]

[0041] B. Each location can only hold one cargo:

[0042]

[0043] C. Stereoscopic warehouse condition restrictions:

[0044]

[0045] The constraint condition ensures the feasibility and practicality of the solution in the optimization problem. Constraint condition A ensures that each cargo can only be assigned to one location, avoiding repeated assignment or omission of cargo. Constraint condition B ensures that each location can only hold one cargo, ensuring that the storage space is not occupied by multiple cargos, and also preventing the storage space from being empty. Constraint condition C ensures compliance with the physical restrictions of the stereoscopic warehouse, ensuring that the assignment of cargos complies with the physical size and structural restrictions of the stereoscopic warehouse.

[0046] As a preferred, in the summer vacation stage of the crayfish optimization algorithm, the logistic curve is defined as the scaling factor C2 of the crayfish optimization algorithm, and the Levy flight behavior is defined as the rand random number step of the crayfish optimization algorithm.

[0047] In the crayfish optimization algorithm, when the environmental temperature is greater than 30 degrees, half of the individuals enter the summer vacation stage, and the individuals will move towards the cave to avoid high temperature in this stage.

[0048] The logistic curve as a scaling factor can allow individuals to explore larger steps towards the optimal point in the early iterations of the algorithm. The logistic curve is an S-shaped curve that grows rapidly at first and then slowly, which makes the algorithm quickly explore the solution space in the early stage and then search in detail to achieve a more refined exploration of the optimal point. The Levy flight as a step replacement has the characteristic of occasional long jumps, which helps the algorithm to jump out of the local optimal solution and enhance the global search ability, thereby improving the probability of finding the global optimal solution. Through the improved scaling factor and step strategy, the algorithm can more effectively balance the conflict between different objectives in multi-objective optimization problems and find better trade-off solutions. The long jump characteristic of the Levy flight significantly improves the global search ability of the algorithm and reduces the risk of the algorithm falling into a local optimal solution. The logistic curve parameter control is flexible, making the convergence process of the algorithm more smooth and natural, which helps the algorithm to approach the optimal solution stably. The movement formula is:

[0049]

[0050] The scaling factor C2 formula is:

[0051]

[0052] Levy(n, m, β) is an n x m-dimensional Levy step matrix with parameter β, and its formula is:

[0053]

[0054] u n×m ~ N(0, σ u )

[0055] v n×m ~ N(0, 1)

[0056] Where:

[0057]

[0058] The improvement of the scaling factor C2 allows individuals in the early iterations of the algorithm to explore larger steps towards the optimal point, accelerating the convergence speed and exploration ability of the algorithm. In the later stage, the step size is gradually reduced to achieve a more refined exploration of the optimal point. The logistic curve parameter control is flexible compared to the linear scaling factor of the basic crayfish optimization algorithm, and the convergence process is more smooth and natural. The occasional long jump characteristic of the Levy flight significantly improves the global search ability of the algorithm and prevents falling into a local optimal solution.

[0059] As a preference, the crayfish optimization algorithm comprises an elite individual probability update phase, selectively moving other individual positions closer to the current optimal individual according to an update probability, which increases over the iteration process.

[0060] The update probability is specifically formulated as:

[0061]

[0062] Where ind is the probability formula:

[0063]

[0064] By gradually increasing the update probability, the algorithm is more inclined to approach the optimal solution in later iterations, which helps the algorithm converge faster. In early iterations, the lower update probability helps maintain the diversity of the population and avoid premature convergence to local optimal solutions. The gradual increase of the update probability allows the algorithm to balance between exploration and exploitation.

[0065] As a preference, the crayfish optimization algorithm comprises a boundary correction phase, in which individuals that exceed the upper and lower bounds are reassigned during the iteration process.

[0066] The crayfish optimization algorithm enters an improved boundary correction phase, which enhances the global search capability compared to the basic crayfish optimization algorithm that directly sets the values of individuals exceeding the bounds to the boundaries. The boundary correction phase ensures that all individuals do not exceed these constraints, thus guaranteeing that they are legal solutions. By reassigning individuals that exceed the boundaries to the vicinity of the boundaries, the boundary correction phase helps maintain the integrity of the search space and prevents individuals from being "lost" during the search process. When individuals exceed the boundaries and are reassigned, they may be placed in different regions of the search space, which helps the algorithm explore new regions and enhance global search capability. If individuals concentrate in a certain region during the search process, it may lead to premature convergence of the algorithm. The boundary correction helps avoid this situation by reassigning individuals.

[0067] Compared with the prior art, the beneficial effects of the present application are reflected in:

[0068] 1.The present application establishes a multi-objective storage location allocation mathematical model, providing an optimized allocation mathematical model for the storage and management of goods in a warehouse. The model considers multiple optimization objectives related to storage location allocation, and through linear weighting, it converts multiple single-objective functions into a comprehensive multi-objective function. This allows for a more comprehensive consideration of various factors, resulting in a relatively optimal storage location allocation scheme. Additionally, the weights of individual objectives can be adjusted flexibly according to actual application conditions to reflect the relative importance of different objectives, enhancing the adaptability and flexibility of the storage location allocation method. Compared to existing genetic algorithms, the use of crayfish optimization algorithm for solving the multi-objective storage location allocation mathematical model provides higher search capability and computational efficiency, enabling faster identification of storage location allocation schemes that meet multi-objective requirements, thereby meeting the efficient needs of warehouse storage and management, reducing warehouse logistics costs, and improving the safety and stability of warehouse operation.

[0069] 2.The present application improves the crayfish optimization algorithm during the summering phase, allowing individuals to perform large-step exploration at the initial stage of the algorithm, which helps quickly cover the solution space and find potential optimal solution regions. As iteration progresses, the growth of the Logistic curve slows down, allowing individuals to search more meticulously in the optimal solution region. The long jump characteristic of Levy flight helps individuals escape local optimal solutions, enhancing the algorithm's global search capability and improving the probability of finding a global optimal solution. This improved strategy allows the algorithm to balance between exploring new regions and delving into known regions, improving search efficiency and effectiveness.

[0070] 3.The present application improves the crayfish optimization algorithm during the elite individual probability update phase by gradually increasing the probability of approaching the optimal individual. This makes the algorithm more inclined to refine the search in later iterations, accelerating the convergence speed. In early iterations, the lower update probability helps maintain the diversity of the population and avoid premature convergence to local optimal solutions. By gradually guiding the population towards the optimal solution, the quality of the final solution can be improved, making the storage location allocation more reasonable and efficient.

[0071] 4.The present application improves the crayfish optimization algorithm during the boundary correction phase, ensuring that all individuals do not exceed the boundaries of the search space, thereby guaranteeing that they are legal solutions. By redistributing individuals that exceed the boundary to the vicinity of the boundary, the integrity of the search space is maintained, preventing individuals from being "lost" during the search process. When individuals exceed the boundary and are redistributed, they may be placed in different regions of the search space, which helps the algorithm explore new regions and enhance global search capability. Boundary correction helps avoid premature convergence by redistributing individuals, thereby avoiding premature convergence. BRIEF DESCRIPTION OF DRAWINGS

[0072] Figure 1 Flow chart of the warehouse position allocation optimization method for the embodiment 1 of the present application;

[0073] Figure 2 Flow chart of the algorithm for the embodiment 1 of the present application;

[0074] Figure 3 Schematic diagram of the randomly generated partial existing warehouse position data for the embodiment 1 of the present application;

[0075] Figure 4 Schematic diagram of the optimal warehouse position allocation result data for the embodiment 1 of the present application;

[0076] Figure 5 Schematic diagram of the warehouse position allocation result for the embodiment 1 of the present application;

[0077] Figure 6 Optimal fitness curve of the algorithm running result for the embodiment of the present application. DETAILED DESCRIPTION

[0078] In order to make the technical means, creative features, purposes and effects of the invention easy to understand, the present application will be further described in conjunction with specific drawings. However, the present application is not limited to the following embodiments.

[0079] It should be noted that the structure, proportion, size, etc. shown in the drawings attached to the present specification are only used to cooperate with the content disclosed in the specification for understanding and reading by those skilled in the art, and do not define the limiting conditions for the implementation of the present application, so they do not have technical substantive significance. Any modification of structure, change of proportion relationship or adjustment of size, without affecting the effects and purposes that can be achieved by the present application, should still fall within the scope of the technical content disclosed by the present application.

[0080] Embodiment 1:

[0081] As shown in a warehouse position allocation optimization method based on crayfish optimization algorithm, comprising the steps of: Figure 1

[0082] S1, obtaining the information of the three-dimensional warehouse and the information of the goods;

[0083] S2, establishing a multi-objective warehouse position allocation mathematical model;

[0084] Respectively establishing a single-objective warehouse position allocation mathematical model; converting the single-objective warehouse position allocation mathematical model into a multi-objective warehouse position allocation mathematical model through linear weighting method;

[0085] S3, calculating the optimal warehouse position allocation result;

[0086] Solving the multi-objective warehouse position allocation mathematical model through crayfish optimization algorithm to obtain the optimal warehouse position allocation result.​

[0087] In step S1, collecting the information of the stereoscopic warehouse and the goods is the basis for subsequent model establishment and algorithm application. The information of the stereoscopic warehouse includes the size, structure, location layout, and number of the stereoscopic warehouse. The information of the goods includes the size, weight, type, and frequency of the goods.

[0088] In step S2, there are multiple single-target location allocation mathematical models, and any target that has an impact on the location allocation result is included. If F1, F2, F3, and F4 represent a single-target location allocation mathematical model, respectively, and multiple single-target location allocation mathematical models are linearly weighted, the multi-target location allocation mathematical model is represented as:

[0089] min F = α1 × F1 + α2 × F2 + α3 × F3 + α4 × F4

[0090] Wherein, α1, α2, α3, and α4 represent the weights of the single targets F1, F2, F3, and F4, respectively. The selection of the weights reflects the relative importance of different single targets, and the distribution of the weights should be determined according to the importance of each target in the actual problem. The multi-target function can more comprehensively consider different factors in the optimization problem. The reason is that the location allocation optimization problem often involves multiple targets, and there may be conflicts or trade-offs between these targets. The multi-target function can reflect this complexity and allow the algorithm to consider multiple targets simultaneously. At the same time, in many cases, there may be conflicts between different targets. For example, improving the efficiency of warehouse access may conflict with reducing the stability of the shelf center of gravity. The multi-target function allows the algorithm to find a balance point between these targets.

[0091] In step S3, the crayfish optimization algorithm is a meta-heuristic optimization algorithm inspired by the social behavior of crayfish in nature. Crayfish search for food resources within their perception range and move towards areas with higher food concentration, similar to the algorithm searching for better target function values. The crayfish optimization algorithm is a prior art, but there is no existing technical solution to use the crayfish optimization algorithm to solve the storage and management of goods in a stereoscopic warehouse.

[0092] The single-target location allocation mathematical model includes the access efficiency model F1, the overall shelf center of gravity model F2, the single-sided shelf anti-overturning model F3, and the double-stretch location priority same goods model F4.

[0093] The access efficiency model F1 is a mathematical model used to evaluate and optimize the efficiency of goods access in a stereoscopic warehouse. The purpose of this model is to reduce the access time of goods in the stereoscopic warehouse and improve the efficiency of stereoscopic warehouse operations. By optimizing this single-target function, we can find the best storage location for goods, thereby improving the access efficiency of the stereoscopic warehouse. The specific expression of the access efficiency model F1 is as follows:

[0094]

[0095] where m, n, l are the number of bin positions in the X, Y, Z axis of the stereoscopic warehouse, v x , v y , v z are the speeds of the stacker running in the X, Y, Z coordinate axis, l t is the length of the bin, w t is the width of the bin, i, j, k are the coordinates of the bin X, Y, Z axis, P ijk represents the frequency of the warehouse goods in and out of the warehouse, and represents whether there is a warehouse task, and the in-out efficiency model F1 is the sum of the product of the running time of all warehouse tasks and the frequency of the goods in and out of the warehouse.

[0096] The overall low center of gravity of the rack model F2 is used to optimize the stability of the stereoscopic warehouse rack, especially for storing heavy or large goods, the core goal of which is to reduce the overall center of gravity of all goods on the rack to enhance the stability and safety of the rack. By reducing the center of gravity, the risk of overturning of the rack due to uneven distribution of goods is reduced, ensuring that the rack can remain stable when subjected to external forces (such as earthquakes, impacts). The specific expression of the overall low center of gravity of the rack model F2 is as follows:

[0097]

[0098] where M ijk is the weight of the goods in the bin, h t is the height of the goods in the bin, and the sum of the product of all goods weight and its height in the bin is calculated, and the overall rack center of gravity is minimized to improve the stability of the rack.

[0099] The single-sided rack anti-overturning model F3 is used to ensure the stability of the rack, which mainly aims to reduce the risk of overturning of the rack due to uneven distribution of weight by reasonably distributing the weight of the goods. The single-sided rack anti-overturning model F3 balances the weight on both sides of the rack to reduce the possibility of overturning due to uneven weight distribution, and reasonably distributes the goods to fully utilize the storage space of the rack while maintaining the stability of the rack. The specific expression of the single-sided rack anti-overturning model F3 is as follows:

[0100]

[0101] This function respectively calculates the moment of the moment generated by the goods on both sides of the rack in the Y axis relative to the center of the respective rack, and minimizes the function to prevent high-level rack overturning caused by excessive weight.

[0102] The double-stretch location priority same kind of goods model F4 is used for optimizing the location allocation of the stereoscopic warehouse, and aims to store the same kind of goods in the double-stretch location which can be accessed from both sides. By storing the same kind of goods in the double-stretch location, the access efficiency of the goods can be improved because the goods can be accessed from both sides at the same time. The operation of moving the goods due to improper storage position is reduced, thereby reducing the labor intensity and operation cost. The storage mode can improve the operation efficiency of the stereoscopic warehouse, reduce the operation of reversing the warehouse, and reduce the error caused by the confusion of the goods. The specific expression of the double-stretch location priority same kind of goods model F4 is as follows:

[0103]

[0104] Wherein, δ is an indicator function, which returns 1 when the condition in the bracket is met, and returns 0 otherwise, D is the set of all double-stretch goods, and the function compares the goods on all shelves. When the goods on the two shelves of the same double-stretch location are different, a penalty is imposed, and the value of the function is reduced to reduce the operation of reversing the warehouse.

[0105] The multi-objective location allocation mathematical model is obtained by linear weighting of the above-mentioned four single-objective location allocation mathematical model functions, and the formula is as follows:

[0106]

[0107] minF = α1×F1+α2×F2+α3×F3+α4×F4

[0108] Wherein 0<α1<1, 0<α2<1, 0<α3<1, 0<α4<1, α1+α2+α3+α4=1; the constraint condition of the multi-objective location allocation mathematical model is as follows:

[0109] A. Each goods can only be allocated to one location:

[0110]

[0111] B. Each location can only store one goods:

[0112]

[0113] C. Stereoscopic warehouse condition restrictions:

[0114]

[0115] The constraints in the optimization problem ensure the feasibility and practicality of the solution. Constraint A ensures that each item can only be assigned to one location, avoiding duplication or omission of assignments. Constraint B ensures that each location can only hold one item, ensuring that the location is not occupied by multiple items, and also preventing the location from being empty. Constraint C ensures compliance with the physical constraints of the AS / RS, and the AS / RS condition constraints ensure that the assignment of items complies with the physical size and structural constraints of the AS / RS.

[0116] As shown in Figure 2 , the crayfish optimization algorithm in step S3 includes the following steps:

[0117] S3.1, initialization of parameters: set the number of iterations of the algorithm, the population size, the size of the AS / RS, the size of the location, the speed of the stacker in each direction, etc.

[0118] S3.2, randomly generate some existing items occupying locations to simulate the conditions when the real warehouse task is performed, and the generated part of the existing locations are as shown in Figure 3 .

[0119] S3.3, adopt real number coding to take the ratio of the location index of the to-be-warehoused task to all locations as the individual value, and the number of to-be-warehoused tasks as the individual dimension. After randomly generating the initial population, calculate the initial fitness value and record the population optimal individual and its fitness value, and assign the same value to the global optimal individual and its fitness value.

[0120] S3.4, start iteration, first calculate and according to the formula, take their mean as the position of the cave, and generate the environment temperature temp to determine the subsequent individual behavior.

[0121] S3.5, when the environment temperature is greater than 30 degrees, half of the individuals enter the summering stage, and the individuals will move towards the cave to avoid high temperature. Here, improvements are made to the crayfish optimization algorithm, mainly replacing the scaling factor C2 with a logistic curve and replacing the rand random number step with a Levy flight, and the moving formula is:

[0122]

[0123] The C2 formula is:

[0124]

[0125] Levy(n, m, β) is an n x m-dimensional Levy step matrix with parameter β, and its formula is:

[0126]

[0127] u n×m ~ N(0, σ u )

[0128] v n×m ~N(0, 1)

[0129] wherein:

[0130]

[0131] The improvement of the scaling factor C2 makes the individuals in the initial stage of algorithm iteration enter the exploration stage with a larger step size towards the optimal point, which accelerates the convergence speed and the exploration ability of the algorithm, while gradually reducing the step size in the later stage to realize more detailed exploration of the optimal point. The logistic curve parameter control of the linear scaling factor of the basic COA algorithm is flexible, and the convergence process is more smooth and natural. The occasional long jump of the Levy flight significantly improves the global search ability of the algorithm and prevents falling into local optimal solution.

[0132] S3.6, when the temperature is higher than 30 degrees, the other half of the individuals enter the competition mode. The difference between the current individual and another random individual is used as the disturbance and the position X of the cave shade The addition simulates the competition of crayfish individuals for the same cave, increases the variation of individuals, and prevents falling into local optimal solution. The formula is:

[0133]

[0134] S3.7, when the temperature is lower than 30 degrees, the individual enters the feeding stage. First, the food size is defined according to formula 14.

[0135]

[0136] The food Q measures the size of the current individual's fitness relative to the optimal individual's fitness. When the food is too large (i.e., the current individual is too different from the optimal individual), the crayfish individual decomposes the food according to formula 15:

[0137]

[0138] By reducing the attractiveness of food to individuals, the over-reliance of individuals on the current optimal individual is reduced, prompting individuals to explore other possible solutions. Then the individual will simulate the crayfish alternatingly using the second and third pairs of claws to eat, and the formula is:

[0139]

[0140] When the food size is appropriate, the crayfish will directly eat, and the specific formula is:

[0141]

[0142] where p is the probability of distribution related to temperature, which affects the speed of individual eating (i.e. the step length of search) by temperature, and the specific formula is:

[0143]

[0144] S3.8, then the algorithm enters the improved elite individual probability update stage, and the individual position is selectively close to the current optimal individual by probability, and the probability is constantly improved with the iteration, and the specific formula is:

[0145]

[0146] where the ind probability formula is:

[0147]

[0148] S3.9, then the algorithm enters the improved boundary correction stage, and the individuals exceeding the upper and lower bounds in the iteration process will be redistributed to the vicinity of the boundary with a certain probability, and the strategy of directly setting the value of the out-of-bound individual to the boundary in the basic COA enhances the global search ability.

[0149] S3.10, then compare the current position of each individual with their respective best position and the historical optimal of the group, dynamically update the optimal solution of fitness, record and iterate until the termination condition is met. The optimal warehouse task slot allocation under multi-objective is obtained, and part of the warehouse slot allocation is as shown in Figure 4 . The slot occupation after allocation is as shown in Figure 5 , wherein the gray slot is the original goods, and the green slot is the warehouse task. The fitness function decline curve of COA, IOC, GA and PSO algorithm for solving the same mathematical model is as shown in Figure 6 .

Claims

1. A warehouse location allocation optimization method based on crayfish optimization algorithm, characterized in that, The method comprises the steps of: S1, obtaining stereoscopic warehouse information and cargo information; S2, establishing a multi-objective warehouse location allocation mathematical model; S3, calculating an optimal warehouse location allocation result; The multi-objective warehouse location allocation mathematical model is solved by a crayfish optimization algorithm to obtain the optimal warehouse location allocation result. The crayfish optimization algorithm solving in the step S3 comprises the following steps: S3.1, initializing parameters: setting the number of algorithm iterations, population size, stereoscopic warehouse size, warehouse location size, and speed of the stacker in each direction; S3.2, randomly generating part of the existing cargo occupied warehouse to simulate the conditions when the real warehousing task is performed; S3.3, adopting real number coding to take the ratio of the warehouse location index of the to-be-warehoused task to all warehouses as an individual value, and taking the to-be-warehoused task quantity as an individual dimension; after randomly generating an initial population, the initial fitness value is calculated and the population optimal individual and the fitness value thereof are recorded, and the values are assigned to the global optimal individual and the fitness value thereof; S3.4, starting iteration, first calculating according to the formula, and taking the mean value as the position of the cave; after the environmental temperature temp is determined, the subsequent individual behavior is generated; S3.5, when the environmental temperature is greater than 30 degrees, half of the individuals enter the summering stage, and the individuals will move towards the cave to avoid high temperature; improvements are made to the crayfish optimization algorithm, mainly replacing the scaling factor C2 with a logistic curve and replacing the rand random number step with a Levy flight, and the moving formula is: The C2 formula is: Levy(n, m, β) is an n*m-dimensional Levy step matrix with a parameter β, and the formula is: Wherein: u n×m ~N(0,σ u ) v n×m ~N(0,1) S3.7, when the temperature is lower than 30 degrees, the individual enters the feeding stage, and the food size is first defined according to the formula: S3.6, when the temperature is higher than 30 degrees, the other half of the individual enters the competition mode, and the stage takes the difference between the current individual and another random individual as the disturbance and the position X of the cave shade The addition simulates the competition of crayfish individuals for the same cave, increases the variation of individuals, and prevents falling into local optimal solution, and the formula is: The food Q measures the size of the current individual fitness relative to the optimal individual fitness; when the food is too large, the crayfish individual decomposes the food according to the formula: By reducing the attraction of food to the individual, the over-reliance of the individual on the current optimal individual is reduced, prompting the individual to explore other possible solutions; then the individual simulates the crayfish alternatingly using the second and third pairs of claws to eat, and the formula is: When the food size is appropriate, the crayfish will directly eat, and the specific formula is: Wherein p is a distribution probability related to the temperature, and the temperature affects the speed of the individual eating, and the specific formula is: S3.8, then the algorithm enters the improved elite individual probability updating stage, and the individual position is selectively close to the current optimal individual through probability, and the probability is continuously improved with the iteration, and the specific formula is: Wherein the ind probability formula is: S3.9, then the algorithm enters the improved boundary correction stage, and the individuals exceeding the upper and lower boundaries in the iteration process will be re-distributed to the vicinity of the boundary with a certain probability, and the strategy of directly setting the value of the individual exceeding the boundary to the boundary in the basic COA strengthens the global search ability. ​ S3.10, then compare each individual's current position with their respective optimal position and the group historical optimum, dynamically update the fitness optimal solution, record and iterate until the termination condition is met, and obtain the optimal warehouse task bin allocation under multi-objective.

2. The method of claim 1, wherein, The single-target bin allocation mathematical model comprises an in-out warehouse efficiency model F1.

3. The method of claim 2, wherein, The single-target bin allocation mathematical model comprises a whole shelf gravity center low model F2.

4. The method of claim 3, wherein, The single-target bin allocation mathematical model comprises a single-side shelf anti-overturning model F3.

5. The method of claim 4, wherein, The single-target bin allocation mathematical model comprises a double-stretch bin priority same kind of goods model F4.

6. The method of claim 1, wherein, The step S2 comprises adding constraint conditions to the multi-objective bin allocation mathematical model.

Images