A WMS storage space allocation method for a four-way densely populated warehouse

Abstract

The present invention relates to the technical field of automated warehouse system development, and in particular to a WMS storage location allocation method for a four-way vehicle-intensive warehouse, specifically a storage location allocation system and method for a four-way vehicle-intensive warehouse, wherein the system comprises: a storage location information maintenance server, a storage location allocation server, a model storage database, and a four-way intelligent shuttle system, wherein the storage location allocation server interacts with the storage location information maintenance server, the model storage database, and the four-way intelligent shuttle system for data respectively; the storage location allocation server interacts with the conveying system for data transportation outside the dense warehouse. The present invention achieves the beneficial technical effect of automatically outputting a storage location allocation plan through a constructed storage location allocation model, and achieves the beneficial technical effect of automatically executing the storage location allocation plan through data interaction between the constructed storage location allocation server and the four-way intelligent shuttle system.

Description

Technical Field

[0001] The present invention relates to the technical field of automated warehouse system development, and in particular to a WMS storage location allocation method for a four-way vehicle-intensive warehouse. Background Art

[0002] The storage space allocation of the four-way vehicle-intensive warehouse (Automated Storage and Retrieval System, AS / RS) and the Warehouse Management System (WMS) has become the cutting-edge technology of modern warehouse management.

[0003] The automated warehouse system introduces AS / RS technology, which fully automates the storage, retrieval and movement of goods through unmanned guided robots, conveyor systems and intelligent control systems.

[0004] The currently disclosed storage location allocation technologies are as follows: Xu Rui and others from the School of Business of Hohai University published an article entitled "Integrated Optimization of Scheduling and Location Allocation of Automated Warehouses under Classified Storage" in "Industrial Engineering and Management" in 2023. They adopted the integrated optimization problem of AS / RS's incoming storage task scheduling and location allocation with a classified storage strategy, proposed a mathematical programming heuristic algorithm combining integer programming and cultural gene algorithm, designed a new incoming storage task matching index based on Hamming distance, and converted this problem into an assignment problem to solve the incoming storage task sorting problem.

[0005] In 2022, Chen Jianxin and others from the 713th Institute of China Shipbuilding Industry Corporation published an article titled "Optimization Algorithm of Multi-vehicle Automated Storage and Retrieval System Based on Monte Carlo Tree Search" in "Ship Science and Technology". They established an optimization algorithm for multi-vehicle automated storage and retrieval system based on Monte Carlo tree search. The algorithm establishes a cargo space optimization model and combines it with an improved Monte Carlo tree search algorithm to solve cargo space optimization.

[0006] While the aforementioned methods each address the problem of allocating goods in their own unique ways, while the four-way intensive warehouse has significantly improved warehouse automation compared to the four-way intensive warehouse, realizing its full potential lies in intelligent location allocation. Traditional static location allocation cannot adapt to rapidly changing inventory demands, making dynamic location allocation technology essential. The dynamic location allocation strategy of the four-way intensive warehouse intelligently determines the storage location of goods by comprehensively considering inventory data, order requirements, and warehouse space availability. This not only improves the storage efficiency of the intensive warehouse, but also speeds up the retrieval of goods and reduces the error rate of retrieval. Summary of the Invention

[0007] The present invention constructs a storage location allocation system for a four-way dense vehicle warehouse and designs a storage location allocation method for the four-way dense vehicle warehouse, which can realize automatic output of storage location allocation scheme on one hand and automatic execution of storage location allocation scheme on the other hand.

[0008] A method for allocating storage spaces in a four-way densely populated parking garage comprises the following steps:

[0009] Step S1: Establish a data interaction protocol between the storage location allocation server Sla and the four-way intelligent shuttle system Sagv. The specific process is as follows:

[0010] The location allocation server Sla sets the following initialization parameters: p is a prime number, g is a generator on the finite field GF(p),

[0011] Four-way intelligent shuttle AGV a Number its machine N m-AGVa The data is sent to the location allocation server Sla, which performs the following operations:

[0012] (1) Randomly select an interaction parameter

[0013] (2) Calculation of interaction parameters

[0014] (3) Set interaction parameter group I = {four-way intelligent shuttle AGV a 、Machine number N m-AGVa , interaction parameter P AGVa-Ⅰ , interaction parameter P AGVa-Ⅱ}Store in the model storage database DBms;

[0015] (4) Set interaction parameter group II = {interaction parameter P AGVa-Ⅰ , interaction parameter P AGVa-Ⅱ Return to the four-way intelligent shuttle AGV a ;

[0016] Step S2: When the conveyor system Ts transports the incoming materials listed on the bill of materials to the entrance of the intensive storage, the conveyor system Ts establishes a communication connection with the storage location allocation server Sla and sends the bill of materials with the approval tag and the machine number to the storage location allocation server Sla. The storage location allocation server Sla reads the incoming materials from the bill of materials and extracts the basic attributes and usage frequency of the incoming materials as input parameters.

[0017] Step S3: The storage location information maintenance server Slim encapsulates the storage location information to obtain a storage location data group with a time tag, and then sends it to the storage location allocation server Sla. The storage location allocation server Sla reads the storage location idle status and storage location idle position in the storage location data group and uses them as input parameters;

[0018] Step S4: The storage location allocation server S1a first assembles the input parameters in steps S2 and S3 to obtain a data parameter group, and then sends the data parameter group to the model storage database DBms;

[0019] The model storage database DBms first inputs the two extracted input parameters into the storage location allocation model to output the storage location allocation plan, and then returns the storage location allocation plan to the storage location allocation server Sla;

[0020] Step S5: The storage location allocation server S1a issues a material warehousing task to the four-way intelligent shuttle system Sagv, and the four-way intelligent shuttle vehicle in the four-way intelligent shuttle system Sagv that can execute the material warehousing task responds to the storage location allocation server S1a;

[0021] The warehouse allocation server Sla first prioritizes the execution of tasks for all the four-way intelligent shuttle vehicles based on the warehouse allocation plan, and then selects the optimal four-way intelligent shuttle vehicle, which is recorded as AGV a , for the four-way intelligent shuttle AGV a The specific process of conducting qualification review for task execution is as follows:

[0022] Step S5-1, the storage location allocation server S1a assigns the four-way intelligent shuttle AGV to the a Send qualification review request, four-way intelligent shuttle AGV a answer;

[0023] Step S5-2, four-way intelligent shuttle AGV a Do the following:

[0024] (1) Generate a request task list Lrt and randomly select a request task list code

[0025] (2) Calculate the label coefficient of the request task list Lrt

[0026] (3) Calculate the label parameters of the request task list Lrt

[0027] Among them, function H() is a hash function;

[0028] (4) Calculate the label parameters of the request task list Lrt

[0029] (5) Output audit label LE AGVa ={machine number N m-AGVa , request task list Lrt, label parameters Tag Parameters Send it to the storage location allocation server Sla;

[0030] In step S5-3, the storage location allocation server S1a performs the following operation process, which is as follows:

[0031] (1) From the review tag LE AGVa Extract the machine number N m-AGVa , request task list Lrt, label parameters Tag Parameters

[0032] (2) According to the machine number N m-AGVa Call the interaction parameter group I to the model storage database DBms and extract the interaction parameter P from it AGVa-Ⅱ ;

[0033] (3) Calculate label coefficient

[0034] (4) Calculate the label parameters of the request task list Lrt-a

[0035] (5) Comparison and Are they equal? ​​If they are equal, the four-way intelligent shuttle AGV is requested to perform the task. a Passed the review and obtained the qualification to perform material warehousing tasks;

[0036] Step S5-4, the storage location allocation server S1a sends the storage location allocation plan to the four-way intelligent shuttle AGV a , four-way intelligent shuttle AGV a Execute the storage location allocation plan.

[0037] Preferably, the method for generating the audit label of the bill of materials BOMi in step S2 is as follows:

[0038] The transport system Ts first calls the interactive parameters Regenerate BOM code Then perform the following calculation operations:

[0039] (1) Label coefficient

[0040] (2) Label parameters

[0041] (3) Tag parameters

[0042] Finally, the transport system Ts obtains the audit label of the bill of materials BOMi {label parameter Tag Parameters }.

[0043] Preferably, the packaging processing method of the storage location data group DGt1 with the time tag in step S3 is as follows:

[0044] The storage location information maintenance server Slim first calls the interaction parameters Regenerate the storage location information I at time T li-T Package code Then perform the following calculation operations:

[0045] (1) Packaging coefficient

[0046] (2) Package parameters

[0047] (3) Packaging parameters

[0048] Finally, the location information maintenance server Slim obtains the location data group DG with time tags tl ={Storage location information I li-T , package parameters Package parameters

[0049] Preferably, the method for assembling the data parameter group PGd in step S4 is as follows:

[0050] Location Allocation Server Sla: On the one hand, call the interaction parameters On the other hand, the input parameters in step S2 and step S3 are called, and the input parameter in step S2 is recorded as P i-Ⅰ The input parameter in step S3 is recorded as P i-Ⅱ ; The third aspect generates assembly code Then perform the following calculation steps:

[0051] (1) Assembly coefficient

[0052] (2) Assembly parameters Among them, || is the string connection symbol;

[0053] (3) Assembly parameter P a-PGd-Ⅱ =C PGd +P a-PGd-Ⅰ ×P Sla-Ⅰ ;

[0054] Finally, the storage location allocation server Sla obtains the data parameter group PGd = {input parameter P i-Ⅰ 、Input parameter P i-Ⅱ , assembly parameter P a-PGd-Ⅰ , assembly parameter P a-PGd-Ⅱ}.

[0055] Preferably, the storage location allocation model in step S4 includes: a task classification module, a material classification module and a storage location allocation module, and any module in the storage location allocation model has data interaction with other modules.

[0056] Preferably, the allocation algorithm of the storage location allocation model in step S4 is:

[0057] Step 1, set initialization parameters;

[0058] Step 2, Task Classification: Use the Task Classification module to classify the tasks of materials to be put into storage in the storage location allocation tasks;

[0059] Step 3, Material Classification: Based on the weight of each material's basic attributes and usage frequency, the material classification module classifies the materials into corresponding task categories. In the next step of task dispatching, if the task is already waiting in the queue in the previous scheduling, the material classification module will add the task category.

[0060] In Step 4, the location allocation module performs the following operations:

[0061] Step 4-1, create the initial population: randomly construct the initial population;

[0062] Step 4-2, calculate fitness: take the transportation time and the storage resources required for the material as the optimization objectives, calculate the fitness of each individual, and select the next generation of individuals based on the fitness function;

[0063] Step 4-3: Increase the fitness function of the previously scheduled task gene in the queue to be assigned by 10%;

[0064] Step 4-4, selection operation: from the chromosomes in a population, some chromosomes are selected for reproduction;

[0065] Step 4-5, crossover operation: Parts of the chromosomes are randomly replaced; using the partial matching crossover method, various parts of the parent chromosomes are selected to produce offspring;

[0066] Steps 4-6, mutation operation: relying on the strong randomness of mutation, expand the search range of excellent chromosomes and supplement the crossover operation;

[0067] Steps 4-7: Repeat the above steps until the optimal storage plan is obtained or the process reaches the maximum number of iterations, and then terminate.

[0068] A storage location allocation system for a four-way dense vehicle warehouse, the system is used for the above-mentioned storage location allocation method, comprising: a storage location information maintenance server Slim, a storage location allocation server Sla, a model storage database DBms, and a four-way intelligent shuttle vehicle system Sagv;

[0069] The storage location allocation server Sla interacts with the storage location information maintenance server Slim, the model storage database DBms, and the four-way intelligent shuttle system Sagv respectively;

[0070] The storage location allocation server S1a interacts with the transportation system Ts for data transmission, and the transportation system is used to transport materials outside the dense storage.

[0071] Based on the above-mentioned storage location allocation system for the four-way vehicle dense storage, the storage location information maintenance server Slim is used to maintain the storage location information in the dense storage;

[0072] The model storage database DBms is used to store the location allocation model and interaction parameters;

[0073] The storage location allocation server S1a is used to allocate storage locations for incoming materials, specifically as follows:

[0074] (1) Extract the basic properties and usage frequency of the material and use them as input parameters P i-Ⅰ ;

[0075] (2) Call the storage location information in the storage location information maintenance server Slim, and obtain the storage location idle status and storage location idle position as input parameters P i-Ⅰ ;

[0076] (3) Calling the storage location allocation model in the model storage database DBms;

[0077] (4) Based on the input parameter P i-Ⅰ and input parameter P i-Ⅰ , use the warehouse location allocation model to output an optimal warehouse location allocation plan;

[0078] (5) Assign the storage location allocation plan to the four-way intelligent shuttle vehicle system Sagv;

[0079] The four-way intelligent shuttle vehicle system Sagv is used to implement the storage location allocation plan.

[0080] The application of the above-mentioned storage space allocation method for the four-way vehicle-intensive warehouse in the automated warehouse system.

[0081] The present invention has the following beneficial effects:

[0082] On the one hand, the basic properties and usage frequency of the materials to be stored are provided through the constructed conveying system; on the other hand, the idle status and idle position information of the warehouse are provided through the constructed warehouse information maintenance server; thirdly, the beneficial technical effect of automatically outputting the warehouse allocation plan is achieved through the constructed warehouse allocation model; finally, the beneficial technical effect of automatically executing the warehouse allocation plan is achieved through the data interaction between the constructed warehouse allocation server and the four-way intelligent shuttle system. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] Figure 1 This is an architectural diagram of a storage space allocation system for a four-way densely populated parking garage.

[0084] Figure 2 This is a workflow diagram of a warehouse location allocation strategy model based on genetic algorithm;

[0085] Figure 3 Flowchart of the partial matching crossover algorithm;

[0086] Figure 4 Schematic diagram of a gene position swapping method for a mutation operation. DETAILED DESCRIPTION

[0087] Example 1:

[0088] Construct a four-way vehicle-intensive storage space allocation system, such as Figure 1 As shown, the system includes: a storage location information maintenance server Slim, a storage location allocation server Sla, a model storage database DBms, and a four-way intelligent shuttle system Sagv;

[0089] The storage location allocation server Sla interacts with the storage location information maintenance server Slim, the model storage database DBms, and the four-way intelligent shuttle system Sagv;

[0090] The storage location allocation server Sla interacts with the conveying system Ts for data transport outside the dense warehouse, including transporting incoming materials to the dense warehouse entrance and transferring outgoing materials from the dense warehouse exit to a designated location.

[0091] The storage location information maintenance server Slim is used to maintain the storage location information in the dense warehouse, which specifically includes the storage location occupancy status, the storage location allocated and unoccupied status, the storage location idle status and the storage location idle position;

[0092] The model storage database DBms is used to store the location allocation model and interaction parameters;

[0093] The location allocation server Sla is used to allocate the locations of incoming materials, as follows:

[0094] Step S1-1, extract the basic properties and usage frequency of the material and use them as input parameters P i-Ⅰ ;

[0095] Step S1-2: call the storage location information in the storage location information maintenance server Slim, and obtain the storage location idle status and storage location idle position, which are used as input parameters P i-Ⅰ ;

[0096] Step S1-3, calling the storage location allocation model in the model storage database DBms;

[0097] Step S1-4, based on the input parameter P i-Ⅰ and input parameter P i-Ⅰ , use the warehouse location allocation model to output an optimal warehouse location allocation plan;

[0098] Step S1-5, assigning the storage location allocation plan to the four-way intelligent shuttle vehicle system Sagv;

[0099] The four-way intelligent shuttle vehicle system Sagv is used to implement the storage location allocation plan.

[0100] Example 2:

[0101] The storage location allocation method based on the storage location allocation system for the four-way dense vehicle storage in Example 1 specifically includes the following steps:

[0102] Step S1: Establishing a data interaction protocol between the storage location allocation server S1a and other functional entities and the transportation system in the storage location allocation system. The specific process is as follows:

[0103] The location allocation server Sla implements the following initialization parameter settings: set p to be a prime number, g to be a generator on the finite field GF(p),

[0104] Based on this, define the data interaction parameters between each functional body of the system and the conveying system and the location allocation server Sla, and perform the following operations:

[0105] (1) The conveying system Ts is numbered N m-Ts Send to the storage location allocation server Sla;

[0106] The location allocation server Sla randomly selects an interaction parameter The interaction parameters are calculated based on this

[0107] The storage location allocation server Sla sends the interaction parameter group {machine number N m-Ts, interaction parameter P Ts-Ⅰ , interaction parameter P Ts-Ⅱ} is stored in the model storage database DBms, and the interaction parameters P Ts-Ⅰ and interaction parameter P Ts-Ⅱ Return to the conveying system Ts;

[0108] (2) The storage location information maintenance server Slim sets its equipment number M e-Slim Send to the storage location allocation server Sla;

[0109] The location allocation server Sla randomly selects an interaction parameter The interaction parameters are calculated based on this

[0110] The storage location allocation server Sla sends the storage location information maintenance server Slim's interactive parameter group {device number M e-Slim , interaction parameter P Slim-Ⅰ , interaction parameter P Slim-Ⅱ} is stored in the model storage database DBms, and the interaction parameters P Slim-Ⅰ and interaction parameter P Slim-Ⅱ Send to the storage location information maintenance server Slim;

[0111] (3) The storage location allocation server Sla assigns its equipment number N e-Sla Send to the model storage database DBms;

[0112] The model storage database DBms randomly selects an interaction parameter The interaction parameters are calculated based on this

[0113] Model storage database DBms repository location allocation server Sla interaction parameter group {device number N e-Sla , interaction parameter P Sla-Ⅰ , interaction parameter P Sla-Ⅱ}, and the interaction parameter P Sla-Ⅰ and interaction parameter P Sla-Ⅱ Send to the storage location allocation server Sla;

[0114] (4) When any four-way intelligent shuttle AGV a When requesting the storage location allocation server Sla to execute the transportation task, the four-way intelligent shuttle AGV a Number its machine N m-AGVa Send to the storage location allocation server Sla;

[0115] The location allocation server Sla randomly selects an interaction parameter The interaction parameters are calculated based on this

[0116] The storage location allocation server Sla will use the four-way intelligent shuttle AGV a The interaction parameter group {machine number N m-AGVa , interaction parameter P AGVa-Ⅰ , interaction parameter P AGVa-Ⅱ} is stored in the model storage database DBms, and the interaction parameters P AGVa-Ⅰ and interaction parameter P AGVa-Ⅱ Sent to the four-way intelligent shuttle AGV a ;

[0117] Step S2, the specific implementation process of the four-way vehicle-intensive storage space allocation system is as follows:

[0118] (1) When the transport system Ts is listed in the bill of materials BOMi (BOMi∈{0,1} * ) is transported to the entrance of the intensive warehouse, the transport system Ts establishes a communication connection with the warehouse allocation server Sla, and sends the audit tag {tag parameter Tag Parameters }Bill of Materials BOMi and machine number N m-Ts Send to the storage location allocation server Sla;

[0119] The method for generating the review label of the above-mentioned bill of materials BOMi is as follows:

[0120] Conveyor system Ts generates a bill of materials code Calculate the label coefficient of the bill of materials BOMi based on this

[0121] Furthermore, the transport system Ts performs the following calculations, the specific process is as follows:

[0122] Label parameters of BOMi

[0123] Label parameters of BOMi

[0124] Conveyor system Ts outputs the audit label of the bill of materials BOMi {label parameter Tag Parameters };

[0125] Among them, function H() is a hash function;

[0126] (2) When the storage location allocation server Sla receives the data with the audit tag {tag parameter Tag Parameters }Bill of Materials BOMi and Machine Number Nm-Ts When the location allocation server Sla performs the following operation process, the details are as follows:

[0127] The storage location allocation server Sla allocates the machine number N of the conveyor system Ts according to m-Ts The interaction parameter set {machine number N} of the conveyor system Ts is called from the model storage database DBms. m-Ts , interaction parameter P Ts-Ⅰ , interaction parameter P Ts-Ⅱ}, and extract the interaction parameter P from it Ts-Ⅱ ;

[0128] Based on this, the storage location allocation server Sla performs the following calculations:

[0129] Label coefficient of the bill of materials BOMi

[0130] Label parameters of BOMi

[0131] like The bill of materials BOMi is approved, and then the storage location allocation server Sla reads the material Mi to be put into storage in the bill of materials BOMi;

[0132] The storage location allocation server Sla extracts the basic attributes and usage frequency of the material Mi to be stored and uses them as input parameters P i-Ⅰ ;

[0133] (3) When the location allocation server Sla starts to allocate the location of the incoming material Mi transported by the conveying system Ts, the location allocation server Sla calls the location information in the dense warehouse from the location information maintenance server Slim. The location information is recorded as I li , the warehouse location information maintenance server Slim will store the warehouse location information I li Perform packaging processing to obtain the storage location data group DG with time tags tl , and the location data group DG with time tag tl , and device number M e-Slim Return to the storage location allocation server Sla;

[0134] The packaging processing method of the above-mentioned storage location data group DGtl with time tags is as follows:

[0135] The storage location information maintenance server Slim generates the storage location information I at time T li-T Package code Based on this, calculate the storage location information I at time T li-T Packaging factor

[0136] Furthermore, the location information maintenance server Slim performs the following calculations. The specific process is as follows:

[0137] Storage location information I li-T Package parameters

[0138] Storage location information I li-T Package parameters

[0139] The storage location information maintenance server Slim outputs the storage location data group DG with time stamp tl ={Storage location information I li-T , package parameters Package parameters };

[0140] (4) When the location allocation server Sla receives the location data group DG with time tag returned by the location information maintenance server Slim tl and device number M e-Slim When the location allocation server Sla performs the following operation process, the details are as follows:

[0141] The storage location allocation server Sla is based on the equipment number M e-Slim , call the interactive parameter group {equipment number M e-Slim , interaction parameter P Slim-Ⅰ , interaction parameter P Slim-Ⅱ}, from which the interaction parameter P is extracted Slim-Ⅱ ;

[0142] Based on this, the storage location allocation server Sla performs the following calculations:

[0143] Unblocking coefficient

[0144] Unpack parameter P u-DGtl =H(C u-DGtl , DGtl);

[0145] like This means that the storage location information I at time T li-T After the complete transmission, the storage location allocation server S1a reads the storage location idle status and storage location idle position in the storage location data group DGtl and uses it as the input parameter P i-Ⅱ ;

[0146] (5) The location allocation server Sla takes the input parameter P i-Ⅰ and input parameter P i-Ⅱ Assemble to get a data parameter group PGd, and add the data parameter group PGd and device number N e-SlaSend it to the model storage database DBms, and the model storage database DBms extracts the input parameter P i-Ⅰ and input parameter P i-Ⅱ , input it into the warehouse location allocation model and output the warehouse location allocation plan Pla;

[0147] The model storage database DBms returns the storage location allocation plan Pla to the storage location allocation server Sla;

[0148] The specific assembly method of the above data parameter group PGd is as follows:

[0149] The location allocation server Sla generates an assembly code The assembly coefficient is calculated based on this

[0150] Furthermore, the storage location allocation server S1a performs the following calculations. The specific process is as follows:

[0151] Assembly parameter P a-PGd-Ⅰ =H(C a-CPGd , P i-Ⅰ ||P i-Ⅱ ); where || is the string concatenation symbol;

[0152] Assembly parameter P a-PGd-Ⅱ =C PGd +P a-PGd-Ⅰ ×P Sla-Ⅰ ;

[0153] The storage location allocation server Sla outputs the data parameter group PGd = {input parameter P i-Ⅰ 、Input parameter P i-Ⅱ , assembly parameter P a-PGd-Ⅰ , assembly parameter P a-PGd-Ⅱ};

[0154] (6) When the model storage database DBms receives the data parameter group PGd and equipment number N sent by the storage location allocation server Sla e-Sla When the model storage database DBms performs the following operation process, the details are as follows:

[0155] Model storage database DBms according to device number N e-Sla , call the interactive parameter group {equipment number N e-Sla , interaction parameter P Sla-Ⅰ , interaction parameter P Sla-Ⅱ}, from which the interaction parameter P is extracted Sla-Ⅱ ;

[0156] Afterwards, the model storage database DBms performs the following specific calculations:

[0157] Assembly factor

[0158] Assembly parameter P' a-PGd-Ⅰ =H(C' a-CPGd , PGd);

[0159] If P' a-PGd-Ⅰ =P a-PGd-Ⅰ , it means that the model storage database DBms has received the complete input parameter P sent by the storage location allocation server Sla. i-Ⅰ and input parameter P i-Ⅱ , the model storage database DBms uses the storage location allocation model to output the storage location allocation plan Pla;

[0160] (7) When the location allocation server Sla receives the location allocation plan Pla sent by the model storage database DBms, the location allocation server Sla extracts the <material Mi-location Li> correspondence contained in the location allocation plan Pla;

[0161] The warehouse allocation server Sla issues a material entry task to the four-way intelligent shuttle system Sagv. After that, several four-way intelligent shuttles that can perform the material entry task respond to the warehouse allocation server Sla. The warehouse allocation server Sla prioritizes the task execution of all responding four-way intelligent shuttles based on the warehouse allocation plan Pla, and selects the best four-way intelligent shuttle, which is recorded as AGV a , conduct qualification review on its task execution;

[0162] Storage location allocation server Sla for four-way intelligent shuttle AGV a The specific process of conducting qualification review for task execution is as follows:

[0163] Storage location allocation server Sla to four-way intelligent shuttle AGV a Send qualification review request, four-way intelligent shuttle AGV a answer;

[0164] Four-way intelligent shuttle AGV a Generate a request task list Lrt and randomly select a request task list code

[0165] Based on this, the label coefficient of the request task list Lrt is calculated

[0166] Further, the four-way intelligent shuttle AGV a Perform the following calculations. The specific process is as follows:

[0167] Request tag parameters for task list Lrt

[0168] Request tag parameters for task list Lrt

[0169] Finally, the four-way intelligent shuttle AGV a Output the audit label of the request task list Lrt {label parameter Tag Parameters };

[0170] Four-way intelligent shuttle AGV a will be labeled with the audit {label parameter Tag Parameters }'s request task list Lrt and machine number N m-AGVa Send to the storage location allocation server Sla;

[0171] When the location allocation server Sla receives the four-way intelligent shuttle AGV a Sent with review label {label parameter Tag Parameters }'s request task list Lrt and machine number N m-AGVa When the location allocation server Sla performs the following operation process, the details are as follows:

[0172] The storage location allocation server Sla is based on the machine number N m-AGVa , call the four-way intelligent shuttle AGV to the model storage database DBms a The interaction parameter group {machine number N m-AGVa , interaction parameter P AGVa-Ⅰ , interaction parameter P AGVa-Ⅱ}, from which the interaction parameter P is extracted AGVa-Ⅱ ;

[0173] Based on this, the storage location allocation server Sla performs the following calculations:

[0174] Request the label coefficient of the task list Lrt-a

[0175] Request tag parameters for task list Lrt-a

[0176] like Then request the four-way intelligent shuttle AGV to perform the task a Obtained the qualification to perform material warehousing tasks;

[0177] Afterwards, the location allocation server Sla transmits the location allocation plan Pla to the four-way intelligent shuttle AGV a , four-way intelligent shuttle AGV a Execute the storage location allocation plan Pla and transport the material Mi to the corresponding storage location Li.

[0178] Example 3:

[0179] For the location allocation model in the model storage database DBms, this embodiment designs a location allocation strategy model based on genetic algorithm, which is used for the allocation of dense warehouse locations. The specific process is as follows: Figure 2 As shown;

[0180] To describe a four-way densely populated warehouse, this embodiment initializes the warehouse as follows: Assume that the warehouse has H rows of shelves, each row has I columns and J layers. Each location is numbered N, where N is a positive integer. The location number N is converted into spatial coordinates, which are stored as (x, y, z).

[0181] For example, let the location number of material i be N. The coordinates of the location are expressed as follows:

[0182]

[0183] y = (N-1)%I+1 (3-2)

[0184]

[0185] in, Indicates rounding up operation, % indicates remainder operation; H, I, and J are the number of rows, columns, and layers of the warehouse respectively;

[0186] This embodiment sets the following conditions:

[0187] (1) The entrance and exit of the warehouse are on the same side;

[0188] (2) Goods are stored on pallets and are suitable for the cargo space;

[0189] (3) The types of products stored are known;

[0190] (4) One cargo space can only store goods on one pallet;

[0191] (5) The mass of the product is evenly distributed, and the center of gravity is located at the geometric center of the cargo space;

[0192] (6) The time for starting, braking, turning and picking the machine is negligible;

[0193] (7) The machine runs at a constant speed in the horizontal and vertical directions, ignoring the acceleration and deceleration process;

[0194] Based on the above initialization condition settings, this embodiment designs a warehouse location allocation strategy model based on a genetic algorithm. The warehouse location allocation model includes a task classification module, a material classification module, and a warehouse location allocation module. Each module in the warehouse location allocation model interacts with other modules in terms of data.

[0195] The specific implementation process of the location allocation model is as follows:

[0196] Step S1, initialization parameter settings: task category C = 3, initial population size α = 100, crossover probability P c =0.8, mutation probability P m =5%, maximum number of iterations λ max =500;

[0197] Step S2, task classification: using the task classification module to classify the tasks of materials to be put into storage in the storage location allocation tasks;

[0198] Step S3: The material classification module classifies the materials into corresponding task categories based on the basic attributes and usage frequency weights of each material;

[0199] In the next step of task dispatch, if the task is already waiting in the queue in the previous dispatch, the material classification module will add the task category;

[0200] Step S4: The location allocation module performs the following operations:

[0201] Step S4-1, creating an initial population: randomly constructing an initial population;

[0202] Step S4-2, calculate fitness: take the transportation time and the storage resources required for the material as the optimization objectives, calculate the fitness of each individual, and select the next generation of individuals based on the fitness function;

[0203] Step S4-3: If there are previously scheduled tasks in the queue to be assigned, the fitness function of the gene will be increased by 10%. The specific implementation steps are as follows:

[0204] Taking the transportation time and the required storage space resources of the materials as the optimization target, the purpose is to minimize them;

[0205]

[0206] Among them, SC [r] Indicates the size of category r, q is the category number, p is the material number, TC [q,p] represents the material p in category q, and f is a natural number constant. The purpose of adding f is to prevent the fitness function value from tending to infinity when the objective function approaches 0;

[0207] If a task i is waiting in queue Q, the objective function will be improved by multiplying it by 0.9, that is, when there are still materials with unassigned locations in the previous batch of material location allocation tasks, the fitness of the gene will increase by 10%;

[0208]

[0209] Among them, object function represents the objective function, and the minimum level of its value represents the optimality of the chromosome; MM(i) is the minimum number of storage resources required for any material to be put into storage, so that the number of resources required is less and more suitable for storage needs; TT(i) represents the transportation time of the material into storage, which should be minimized;

[0210] Therefore, the calculation method of the fitness function fitness is as shown in formula (3-6):

[0211]

[0212] After determining the fitness of each gene, by asking the material storage task i whether there are remaining materials in the previous allocation queue, if there are remaining materials in the allocation queue, its fitness will be increased by 10% to give it a chance to obtain resources to prevent the occurrence of "deadlock" phenomenon;

[0213] Step S4-4, selection operation: Combining the roulette selection strategy and the tournament selection strategy, this embodiment designs a roulette-tournament selection strategy. This strategy selects excellent individuals while preserving genetic diversity. The specific steps are as follows:

[0214] (1) Sort the fitness of each individual from best to worst, and use the tournament strategy to select the top 30% of the best individuals from the population;

[0215] (2) Among the remaining 70% of individuals selected in step (1), adjacent individuals are compared in pairs using a roulette wheel strategy to randomly select individuals;

[0216] (3) The individual composition of the new generation population obtained through steps (1) and (2) is the same as that of the original population;

[0217] Step S4-5, crossover operation: The model randomly selects two crossover points to determine the crossover region. After performing the crossover, two invalid chromosomes are obtained, and some genes may be repeated.

[0218] Furthermore, in order to repair chromosomes, a matching relationship is established for each chromosome within the crossover region, and then the conflict can be eliminated by applying this matching relationship to the duplicate genes outside the crossover region, e.g. Figure 3 As shown in the figure, the specific cross operation process is as follows:

[0219] (1) Parent selection: The individual selects two individuals from the parent population as the parent P i ;

[0220] Where i is the number of parent generations, i = 1, 2, ..., n1, n1 is the number of parent populations;

[0221] (2) Select intersection points: Randomly select two intersection points, which are two integer positions in this embodiment, to specify the intersection part;

[0222] (3) Create child D1: Copy the part between the intersection points in P1 to D1, keeping the order unchanged; then, insert the elements in P2 that do not appear in D1 into D1 according to the order in P2 to replace the corresponding positions in D1;

[0223] (4) Create child D2: Copy the part between the intersection points in P2 to P2, keeping the order unchanged; then, insert the elements in P1 that do not appear in D2 into D2 according to the order in P1 to replace the corresponding positions in D2;

[0224] (5) Conflict detection: Establish a mapping relationship based on the two groups of genes exchanged;

[0225] Further, if Figure 3 As shown in the figure, the partial matching crossover algorithm: taking the mapping relationship 7-5-2 as an example, in the second step result, there are two genes 7 in D1, which are transformed into gene 2 through the mapping relationship, and so on until there is no conflict;

[0226] Finally, all conflicting genes will be mapped to ensure that the new pair of offspring genes are conflict-free;

[0227] Step S4-6, mutation operation: Figure 4 As shown, first, the population is selected according to the determined mutation probability P m Select the parent chromosome individual that needs to be mutated, and then use a random function to determine the positions m and n of the mutated gene fragments that need to be swapped, completing the genetic material swap to generate a new daughter chromosome;

[0228] Step S4-7, repeat the above steps until the optimal storage solution is found or the process reaches the maximum number of iterations, and then terminate.

[0229] Example 4:

[0230] For the task classification module in the warehouse allocation model, this embodiment independently designs a method for classifying incoming material tasks. This method uses a feedforward neural network to classify incoming material tasks in the warehouse allocation task.

[0231] The feedforward neural network consists of three layers: input layer, hidden layer and output layer. Its main features are as follows:

[0232] Input layer: receives historical incoming material data or features, with each input point corresponding to one feature;

[0233] The nodes in the input layer transmit features to the hidden layer;

[0234] Hidden layer: The core of a feedforward neural network, consisting of one or more layers. Each hidden layer contains multiple nodes that process the input historical incoming material data through weights and activation functions. The hidden layer learns the complex features and patterns of the historical incoming material data.

[0235] The hidden layer transmits the learned complex features of historical incoming material data to the output layer;

[0236] Output layer: contains multiple neurons, each of which corresponds to a model output. The activation value of the neurons in the output layer represents the model's classification result for the material warehousing task.

[0237] The specific process of the feedforward neural network is as follows:

[0238] Step S1, data preparation: partition the historical incoming material data; 70% of the data is used for training, 15% for validation, and 15% for testing; randomly select the data for training, validation, and testing to improve the performance of training, validation, and testing;

[0239] Step S2, training: The training steps of the feedforward neural network on the historical incoming material data are as follows:

[0240] (1) Forward propagation: The input data is passed from the input layer through the hidden layer to the output layer, and the weights and activation function Sigmoid are applied at the same time to generate the model's prediction of the input data;

[0241]

[0242]

[0243] Among them, W [i] For X [i] The weight of X [i] Is the input of the Sigmoid function;

[0244] (2) Calculating loss: Calculating the gap between the model’s predictions and the actual labels is called a loss function. That is, using a loss function to calculate the gap between the network’s predictions and the actual target labels. This method uses cross-entropy as the loss function.

[0245] cross-entropy=﹣∑(y i *log(p i )) (4-3)

[0246] Among them, y represents the actual category label, p represents the model's predicted probability for each category, and y i Indicates the i-th element of the actual category, p i Represents the model's predicted probability for the i-th category;

[0247] (3) Backpropagation: Calculate the gradient of the loss function with respect to weights and biases, and then use gradient descent to update the weights and biases to reduce the loss; use the scaled conjugate gradient method to update the weights and biases of the data;

[0248] (4) Repeat training: Repeat the above steps and terminate training in the following cases: the maximum number of iterations has been created; the time exceeds the maximum level; the performance of the network is lower than a threshold; the gradient of the performance graph is lower than the minimum value; the performance has declined since the last time.

[0249] Example 5:

[0250] For the material classification module in the location allocation model, this embodiment independently designs a dynamic material classification method, which calculates the material weight based on its basic attributes and usage frequency, and dynamically assigns an initial weight to each parameter based on system conditions;

[0251] TW [i] =[WP(A)×A [i] ]+[WP(F)×F [i] ] (5-1)

[0252] Among them, WP(A) is the material attribute A [i] The weight of WP(F) is the material usage frequency F [i] The weight of

[0253] The weight of a material is compared with the average weight of tasks in different categories, and tasks with similar weights are grouped into one class;

[0254]

[0255] Among them, TW [i] is the weight of material i, CW [r] is the average weight of category r, TC [r] is the task in category r, ε is the maximum range of the difference between the set material weight and the average weight of the category tasks;

[0256] According to the weight attribute of each task, it is added to the categories set in Example 4, and 3 categories are set;

[0257] Furthermore, in order to complete efficient resource allocation tasks under limited storage resources, this embodiment independently designs a method for dynamic task classification. That is, if a material allocation task is already waiting in the queue of a previous schedule, the task category will be changed to fully utilize the existing storage resources and save allocation time.

[0258] That is: if C(i) has been waiting in the queue Q of the previous step and i>1, the task will be transferred to C(i-1), that is, C(i)→C(i-1);

[0259] When all tasks are placed in the appropriate categories, the WMS compares the number of free slots in the warehouse, free_slots, with C1_count; if free_slots ≤ C1_count, C1 is sent to the location allocation algorithm for allocation; otherwise, C2 is also sent to the location allocation algorithm for allocation;

[0260] If free_slots>0 after all tasks in C1 and C2 are sent, C3 will continue to be sent;

[0261] Where C(i) represents the category of task i and takes the value of 1, 2 or 3;

[0262] Q represents the queue of the previous step, which contains tasks to be processed;

[0263] C1 represents the set of tasks of category 1;

[0264] C2 represents the set of tasks of category 2;

[0265] C3 represents the task set of category 3;

[0266] C1_count represents the number of tasks in C1;

[0267] free_slots indicates the number of free slots in the warehouse.

Claims

1. A method for allocating storage spaces in a four-way densely populated parking garage, characterized in that: The following steps are involved: Step S1: Establish a data interaction protocol between the storage location allocation server Sla and the four-way intelligent shuttle system Sagv. The specific process is as follows: The location allocation server Sla sets the following initialization parameters: p is a prime number, g is a generator on the finite field GF(p), Four-way intelligent shuttle AGV a Number its machine N m-AGVa The data is sent to the location allocation server Sla, which performs the following operations: (1) Randomly select an interaction parameter (2) Calculation of interaction parameters (3) Set interaction parameter group I = {four-way intelligent shuttle AGV a 、Machine number N m-AGVa , interaction parameter P AGVa-Ⅰ , interaction parameter P AGVa-Ⅱ }Store in the model storage database DBms; (4) Set interaction parameter group II = {interaction parameter P AGVa-Ⅰ , interaction parameter P AGVa-Ⅱ Return to the four-way intelligent shuttle AGV a ; Step S2: When the conveyor system Ts transports the incoming materials listed on the bill of materials to the entrance of the intensive storage, the conveyor system Ts establishes a communication connection with the storage location allocation server Sla and sends the bill of materials with the approval tag and the machine number to the storage location allocation server Sla. The storage location allocation server Sla reads the incoming materials from the bill of materials and extracts the basic attributes and usage frequency of the incoming materials as input parameters. Step S3: The storage location information maintenance server Slim encapsulates the storage location information to obtain a storage location data group with a time tag, and then sends it to the storage location allocation server Sla. The storage location allocation server Sla reads the storage location idle status and storage location idle position in the storage location data group and uses them as input parameters; Step S4: The storage location allocation server S1a first assembles the input parameters in steps S2 and S3 to obtain a data parameter group, and then sends the data parameter group to the model storage database DBms; The model storage database DBms first inputs the two extracted input parameters into the storage location allocation model to output the storage location allocation plan, and then returns the storage location allocation plan to the storage location allocation server Sla; Step S5: The storage location allocation server S1a issues a material warehousing task to the four-way intelligent shuttle system Sagv, and the four-way intelligent shuttle vehicle in the four-way intelligent shuttle system Sagv that can execute the material warehousing task responds to the storage location allocation server S1a; The warehouse allocation server Sla first prioritizes the execution of tasks for all the four-way intelligent shuttle vehicles based on the warehouse allocation plan, and then selects the optimal four-way intelligent shuttle vehicle, which is recorded as AGV a , for the four-way intelligent shuttle AGV a The specific process of conducting qualification review for task execution is as follows: Step S5-1, the storage location allocation server S1a assigns the four-way intelligent shuttle AGV to the a Send qualification review request, four-way intelligent shuttle AGV a answer; Step S5-2, four-way intelligent shuttle AGV a Do the following: (1) Generate a request task list Lrt and randomly select a request task list code (2) Calculate the label coefficient of the request task list Lrt (3) Calculate the label parameters of the request task list Lrt Among them, function H() is a hash function; (4) Calculate the label parameters of the request task list Lrt (5) Output audit label LE AGVa ={machine number N m-AGVa , request task list Lrt, label parameters Tag Parameters }, send it to the storage location allocation server Sla; In step S5-3, the storage location allocation server S1a performs the following operation process, which is as follows: (1) From the review tag LE AGVa Extract the machine number N m-AGVa , request task list Lrt, label parameters Tag Parameters (2) According to the machine number N m-AGVa Call the interaction parameter group I to the model storage database DBms and extract the interaction parameter P from it AGVa-Ⅱ ; (3) Calculate label coefficient (4) Calculate the label parameters of the request task list Lrt-a (5) Comparison and Are they equal? ​​If they are equal, the four-way intelligent shuttle AGV is requested to perform the task. a Passed the review and obtained the qualification to perform material warehousing tasks; Step S5-4, the storage location allocation server S1a sends the storage location allocation plan to the four-way intelligent shuttle AGV a , four-way intelligent shuttle AGV a Execute the storage location allocation plan.

2. The method for allocating storage spaces for a four-way densely populated parking garage according to claim 1, characterized in that: The method for generating the review label of the bill of materials BOMi in step S2 is as follows: The transport system Ts first calls the interactive parameters Regenerate BOM code Then perform the following calculation operations: (1) Label coefficient (2) Label parameters (3) Tag parameters Finally, the transport system Ts obtains the audit label of the bill of materials BOMi {label parameter Tag Parameters }.

3. The method for allocating storage spaces for a four-way densely populated parking garage according to claim 1, characterized in that: The packaging processing method of the storage location data group DGt1 with time tags in step S3 is as follows: The storage location information maintenance server Slim first calls the interaction parameters Regenerate the storage location information I at time T li-T Package code Then perform the following calculation operations: (1) Packaging coefficient (2) Package parameters (3) Packaging parameters Finally, the location information maintenance server Slim obtains the location data group DG with time tags tl ={Storage location information I li-T , package parameters Package parameters }.

4. The method for allocating storage spaces for a four-way densely populated parking garage according to claim 1, characterized in that: The method for assembling the data parameter group PGd in step S4 is as follows: Location Allocation Server Sla: On the one hand, call the interaction parameters On the other hand, the input parameters in step S2 and step S3 are called, and the input parameter in step S2 is recorded as P i-Ⅰ The input parameter in step S3 is recorded as P i-Ⅱ ; The third aspect generates assembly code Then perform the following calculation steps: (1) Assembly coefficient (2) Assembly parameters Among them, || is the string connection symbol; (3) Assembly parameter P a-PGd-Ⅱ =C PGd +P a-PGd-Ⅰ ×P Sla-Ⅰ ; Finally, the storage location allocation server Sla obtains the data parameter group PGd = {input parameter P i-Ⅰ 、Input parameter P i-Ⅱ , assembly parameter P a-PGd-Ⅰ , assembly parameter P a-PGd-Ⅱ }.

5. The method for allocating storage spaces in a four-way densely populated parking garage according to any one of claims 1 to 4, characterized in that: The storage location allocation model in step S4 includes: a task classification module, a material classification module and a storage location allocation module. Any module in the storage location allocation model has data interaction with other modules.

6. The method for allocating storage spaces for a four-way densely packed parking garage according to claim 5, characterized in that: The allocation algorithm of the storage location allocation model in step S4 is: Step 1, set initialization parameters; Step 2, Task Classification: Use the Task Classification module to classify the tasks of materials to be put into storage in the storage location allocation tasks; Step 3, Material Classification: Based on the weight of each material's basic attributes and usage frequency, the material classification module classifies the materials into corresponding task categories. In the next step of task dispatching, if the task is already waiting in the queue in the previous scheduling, the material classification module will add the task category. In Step 4, the location allocation module performs the following operations: Step 4-1, create the initial population: randomly construct the initial population; Step 4-2, calculate fitness: take the transportation time and the storage resources required for the material as the optimization objectives, calculate the fitness of each individual, and select the next generation of individuals based on the fitness function; Step 4-3: Increase the fitness function of the previously scheduled task gene in the queue to be assigned by 10%; Step 4-4, selection operation: from the chromosomes in a population, some chromosomes are selected for reproduction; Step 4-5, crossover operation: Parts of the chromosomes are randomly replaced; using the partial matching crossover method, various parts of the parent chromosomes are selected to produce offspring; Steps 4-6, mutation operation: relying on the strong randomness of mutation, expand the search range of excellent chromosomes and supplement the crossover operation; Steps 4-7: Repeat the above steps until the optimal storage plan is obtained or the process reaches the maximum number of iterations, and then terminate.

7. A storage location allocation system for a four-way densely populated parking garage, the system being used to execute the storage location allocation method described in claim 1, characterized in that: include: Storage location information maintenance server Slim, storage location allocation server Sla, model storage database DBms, four-way intelligent shuttle system Sagv; The storage location allocation server Sla interacts with the storage location information maintenance server Slim, the model storage database DBms, and the four-way intelligent shuttle system Sagv respectively; The storage location allocation server S1a interacts with the transportation system Ts for data transmission, and the transportation system is used to transport materials outside the dense storage.

8. The storage space allocation system for a four-way dense vehicle storage according to claim 7 is characterized in that: The storage location information maintenance server Slim is used to maintain the storage location information in the dense warehouse; The model storage database DBms is used to store the location allocation model and interaction parameters; The storage location allocation server S1a is used to allocate storage locations for incoming materials, specifically as follows: (1) Extract the basic properties and usage frequency of the material and use them as input parameters P i-Ⅰ ; (2) Call the storage location information in the storage location information maintenance server Slim, and obtain the storage location idle status and storage location idle position as input parameters P i-Ⅰ ; (3) Calling the storage location allocation model in the model storage database DBms; (4) Based on the input parameter P i-Ⅰ and input parameter P i-Ⅰ , use the warehouse location allocation model to output an optimal warehouse location allocation plan; (5) Assign the storage location allocation plan to the four-way intelligent shuttle vehicle system Sagv; The four-way intelligent shuttle vehicle system Sagv is used to implement the storage location allocation plan.

9. Application of the storage space allocation method for a four-way dense vehicle storage according to claim 1 in an automated warehouse system.

Images